Está en la página 1de 348

The Control Techniques Drives and Controls Handbook

Bill Drury

The Institution of Electrical Engineers

Published by: The Institution of Electrical Engineers, London, United Kingdom 2001: The Institution of Electrical Engineers This publication is copyright under the Berne Convention and the Universal Copyright Convention. All rights reserved. Apart from any fair dealing for the purposes of research or private study, or criticism or review, as permitted under the Copyright, Designs and Patents Act, 1988, this publication may be reproduced, stored or transmitted, in any forms or by any means, only with the prior permission in writing of the publishers, or in the case of reprographic reproduction in accordance with the terms of licences issued by the Copyright Licensing Agency. Inquiries concerning reproduction outside those terms should be sent to the publishers at the undermentioned address: The Institution of Electrical Engineers, Michael Faraday House, Six Hills Way, Stevenage, Herts. SG1 2AY, United Kingdom While the author and the publishers believe that the information and guidance given in this work are correct, all parties must rely upon their own skill and judgment when making use of them. Neither the author nor the publishers assume any liability to anyone for any loss or damage caused by any error or omission in the work, whether such error or omission is the result of negligence or any other cause. Any and all such liability is disclaimed. The moral right of the author to be identified as author of this work has been asserted by him in accordance with the Copyright, Designs and Patents Act 1988.

British Library Cataloguing in Publication Data Control Techniques drives and controls handbook. (lEE power series; no. 35) 1. Electric motors 2. Electric controllers 3. Electric driving I. Drury, W. I1. Control Techniques Drives PIc 621.4' 6 ISBN 0 85296 793 4

Typeset by Newgen Imaging Systems, India Printed in England by Cambridge University Press, Cambridge


Variable-speed drives remain a key component of the boom in all aspects of automation and energy saving which is becoming of ever greater importance throughout the world. The words of Harry Ward Leonard first uttered on 18 November 1896 in his paper entitled 'Volts versus o h m s - s p e e d regulation of electric motors' still hold true: 'The operation by means of electric motors of elevators, locomotives, printing presses, travelling cranes, turrets on men-of-war, pumps, ventilating fans, air compressors, horseless vehicles and many other electric motor applications too numerous to mention in detail, all involve the desirability of operating an electric motor under perfect and economical control at any desired speed from rest to full speed'. It can, and should, be argued that electrical variable-speed drives have facilitated the automation revolution. They have, like so many enabling technologies, developed rapidly, fuelled by their success, stretched by demands never dreamed possible a generation earlier. The development cycle of drives products is now such that product ranges have expected lifetimes of only three to five years - a problem in itself to many OEM customers whose own products have a much longer design life. The world of variable-speed drives is an exciting and rapidly moving one. To predict the future and the pace of development is difficult. A historical perspective is helpful and, for those who need any convincing, shows how quickly things are moving: 1820 Oersted was the first to note that a compass needle is deflected when an electric current is applied to a wire close to the compass - the fundamental principle of an electric motor. Faraday built two devices to produce what he called electromagnetic rotation: that is a continuous circular motion from the circular magnetic force around a wire. This was the initial stage of his pioneering work.

Figure P. 1 Michael Faraday (1791-1867)





Arago discovered that if a copper disc is rotated rapidly beneath a suspended magnet, then the magnet also rotates in the same direction as the disc. Babbage and Herschel demonstrated the inversion of Arago's experiment by rotating a magnet beneath a pivoted disc causing the disc to rotate. This was truly induced rotation and just a simple step away from the first induction motor, a step which was not taken for half a century. Using an induction ring, Faraday made one of his greatest discoveries - electromagnetic induction: the induction of electricity i n a wire by means of the electromagnetic effect of a current in another wire.



1832 1838









The induction ring was the first electric transformer. In a second series of experiments in the same year he discovered magneto-electric induction: the production of a steady electric current. To do this, Faraday attached two wires through a sliding contact to a copper disc, the first commutator, an approach suggested to him by Amp+re. By rotating the disc between the poles of a horseshoe magnet he obtained a continuous direct current. This was the first generator. Faraday's scientific work laid the foundations of all subsequent electro-technology. From his experiments came devices which led directly to the modem electric motor, generator and transformer. Pixii produced the first magneto-electric machine. Lenz discovered that a D.C. generator could be used equally well as a motor. Jacobi used a battery-fed D.C. motor to propel a boat on the River Neva. Interestingly, Jacobi himself pointed out that batteries were inadequate for propulsion- a problem which is still being worked on today. Wheatstone and Cooke patented the use of electromagnets instead of permanent magnets for the field system of the dynamo. Over twenty years were to elapse before the principle of self excitation was to be established by Wilde, Wheatstone, Varley and the Siemens brothers. Gramme introduced a ring armature somewhat more advanced than that proposed by Pacinotte in 1860, which led to the multibar commutator and the modem D.C. machine. Gramme demonstrated, at the Vienna Exhibition, the use of one machine as a generator supplying power over a distance of 1 km to drive a similar machine as a motor. This simple experiment did a great deal to establish the credibility of the D.C. motor. Bailey developed a motor in which he replaced the rotating magnet of Babbage and Herschel by a rotating magnetic field, produced by switching of direct current at appropriately staggered intervals to four pole pieces. With its rotation induced by a rotating magnetic field it was thus the first commutatorless induction motor. Ferraris produced a motor in which a rotating magnetic field was established by passing single-phase alternating current through windings in space quadrature. This was the first alternating-current commutatorless induction motor, a single-phase machine which Dobrowolsky later acknowledged as the inspiration for his polyphase machine. Tesla developed the first polyphase induction motor. He deliberately generated four-phase polyphase currents and supplied them to a machine which had a four-phase stator. He used several types of rotor, including one with a soft-iron salient-pole construction - a reluctance m o t o r - and one with two short-circuited windings in space quadrature - the polyphase induction motor. Dobrowolsky, working independently from Tesla, introduced the three-phase squirrel-cage induction motor. Dobrowolsky introduced a three-phase induction motor with a polyphase slip-ring rotor into which resistors could be connected for starting and control.

The speed of these motors depends fundamentally upon pole number and supply frequency. Rotor resistance control for the slip-ring motor was introduced immediately, but this is equivalent to armature resistance control of a D.C. machine and is inherently inefficient. By 1890 there was a well established D.C. motor, D.C. central generating stations, three-phase A.C. generation and a simple three-phase motor with enormous potential but which was inherently a single-speed machine. There was as yet no way of efficiently controlling the speed of a motor over the full range, from zero to top speed.

V O L T S VS. OttMS.


Figure P.2 1l Oth meeting of the American Institute of Electrical Engineers, New York, 18 November 1896



The work of Ward Leonard clearly marks the birth of efficient, wide-range, electrical variable-speed drives. The system he proposed was of course based upon the inherently variable-speed D.C. machine (which had hitherto been controlled by variable armature resistors). His work was not universally accepted at the time and attracted much criticism, understandably, as it required three machines of similar rating to do the job of one. Today, however, all D.C. drives are based upon his control philosophy, with only the implementation changing from multimotor schemes through the era of grid-controlled mercury-arc rectifiers to thyristors and, more re-cently, in demanding dynamic applications, to bipolar transistors, field-effect transistors (FETs), insulated-gate bipolar transistors (IGBTs)... Kramer made the first significant move with respect to frequency changing in 1904 by introducing a D.C. link between the slip rings and the A.C. supply. This involved the use of two A.C. ~ D.C. motor sets. The D.C. link was later to become a familiar sight in many A.C. drive technologies. Subsequent advances in A.C. motor speed control were based upon purely electrical means of frequency and voltage conversion. Progress has followed the advances in the field of semiconductors (power and signal/ control).







Otoichs~,rom motor. C,;lkerumform~r, Ahb. !1.

l~,i~enerrofulbG~ H B'aliptstromerrefd n~. Abb. 12.

Neue Methoden zur Regelung yon Asynchronmotoren und ihre Anwendung ffir verschiedene Zwecke. ~)
Vou Ch. Krilmer.

Figure P.3 Elektrotechnische Zeitschrift, volume 31, 30 July 1908

Schrage introduced a system based upon an induction motor with a commutator on the rotor. This machine proved to be very popular, requiring no auxiliary machines, and was very reliable. It found large markets particularly in the textile industry and some other niche applications, and is still sold today although in rapidly reducing numbers. 1923 The introduction of the ignitron made controlled rectification possible. The thyratron and grid-controlled mercury rectifiers made life easier in 1928. This made possible the direct control of voltage applied to the armature of a D.C. machine so as to apply the philosophy of Ward Leonard control without additional machines. 1930 The ideas of inversion (D.C. to variable frequency/ voltage A.C. which is the basis for the present day inverter) had been established, the use of forced commutation by means of switched capacitors was introduced. 1931 Direct A.C. to A.C. conversion by means of cycloconverters was introduced for the railway service. 1932 Nyquist stability criterion developed. 1938 Bode stability criterion developed. 1950 The introduction of silicon into power switches replacing the bulky and relatively inefficient mercury-arc rectifiers (MAR). By 1960 thyristors (SCRs) had become available and the key enabling technology for drives had arrived. D.C. drives and cycloconverters quickly embraced the new silicon technology, at first using techniques with origins in the MAR forerunners. The faster switching performance of the new silicon, however, opened many new doors notably in the field of forced commutation-the way was clear for commercial variablefrequency inverters. 1957 Back to back reversing D.C. drive introduced. 1960s Power semiconductor voltage and current ratings grow and performance characteristics improve. Inverters became commercially viable, notably in 1911

industries such as textiles where a single (bulk) inverter was used to feed large numbers of induction motors (or reluctance motors, despite their low power factor, where synchronisation was required). 1963 Gain-bandwidth relationships of power converters investigated. 1970 The 1970s saw a new and very significant revolution hit the variable-speed drives market - packaging. Up until this time the static variable-speed drive design process had essentially concentrated on performance/functionality. Both A.C. and D.C. drives of even low rating were broadly speaking custom built/ hand crafted. This approach resulted in bulky, highcost drives the very uniqueness of which often compromised reliability and meant service support was difficult. The drives industry was not fulfilling its potential. 1970s A.C. motor drives had made great advances in terms of performance but still lacked the dynamic performance to really challenge the D.C. drive in demanding process applications. Since the early 1970s considerable interest was being generated in the field-oriented control of A.C. machines. This technique pioneered by Blaschke and further developed by Leonhard opened up the opportunity for A.C. drives not only to match the performance of a D.C. drive but to improve upon it. The processing requirements were such that in its early days commercial exploitation was restricted to large drives such as mill motor drives, boilerfeed pump drives. Siemens was very much in the forefront of commercialising field orientation and was also rationalising the numerous alternative drive topologies which had proliferated and, although stimulating to the academic, were confusing to drive users: D.C. drives single converter double converter - circulating current free circulating current

A.C. drives voltage (phase) control voltage-fed inverters quasi square V/f - quasi square V/f with D.C. link chopper - pulse-width modulated (PWM)

current-source inverters induction motor synchronous machine


static Kramer drive cycloconverter 1972 Siemens launched the SIMOPAC integrated motor with ratings up to 70 kW. This was a D.C. motor with integrated converter including line reactors! A new approach to drives in terms of packaging. Utilising 19-inch rack principles, a cubicle mounting standard well used in the process industry, compact, high-specification ranges of D.C. drives in modular form became available off the shelf. Companies such as AEG, Thorn Automation, Mawdsley's and Control Techniques pioneered this work. A new era of drive design had started.




Figure P.4 D.C. drive module (Control Techniques)




Further advances in packaging design were made possible by the introduction of isolated thyristor packages. In 1983 plastic mouldings made their first significant impact in drives. Bipolar transistor technology also arrived, which eliminated bulky auxiliary commutation circuits. Takahashi and Noguchi published a paper on direct torque control (DTC) in the IEEE. (This date is included not because of its technical significance rather as a point of interest as DTC has received much attention recently.)

Figure P.6 Digital D.C. drive with microprocessor and ASIC (Control Techniques)
1986 Great advances were being made at this time in the field of microprocessors making possible costeffective digital drives at low powers. Further drives were introduced containing application-specific integrated circuits (ASIC), which up to that time had only been used in exceptionally large volume/ domestic applications. Further, new plastic materials were introduced which gave structural strength, weight, size, assembly and cost advantage. IGBT technology was introduced to the drives market. IGBTs heralded the era of relatively quiet variable-speed drives (and introduced a few problems, some of which have led to substantial academic activity and a very few of which have required more pragmatic treatment). The first implementation of the field-orientation or flux-vector drive was introduced to the high-volume, lower power market. It found immediate application in machine tool spindle drives and has grown rapidly in application (and rating) since. It should be said that the name vector has been prostituted by some in the drives industry with voltage vector and other such names/techniques, causing confusion and frustration to customers. The trend to smaller drive products which were also simpler to design was given a significant boost by Mitsubishi which introduced intelligent power modules, integrating into the semiconductor package necessary gate drive and protection functions. A new packaging trend e m e r g e d - the bookform shape which had previously been applied to servo drives was now being applied to the broader





Figure P.5 Plastic mouldings introduced into drives (Control Techniques)



i! i i !i !~i!~i~'i!!i!i!i~i !!ii ~i ~i~~' !i iii~?'.....




~ !iI. . . . . ii

i,~I ~

. ~ ~

~ ~ .... i~ ~I~I~I

i i ~ ~i i ~ !!~i,~i~ ~iill iii~~:i~ii!~ i!~!!: ~ i~ ~ I~~I~ '


P.7 I G B T p o w e r Techniques)


in an A.C.


(Control F i g u r e P.9 B o o k f o r m shape of drive (Control Techniques)





i~,~ii i~,~,+~i>~,~ii~iiii~ii~i~!ii '~'

F i g u r e P.8 V e c t o r d r i v e ( C o n t r o l


industrial A.C. drives market. The trend continues today but there is not a consensus that this is the most suitable shape for all market segments. Another innovation in packaging - at the low-power end of the spectrum when a DIN rail mounting 0.4 kW inverter package, similar to that used widely in equipment such as contactors and control relays, was launched. The first drive with a built-in supply side filter fully compliant with, the then impending, EU regulations on conducted EMC was introduced. The first truly universal drive was launched which met the diverse requirements of a general purpose open-loop vector drive, a closed-loop flux-vector drive, a servo drive and a sinusoidal supply converter with the selection purely by parameter selection. The integrated D.C. motor launched in 1972 was not a great commercial success - much has been learned since those days. In 1998 integrated A.C. motor drives were introduced onto the market. These products are, for the most part, open-loop inverterdriven induction motors and were initially targeted on replacing mechanical variable-speed drives. A radical servo drive was introduced with the position and speed loop embedded in the encoder housing on the motor itself. This brought with it the advantage of processing the position information close to the source thereby avoiding problems of noise etc., and allowed dramatic improvements in control resolution, stiffness of the drive and reduced the number of wires between the drive and the motor.



Figure P. 12 Integrated A.C. motor (courtesy Leroy Somer)

Figure P. 10 DIN rail mounting drive with built-in EMC filter (Control Techniques)

A review of the time lines presented above illustrates that development within the drives industry continues at an ever increasing pace. Fundamental changes in the product, from a customer perspective, are still emerging, accessing ever more applications driven by automation and quality. This book covers the present state of development, or rather commercial exploitation, of industrial A.C. and D.C. variable-speed drives and associated systems. It is intended primarily for the use of professional engineers who specify or design systems which incorporate variable-speed drives. The theory of both the driven motor and the drive is explained in practical terms, with reference to fundamental theory being made only where necessary. Information on how to apply drive systems is included, as are examples of what is available within commercially offered drives and indications of what can be achieved using them. Emphasis is placed on industrial drives in the range 0.37 kW to 1 MW. The practical emphasis of the book has led to two unfortunate but I fear unavoidable consequences. First, some of the theory behind the technology contained in the book has had to be omitted or abridged in the interests of simplicity and volume. Second, in such a practical book it has proved difficult to avoid reference to proprietary equipment. In such circumstances a tendency towards referencing the products of Control Techniques is inevitable. It should be clear to readers that these products are described for illumination and explanation of the technology. The lEE, publisher of this book, does not endorse these products or their use in any way. This edition of the Control Techniques Drives and Controls Handbook has been created with contributions from engineers both within Control Techniques itself as well as sister companies within the family that is Emerson. I would in particular like to thank Dr Pete Barrass, Ray Brister,

Figure P. 11 Universal A.C. drive modules (Control Techniques)



Figure P. 13 Speed loop motor (Control Techniques)

Dr Mike Cade, Vikas Desai, Dr Colin Hargis, Jim Lynch, John Orrells, Bleddyn Powell, Alex Rothwell, Michael Turner and Peter Worland.

Prof Bill Drury


Preface I Industrial motors


1 D.C. motors General Fundamental equations and performance Wound-field motors Permanent-magnet motors Operating principles Commutation Rotation Compensation D.C. PM commutation Construction of the D.C. motor D.C. motor frame D.C. motor armature Brush gear Mountings D.C. PM design Rotor inertia Permanent-magnet materials 2 A.C. induction motors General Fundamental equations and performance Electrical characteristics of induction motors Torque characteristics Voltage-frequency relationship Increased voltage Reduced frequency Slip-ring induction motor Speed-changing motors A.C. induction motor construction 3 A.C. synchronous motors General A.C. synchronous motor construction 4 Brushless servomotors General Principles of operation of brushless servomotors Introduction Torque constant

1 1

2 2 3 4 4 5 5 6 6 6 6 7 7 7 8 8 8 8 9 9 10 11 12 12 13 13 14 15 15 17 17 17 18 18 18

Relationships between torque and back e.m.f. constant Stationary torque characteristics Construction of brushless servomotors Stator structure Rotor structure 5 Reluctance motors 6 A.C. commutator motors 7 Mechanical and environmental Mounting of the motor General IEC 60034-7 standard enclosures NEMA standard enclosures ~ Degree of protection General IEC 60034-5 US practice Cooling General Air filters Duty cycles Continuous duty - S 1 Short-time duty- $2 Intermittent d u t y - $3 Intermittent duty with starting- $4 Intermittent duty with starting and electric braking- $5 Continuous operation periodic d u t y - $6 Continuous operation periodic duty with electric braking- $7 Continuous operation periodic duty with related load speed changes- $8 Duty with nonperiodic load and speed variations- $9 Duty with discrete constant loads - S 10 Terminal markings and direction of rotation General IEC 60034-8 NEMA Ambient conditions Introduction

19 19 20 20 20 21 22 22 22 22 22 22 24 24 24 25 26 26 27 27 28 28 28 29 29 29 29 30 30 30 30 30 30 32 33 33



Temperature Altitude Power supply system Noise and vibration General Vibration Noise Motors for special applications Geared motors Brake motors Torque motors Motors for hazardous locations General CENELEC North American standards Testing authorities 8 Effects of semiconductor power converters General Drive converter effects upon D.C. machines Drive converter effects upon A.C. machines Introduction Machine rating - thermal effects Machine insulation Beating currents Overspeed Motors for hazardous locations

33 34 34 34 34 34 35 37 37 37 37 37 37 37 39 39 40 40 40 40 40 40 40 46 48 48

2 Power electronics
1 Power semiconductor devices General Diode rectifier Thyristor Thyristor gating requirements Power losses and current ratings Surge current ratings High-frequency current operation Gate turn-off thyristor Switching characteristics and gate drive Snubber design Voltage and current ratings Bipolar Transistor Voltage ratings Current ratings Switching characteristic and base drive Safe operating areas Short-circuit performance MOSFET Voltage and current ratings Switching performance Safe operating area Parasitic diode Insulated-gate bipolar transistor Operation Voltage and current ratings Switching behaviour and gate drive Safe operating area Short-circuit performance Series and parallel operation Integrated-gate commutated thyristor Voltage and current ratings Switching behaviour and gate drive Other power devices and materials

51 51 51 52 54 54 55 56 56 57 57 57 57 58 58 59 60 61 61 62 62 62 62 63 63 63 63 64 64 64 65 65 65 66

MOS-controlled thyristor MOS tum-off thyristor Silicon carbide Power device packaging Pressure contact packages Large wire-bonded packages for power modules Small wire-bonded packages for discrete devices Applications 2 Drive converter circuits A.C. to D.C. power conversion General Converters for connection to a single-phase supply Converters for connection to a three-phase supply Voltage tipple characteristics Practical effects D.C. motor drive systems D.C. to D.C. power conversion General Step-down D.C.-D.C. converters Step-up D.C.-D.C. converters A.C. to A.C. power converters with intermediate D.C. link General Voltage source inverters Current source inverters Direct A.C. to A.C. power converters General Soft starter/voltage regulator Cycloconverter Static Scherbius drive Matrix converter

66 66 67 67 67 67 69 70 72 72 72 73 74 76 76 76 79 79 79 81 81 81 81 83 85 85 85 86 86 87

3 Speed and position feedback devices

1 D.C. tachometer generator General Output voltage ripple Temperature effects Linearity and load effects Stability of the output Maximum terminal voltage Maximum operating speed Mechanical construction 2 A.C. tachometer generator 3 Resolver Design principles Synchros Torque synchros Control synchro Resolver General Computing resolvers Phase shifting Brushless resolvers Multipole resolvers A.C. rotary pickoffs Resolver-to-digital conversion 4 Encoder Incremental encoder Absolute encoder Sin/cos encoder

90 90 90 91 91 91 91 91 92 92 92 93 93 93 94 94 94 94 95 95 95 95 96 97 97 98 99



5 Selection of a feedback device for a drive system 6 Mechanical considerations 7 Glossary of terms 4 Drive control 1 General The ideal control system Open-loop control Closed-loop control Criteria for assessing the performance of a closed-loop control system 2 A.C. motor drive control General-purpose open-loop A.C. drive Space-vector modulator and inverter Reference-frame translation Reference-frame generation Current limit Performance and applications Permanent-magnet servodrive Reference-frame generation Current control Speed control Performance and applications Closed-loop induction motor drive Flux calculator and reference-frame generation Flux control Performance and applications Operation without position feedback Four-quadrant operation Reference-frame generation Performance and applications Direct torque control 3 D.C. motor drive control Flux controller Torque controller Performance and applications 4 Analysis of and set up of a speed controller Ideal speed controller Calculating the required gains Nonideal effects in a real speed controller

99 100 101 103 103 103 103 104 104 105 106 106 107 107 108 108 109 109 109 110 110 110 111 111 111 111 112 112 112 113 114 114 115 116 116 116 117 119

Polyphase switched-reluctance machines Losses in the switched-reluctance motor Excitation frequency Power electronics for the switchedreluctance motor Power supply and front-end bridge Power switching stage Single-switch-per-phase circuits Multiple-phase operation Single-switch circuit using bifilar winding Two-switch asymmetrical bridge Advantages of the switched-reluctance system Rotor construction Stator construction Electronics and system-level benefits Disadvantages of the switched-reluctance system Torque tipple Acoustic noise 3 Stepper-motor drives Stepping-motor principles The permanent-magnet motor The VR motor The hybrid motor Stepping-motor drive circuits (logic modes) Unipolar switching Bipolar switching High-speed stepping- L/R drives Chopper drives Bilevel drives Application notes Effect of inertia Resonance Special products Stepper/encoders Space-rated steppers Fuel-control actuators

127 127 128 128 128 128 128 129 129 130 131 131 131 131 132 133 133 134 134 135 135 135 136 137 137 139 139 139 139 139 139 140 140 140 140 141 141 143 143 143 143 143 143 143 143 144 144 144 144 145 145 145 148 148 148 148 148 148 149 149

6 Practical drives
1 General Digital input Programmability Typical specification Digital output Programmability Typical specification Analogue input Programmability Typical specification Analogue output Programmability Typical specification 2 D.C. drives The technology Drive selection Technical specifications and ratings Drive set up and commissioning Autotune PC-based commissioning tools- MentorSoft Performance Speed and current-loop response Typical applications Low-power analogue D.C. drives

5 Switched-reluctance and stepper-motor

drives 1 General 2 Switched-reluctance motors and controllers Basic principle of the switched-reluctance motor Operation as a motor Operation as a brake or generator To summarise so far Relationship between torque polarity and motoring/generating Control of the machine in practice Low-speed operation What happens as speed is increased? Medium-speed operation How is performance maintained as speed increases? High-speed operation Summary of typical/practical control Control of speed and position 121 121 122 122 123 123 123 124 124 124 124 124 125 125 126 126



The 4Q2 D.C. drive Cheetah-Puma-Lynx 3 A.C. drives Features common to all A.C. drives Power terminal layout Control terminal layout Wiring precautions to prevent electromagnetic compatibility (EMC) issues Open-loop inverters Specifications and ratings Features and options Methods of control Performance of the Commander SE open-loop drive Typical applications The universal A.C. drive The concept of a universal drive Unidrive option modules Open-loop operation Closed-loop operation Servo operation Regeneration mode High-performance servodrives Performance Summary of practical advantages of SLM technology Applications 4 Soft-start A.C. motor control Conventional starting Direct-on-line starting (DOL) Star-delta starting Auto-transformer starting Disadvantages of conventional starting Electronic soft start Typical applications 5 Application boards and software Applications module Software commissioning tools Communications modes Drive set-up wizard Commissioning screen Monitoring screen Parameter list

149 149 151 151 151 151 153 153 153 153 156 156 157 157 157 158 159 159 160 162 163 166 168 169 169 169 169 169 170 170 171 171 172 172 173 173 173 174 174 174 175 175 177 180 180 180 180 180 182 182 183 187 188 189 189 190 191

Summary 2 Network basics Physical layer Network cables and connectors Interface circuits Data encoding Network topology Data-link layer Framing Data model Media access control Error handling Conclusions Application layer Device profile 3 Simple fieldbus systems Modbus Control Techniques' protocol 4 Fieldbus systems Requirements for drive applications Physical layer Error detection Dynamic performance General message services Centralised v e r s u s distributed intelligence Profibus DP Interbus-S CAN DeviceNet CANopen CTNet

191 191 192 192 192 192 193 194 194 194 194 195 195 195 195 196 196 196 197 197 197 197 197 197 197 197 198 198 198 198 198 199 199 200 200 200 201 201 203 203 203 204 204 205 205 206 206 207 207 207 207 207 207 211 211 213 213 213 214

9 Supply harmonics due to drives

1 Overview 2 Regulations Regulations for installations Regulations and standards for equipment 3 Harmonic generation within variable-speed drives A.C. drives D.C. drives Effect of loading 4 The effects of harmonics 5 Calculation of harmonics Individual drives- D.C. Individual drives - A.C. Systems Isolated generators 6 Remedial techniques Connect the equipment to a point with a high fault level (low impedance) Use three-phase drives where possible Use additional inductance Additional A.C. supply-line inductance Additional D.C. inductance Use a higher pulse number (12 pulse or higher) Use a drive with an active input stage Use a harmonic filter

7 Position and motion-control systems

1 General 2 Basics of motion control 3 Typical motion functions Position lock- electronic gearbox Direct positional lock Ramped nonrigid lock Ramped rigid lock Simple single-axis positioning CAM functions Multiaxis positioning 4 Programmability 5 Summary

8 Communications systems
1 Introduction Drive set up Drive control

10 Electromagnetic compatibility (EMC)

1 Introduction General Principles of EMC



EMC regulations 2 Regulations and standards Regulations Standards 3 EMC behaviour of variable-speed drives Immunity Low-frequency emission High-frequency emission 4 Installation rules EMC risk assessment Basic rules Segregation Control of return paths, minimising loop areas Earthing Simple precautions and fixes Full precautions 5 Theoretical background Emission modes Principles of input filters Screened motor cables Ferrite ring suppressors Filter earth leakage current Filter magnetic saturation 6 Additional guidance on cable screening for sensitive circuits Cable screening action Cable screen connections Recommended cable arrangements

214 214 214 215 215 215 216 216 217 217 217 217 217 218 219 219 220 220 221 222 222 222 222 223 223 223 225 227 228 228 228 228 228

11 Systems design
1 General 2 Design matrix 3 Dynamic/resistive braking General D.C. motor braking Example calculation of a brake resistor of a D.C. motor A.C. regeneration and braking Example calculation of a brake resistor of a PWM A.C. induction motor drive system 4 Fusing General Protection of mains-drive and drive-motor cabling Protection of drive components 5 Motor overtemperature protection General Overtemperature protection of a converter-fed motor What can be used What cannot be used 6 A.C. drive motor cabling General Closed loop- induction motor Cable resistance Cable-charging currents Closed l o o p - P M servomotor Cable resistance Cable-charging currents Open-loop current control-induction motor Cable resistance Cable-charging currents

10 231 231 232 233 233 233 233 234 234 235 235 235 236 236 236 236 236 237 237 237 237 237 237


Limits to cable length Example Output chokes for long motor cable applications General Principles Calculations Example Position of chokes in multiple motor configurations Typical recommended cable size Power supply considerations High or low line voltage Supply frequency variations Supply impedance/fault level Low supply impedance High supply impedance Multiple drive installations Thermal design of enclosures General Calculating the size of a sealed enclosure Example Calculating the air flow in a ventilated enclosure Example Installation and maintenance of standard motors and electronic equipment Motors General Storage Installation Maintenance guide Brush gear maintenance Electronic equipment General Siting of equipment Ventilator systems and filters Condensation and humidity Fuses Common D.C. bus configuration of A.C. drives General A simple bulk uncontrolled rectifier Using the mains supply converter in one drive to supply all drives from its D.C. link Effectively hard paralleling of all drive input rectifiers A bulk four-quadrant controlled rectifier feeding the D.C. bus A bulk four-quadrant PWM converter feeding the D.C. bus Note on EMC filters for common D.C. bus systems Mechanical vibration, critical speed and torsional dynamics General Example Causes of shaft vibrations independent of variable-speed drives Subsynchronous vibrations Synchronous vibrations Super-synchronous vibrations Critical speeds Applications where torque ripple excites a resonance in the mechanical system High-performance closed-loop applications

237 240 240 240 240 240 241 241 242 243 243 243 243 243 243 244 244 244 244 244 245 245 246 246 246 246 246 247 247 248 248 248 249 249 249 250 250 250 252 252 252 253 254 255 255 255 255 256 256 256 256 256 257


Limits to dynamic performance System control-loop instability Measures for reducing vibration

257 257 257 259 259 260 260 260 261 261 261 261 262 262 262 262 263 263 263 263 264 264 264 264 264 264 265 265 265 266 266 266 266 267 268 268 269 271 272 272 272 273 274 274 275 275 276 276 276 277 277 278 278 278 278 278 281 284 284

12 Applications
1 Typical load characteristics and ratings Metals industries Plastics Rubber Chemical Materials handling Lift, hoist and crane Concrete pipe manufacture Fans and blowers Pumps Paper and tissue Printing Packaging Engineering industries Wire and cable Hydraulics Electric motors and alternators Textiles Foods, biscuit and confection 2 Techniques common to many applications Special D.C. loads Traction motor field control Battery charging Electrolytic processes Electric heating and temperature control Digital slaving General Drive slaving techniques Principle of digital speed/position following The digital speed/position controller Load sharing General D.C. thyristor converter-fed system A.C. inverter-fed systems High-frequency inverters General Frequency control of A.C. induction motors High-frequency purpose-designed motors High-frequency inverters High-frequency applications Centre winders General Speed or torque control Taper tension Constant torque and field weakening Power requirements for centredriven winders Inertia compensation Loss compensation Flux compensation Drive selection-limiting parameters Sectional drive systems General Theory of operation Using an IEC61131-3 programming tool to configure a sectional drive line Energy saving

General Centrifugal pumps Centrifugal fans and compressors 3 Application principles/examples Cranes and hoists General Planning an installation Slewing control Crane refurbishment for a Norwegian steel wire rope maker Elevators and lifts Lift system description Speed profile generation Load weighing devices Metals and metal forming Winding, crimping and precise cutting Roll feed line Wire and cable manufacture Four-quadrant D.C. drives for a bar mill Wire-drawing machine Paper manufacturing General Sectional drives Loads and load sharing Control and instrumentation Winder drives Brake generator power and energy Unwind brake generator control Coating machines Paper-slitting machine Paper board machine Building materials Brick-handling line Roofing-tile manufacturing plant Textiles Fabric-dyeing machine Quilting machine Plastics extrusion General Basic extruder components Overall extruder performance Energy considerations Motors and controls Food Control of hammermills in animal feed production HVAC Air conditioning for driver and vehicle licensing agency Air-handling units at Oxford Brookes University students' union Steel Main mill drives Auxiliary drives Chemical Enamel painting of fluorescent tubes Marine applications Cable laying Pipe laying Control of lock gates and sluices Printing Real-time registration and shaftless web tensioning control

284 285 288 288 289 289 289 289 290 291 291 292 292 292 292 295 296 296 296 297 297 297 298 299 300 301 302 302 303 303 304 304 305 307 307 308 308 308 310 310 310 311 312 312 313 313 313 314 314 314 315 315 316 316 318 319 320 320



Offset priming presses Stage scenery-film and theatre James Bond film stunts Controlling acoustics Exhibition focal point- the Control Techniques' orchestra Rock concert Millennium Dome aerial ballet

321 322 322 323 323 325 326

Appendix A Standards for drives

A1 IEC (intemational) standards Planned future IEC61800 standards A2 CENELEC (EC) standards A3 British standards A4 IEEE (USA) standards A5 UL (Underwriters' Laboratories, USA) standards A6 Other standards Electricity Association, UK EIA/TIA (previously RS) ANSI

330 332 332 334 335 335 335 335 335 335

Method Calculate speeds and gearing ratio Load, force and torque Power ratings for the motor and drive Inclined conveyor Hoist Data Velocity ratio (VR) Speed and acceleration of the hook Lifting force and torque to accelerate from rest to full speed Lifting force and torque to maintain full speed Required motor power rating Drive module power rating Screw-feed loads

346 346 346 347 348 348 348 348 348 348 349 349 349 349 351 351 351 352 352 353 353 353 354 354 354 354 354 354 355 355 355 355 355 355 355 355 355 356 356 356 356 356 356 356 357

Appendix C Tables
C1 Mechanical conversion tables Length Area Volume Mass Energy Inertia Torque Force Power C2 General conversion tables Length Area Volume Mass Force and weight Pressure and stress Velocity (linear) Velocity (angular) Torque Energy Power Moment of inertia Temperature Flow Torque Force Moment of inertia Linear acceleration C3 Power/torque/speed nomogram

Appendix B Symbols and formulae

B 1 SI units and symbols SI base units Decimal multiples and submultiples Derived units Geometrical units Time-related units Mechanical units B2 Electrical formulae Electrical quantities A.C. three-phase (assuming balanced symmetrical waveform) A.C. single phase Three-phase induction motors Loads (phase values) Impedance A.C. vector and impedance diagrams E.m.f., energy transfer Mean and r.m.s, values, waveform Principles Mean D.C. value R.m.s. value Other waveforms Form factor B3 Mechanical formulae Laws of motion Linear motion Rotational or angular motion Relationship between linear and angular motion The effect of gearing Friction and losses Fluid flow B4 Worked examples of typical mechanical loads Conveyor Data

336 337 337 337 337 337 337 338 338 338 338 338 338 338 338 339 340 340 340 341 341 341 342 342 342 343 343 344 344 345 346 346 346

Appendix D World industrial electricity supplies (< 1 kV) Appendix E Bibliography and further reading Index


363 365

Industrial M o t o r s

i!~i~i!i i 1 ~ i iiili iiiii iiiiiiiiii iiiii 2 3 4 5 6 7 ..!iii.. .... 8



History will recognise the vital role played by D.C. motors in the development of industrial power transmission systems. The D.C. machine was the first practical device to convert

electrical power into mechanical power, and v i c e v e r s a in its generator form. Inherently straightforward operating characteristics, flexible performance and high efficiency encouraged the widespread use of D.C. motors in many types of industrial drive application.

D.C. MOTORS:General

The later developments of the lower-cost A.C. cage motor and, more recently, of electronic variable-frequency control have displaced the D.C. motor to some extent, particularly in the lower kW range. Nevertheless, the advantages associated with the inherently stable and relatively simple to control D.C. machine are indisputable. In its most straightforward form, speed is approximately proportional to armature voltage, torque to armature current and there is a one-to-one relationship between starting torque and starting current. Modem D.C. motors under thyristor control and with sophisticated protection continue to provide very sound industrial variable-speed drive performance. High-performance test rigs and the higher-kW ratings of drives for the printing and paper industries, for high-speed passenger lifts, and drives subject to high transient loading in the metal and plastics industries; all are likely to continue employing the D.C. motor with thyristor control for some considerable time particularly in refurbishment programmes where a D.C. motor exists. The task facing the A.C. drive to completely oust its D.C. competitor is formidable. D.C. machine tool and servo drives based mainly on chopper technology continue to offer high performance at low price and at ratings up to approximately 5 kW (continuous) but here too A.C. technology is making significant inroads. The introduction and development of electronic variablespeed drives continues to stimulate intensive development of motors, both D.C. and A.C. The performance capabilities of both are being extended as a result, and the D.C. motor is likely to find assured specialised applications for the foreseeable future. The majority of standard D.C. motors, both wound field and permanent magnet, are now designed specifically to take advantage of rectified A.C. power supplies. Square, fully laminated frame construction allows minimal shaft centre height for a given power rating, and affords reduced magnetic losses, which in turn greatly improves commutating ability. Over the last few years the use of permanent-magnet motors, usually in the fractional to 3 kW range, has become commonplace in general-purpose drive applications. In this design permanent magnets bonded into the motor frame replace the conventional wound field. The magnets have a curved face to offer a constant air gap to the conventional armature.


v~ E~

l 1
Ea = klndp

Figure 1.1 Shunt-wound D.C. motor

field flux by: (1.1)

where n = speed of rotation, 0 = field flux and kl--motor constant. Also, the applied, or terminal armature voltage Va is given by:

Va = Ea -[- Ia "Ra


where Va - applied armature voltage, Ia - armature current and Ra = armature resistance. Multiplying each side of equation 1.2 by Ia gives:
Vala : Eala + 12Ra


total power supplied = power output + armature losses Interaction of the field flux and armature flux produces an armature torque. Thus: torque M = k2Ifla where k2=constant, armature current. If=main (1.4)

field current and Ia=

This confirms the straightforward and linear characteristic of the D.C. motor and consideration of these simple equations will show its controllability and inherent stability. The speed characteristic of a motor is generally represented by curves of speed against input current or torque and its shape can be derived from equations 1.1 and 1.2:

klnq5 =




The circuit of a shunt-wound D.C. motor, Figure 1.1, shows the armature, armature resistance (Ra) and field winding. The armature supply voltage Va is supplied typically from a controlled thyristor system and the field supply voltage Vf from a separate bridge rectifier. As the armature rotates an e.m.f. Ea is induced in the armature circuit and is called the back e.m.f, since it opposes the applied voltage Va and the flow of current produced by Va. This back e.m.f. Ea is related to armature speed and main

If the flux is held constant, which is achieved by simply holding the field current constant in a properly compensated motor, then:
n -- k3[Va - (IaRa)]

The circuits of shunt-wound and series D.C. motors are shown in Figures 1.1 and 1.2. With the shunt motor the field flux by armature current, and the value exceeds 5 per cent of Va, giving Figure 1.3, where speed remains wide range of load torque. 0 is only slightly affected of IaRa at full load rarely a speed curve shown in sensibly constant over a

The series motor curve, Figure 1.4, shows initial flux increase in proportion to current, falling away owing to

Chapter 1.1


RaI <





Figure 1.2 Schematic of series D.C. motor Figure 1.5 Compound D.C. motor




(3. 1}










, | , , , , , , ,

torque (%)
Figure 1.3 Torque~speed characteristic of a shunt-wound D. C motor










torque (%)
Figure 1.6 Torque~speed characteristic of a compound D.C. motor

Under semiconductor converter control, with speed feedback from a tachogenerator, the shape of the speed/load curve is largely determined within the controller. It has become standard to use a plain shunt D.C. motor on the basis of reduced cost, even though the speed/load curve on openloop control is often slightly rising.










torque (%)
Figure 1.4 Torque~speed characteristic of a series D.C. motor

Permanent-magnet Motors
Compared with standard types of wound-field D.C. motor, the conventional D.C. permanent-magnet (D.C. PM) servomotor is often designed to exhibit extremely good low-speed torque tipple characteristics. However, the high-speed characteristics of a D.C. PM motor are not ideal for all applications. Because of the mechanical and electrical constraints set by the motor commutator, increasing the motor speed with a maintained load characteristic soon reveals the commutation limit of the motor. In practice, each motor is designed to work within a safe area of commutation where the available motor torque reduces as motor speed increases, in the manner of the shunt-wound D.C. motor, Figure 1.3. The torque/speed curve for a conventional D.C. PM motor is shown in Figure 1.7. The area of continuous operation may be defined as the area where the motor can operate on a 24-hour, 100 per cent duty basis with an acceptable temperature rise. The intermittent duty part of the motor characteristic defines the area of operation available on an intermittent basis for acceleration and deceleration periods. This aspect of rating is discussed in the section on Duty Cycles (pp. 27-8). The

magnetic saturation. In addition, the armature circuit includes the resistance of the field winding and the speed becomes roughly inversely proportional to the current. If the load falls to a low value the speed increases dramatically, which may be hazardous: the series motor should not normally be used where there is a possibility of load loss, but because it produces high values of torque at low speed and its characteristic is falling speed with load increase, it is useful in applications such as traction and hoisting or some mixing duties where initial stiction is dominant. The compound-wound D.C. machine combines both shunt and series characteristics; its circuit is shown in Figure 1.5. The exact shape of the torque/speed characteristic is determined by the resistance values of the shunt and series fields. The slightly drooping characteristic, Figure 1.6, has the advantage in many applications of reducing the mechanical effects of shock loading.

D.C. MOTORS: F u n d a m e n t a l Equations and Performance

commutation '~" limit intermittent zone



b ----..~ . . . . ~ a, I torque


Figure 1.7 Characteristic of D.C. PM motor a peak torque b stall torque at 40C

commutation limit is most obviously exceeded if brush sparking becomes excessive. Severe overloading in this range will cause a complete ring of heavy sparking to run around the commutator circumference. This phenomenon is known as brush fire or flashover and must be avoided since it damages both the commutator and brush gear, reducing the life expectancy of the motor considerably. Fortunately, the electronic controller supplying the motor may be easily specified to prevent overloading. Good commutation - the ability to give good armature current reversal without undue sparking at the brushes - is of particular importance. It should be noted that although the presence of sparking does give a very good indication of poor commutation, optimisation of commutation is a very skilled activity. Indifferent commutating performance places D.C. motors at some disadvantage when compared with A.C. cage induction motors, particularly in respect of maintenance requirements and costs. The increasing emphasis being placed upon reduced production down time and improved drive system dependability, combined with a reduced number of hours available for regular maintenance, are factors tending to favour A.C.


Figure 1.8 Interaction of stator and rotor fields producing torque

a stator field b rotor field c resultant field distortion

With separately-excited shunt-wound D.C. motors, a steady D.C. voltage is applied to the field winding and the resulting current produces a magnetic flux in the motor yoke (main frame), the main field poles, the armature and the air gaps, Figure 1.8a. Field strength is determined by the field-winding current. The armature winding occupies the rotor peripheral slots and opposite coils are connected to each other and to a commutator segment. The brushes provide the means of connecting current flow to the rotating armature. This current flow generates an armature flux, Figure 1.8b, with a magnetic axis mechanically fixed at 90 to that of the field flux. The interaction of these two magnetic fluxes generates a flux distortion, Figure 1.8c, so that the armature rotates, endeavouring continuously to correct this distortion and thus providing motive power. Permanent-magnet D.C.

motors act on the same principle except that the field strength is not variable.

In order that the motive torque shall be smooth and continuous, the current distribution in the armature coils must remain constant. This requires that the current in any given coil must reverse as that coil passes from the influence of one pole to the next in rotation. The process of current reversal is known as commutation and must take place during the time that the coil is short circuited by a brush spanning the gap between adjacent commutator segments. Failure to achieve armature current reversal during this short-circuit time will cause sparking at the trailing edge of the brush as the short circuit is broken. Commutator action is shown in detail in Figures 1.9 and 1.10.

Chapter 1.1



//2 Itor

i .........................................

i .............................




' ,




i ', i



', i





i, \i
. .


brush arcing
. . . . . .

contact/~:~ "........... ;eriodoi ........... ~" with short circuit period of segment c _ current reversal

'i i i i



Figure 1.10 Commutator current reversal

Figure 1.9 Reversal o f current flow in one armature coil

segment c makes contact with the brush (Figure 1.9). As the area in contact with segment c increases, the current flowing from the brush via segment c to coil A decreases. If the current through segments b and c is proportional to brush contact areas then the current through coil A decreases linearly to zero (direction left to right) and then builds up linearly in the opposite direction, Figure 1.10. Under these conditions commutation would be linear and sparkless, assisted by the changing ohmic resistance of the carbonbrush/copper segment contact area. Because of the armature coil inductance, a reactive current is induced in the coil undergoing commutation, delaying the current reversal in the short-circuited coil, as shown in Figure 1.10. With heavier motor loading there is an increase of armature current and armature flux, distorting further the flux generated by the field. This distortion, which is producing the rotational torque, acts to degrade commutation and to produce brush sparking. The degree of flux distortion from the neutral flux axis is proportional to armature current. The effect is termed armature reaction. Referring back to Figure 1.8, the brushes are shown on the neutral axis, i.e. at 90 to the flux path and midway between the field poles, assuming no flux path distortion. Figure 1.8c shows that distortion of the flux owing to armature reaction causes the neutral axis to shift. If the brush axis were moved to position X-Y, Figure 1.8c, the brushes would again lie on the neutral axis, that is at 90 to the direction of the main flux path.

There is a physical limit to the speed at which current can be commutated. This commutation limit is often defined as the product of the motor shaft power and the rotational speed. The limit is widely accepted to be approximately 3 x 106 (kW. m i n - 1). Where greater ratings are required more than one armature can be placed on a single shaft, or several motors arranged in series (provided the shafts are rated to transmit the total torque).

Figure 1.9a shows that coil A is about to be short circuited and is carrying half the armature current (/2) from left to right, where the current I flows to the brush. Figure 1.9b shows coil A in the middle of its short-circuit period, where armature current from the brush does not pass through shortcircuited coil A. In Figure 1.9c, coil A is shown immediately following the short circuit, carrying 1/2 current in the right to left direction; the current in coil A has been reversed, or commutated. Any failure of the current to fall through zero and rise to the same value in the reverse direction during this very brief period of commutation results in sparking at the brush as the inductive circuit of the armature coil is switched. The current changes are shown in graphical form in Figure 1.10. The current in coil A is 1/2 from left to right until

To improve commutation, interpoles are introduced to shift the main flux so that it lies on the geometric neutral axis. These interpoles are interposed between the main field poles and their coils carry the full armature current, generating a flux proportional to current (torque load) and reducing the effect of armature reaction. It follows that interpoles are most effective in their corrective action when the field flux is at its full level. The commonly used technique of field weakening progressively

D.C. MOTORS:Operating Principles

reduces this self-corrective effect; hence the quality of commutation can be affected when a motor operates under the weak field conditions sometimes required in system application. An additional method of reducing this troublesome armature reaction, is to neutralise the armature flux directly under the main pole faces by the use of compensating or pole face windings. These windings are placed in slots cut into the main pole air gap faces and carry the full armature current, being connected in series with the armature circuit, and thus giving compensation with current loading. Compensating windings on the pole faces present problems of both mechanical security and electrical insulation, a satisfactory solution often involving significant additional cost. The use of a compensated machine is restricted therefore to duties requiring a very wide range of field operation, such as highspeed machine tool applications and others involving rapidly varying loads and speeds, in which the commutating performance of the compensated motor can be extremely good when compared with that of the more normal noncompensated machine. incorporated into the design, the lower the resulting torque ripple will be. However, in practice, this is limited by physical practicalities, cost and available space. Good commutation in a D.C. PM motor is therefore a compromise.


The principal components of the D.C. motor comprise the rotating shaft-mounted wound armature and connected commutator, the field system surrounding the armature across the motor air gap, the supporting and enclosing end frames with shaft bearings and brush gear to connect the commutator to external terminals. These components and others are to be seen in the sectioned assembly of a typical machine, Figure 1.11.

D.C. Motor Frame

The main flame of a typical modern four-pole D.C. motor is a square, laminated iron flame with a curved inner face. Main pole lamination and coil shunt field assembly are mounted on the inside face at 90 to each other, with interpole assemblies interposed at 45 between. Square-frame motors offer reduced centre height compared with round-flame motors by virtue of the fact that the main field poles mounted on the four flat internal faces are of pancake design and that the four interpoles, mounted on diagonal axes, fit into the internal corners of the square.

D.C. PM Commutation
The commutator also plays a very important role in the functioning of a D.C. PM motor. To reduce torque ripple at the motor shaft it is desirable to keep the magnetic interaction between the rotor and stator as evenly distributed as possible. In this case, however, interpoles and winding compensation are not available options. The prime purpose of the commutator is to switch the current to each set of windings in turn on the armature, and if the optimum relationship between the stator flux and the armature flux is to be maintained for the best current conversion to torque, higher rates of commutation are essential to ensure minimum torque ripple in the motor. This means that the higher the number of commutation segments

D.C. Motor Armature

The armature consists of a cylindrical laminated iron core, shaft mounted, with a series of slots to accommodate the armature winding coils, insulated from each other and the core. The coil ends are connected to insulated segments of the commutator.

Figure 1.11 High-performance D.C. motor (courtesy Leroy-somer)

Chapter 1.1

The commutator, a cylindrical copper segmented construction, is mounted at one end of the armature core, with wedge-section segment bars clamped between commutator end tings, insulated from each other and the metal of the armature. These segments are machined on the outer axial face to provide a smooth concentric surface on which the brushes can bear with a tensioned contact. An armature assembly is shown in Figure 1.12. The armature winding slots are skewed by about one slot pitch to reduce torque ripple (cogging effect) at low speed and slot noise generation at high speed. Such a cogging effect superimposed on the otherwise steady rotation of the motor shaft would be unacceptable in many applications. For example, a machine tool axis drive with such a characteristic could well leave undesirable surface patterning on the workpiece. The precise method of skewing presents the motor manufacturer with two problems. The first is in determining the amount of skew for a particular design of motor. The second is that it is much more difficult to wind the armature on an automatic winding machine when the laminations are skewed in this fashion. Computer-aided design plays a large part in solving the first of these difficulties, and most motor manufacturers have developed standard programmes to speed magnetic calculations. Modem wire insertion machines make relatively easy work of shooting windings into the skewed armature slots. Armature winding coil ends are banded with glass-fibre and epoxy bonding for high speed operation (> 6000 m i n - ~ for many machine tool applications) to prevent armature coil deflection under centrifugal stress. Two balancing discs, one behind each beating, permit dynamic balancing to R or the more stringent S standard over the working speed range of the machine.

Brush Gear
The number of brushes is based upon the full-load current of the armature circuit and the working velocity of the commutator track and the working conditions generally govern the grade of carbon used. Brush pressure is usually fixed and unvarying over the permitted wear length, and is controlled by tensator constant-tension springs. It is very important to use the correct grade of brush!

The typical D.C. motor shown in Figure 1.11 is a drip-proof (IP23) enclosed foot-mounted D.C. motor with side-mounted terminal box and with a feedback tachogenerator coupled to the nondrive end. Access to the brush gear is through the four inspection covers around the commutator end shield, although in practice only three can be used. Any of these three inspection apertures can be used for mounting a force ventilation fan, to allow wide (typically 100: 1) speed range operation at constant load torque. Basic adaptability of the square-frame machine also allows the terminal box to be mounted on any of the three available faces of the machine, and for the machine itself to be mounted by end or foot flange; and since both end shields have the mounting feet attached, they can readily be rotated through 90 to allow side foot mounting.

D.C. PM Design
Industrial D.C. permanent-magnet (D.C. PM) motors tend to offer a smaller frame size for a given kW rating than their wound-field counterparts, although frame lengths can be greater. A D.C. PM servomotor is conventional in most aspects of construction, but usually incorporates technical refinements

Figure 1.12 D.C. motor armature assembly

D.C. MOTORS:Construction of the D.C. Motor

to give performance characteristics of a higher order than those of the general purpose D.C. PM motor. Areas where the motor designer achieves the required sophistication include rotor inertia and choice of magnetic materials.

determined it is possible to partially demagnetise the motor through overcurrent in the armature or with temperatures in excess of the Curie point of the material, with the result of reduced torque and usually a higher-than-rated motor speed on light load, for a given input voltage. Alternative types of magnetic material overcome these problems with very high demagnetisation levels and operating temperature ranges. Such a material is samarium cobalt, which is grouped with the rare-earth category of magnetic materials. There is much development activity in this area and new, improved, materials are appearing continually. Currently, magnets made from rare-earth materials are expensive and their use can double the cost of a motor, but for many applications the choice of such magnetic materials is unarguably justified; the resulting magnetic flux is much denser per unit volume of magnet, and therefore the rotor inertia can be significantly reduced for a given rated output torque. This allows an improved acceleration/deceleration performance per unit armature current. A magnetic material with the cost advantages of an ordinary ferrite magnet and the performance advantages of the rareearth types has been sought. This has now been achieved with the advent of synthetic magnetic materials, principally neodymium iron boron, which is fully acceptable to numerous industries, including the automotive industry, and provides extremely good performance characteristics without the large price premium of the rare-earth materials. The magnet itself is formed from powders drawn from an ever growing range of magnetic materials and is moulded into the required size, shape and profile using a high-temperature specialised sintering process. Delivered to the motor manufacturer in an unmagnetised state, the magnet is first bonded into the motor frame using a two-part adhesive with a suitably high melting point. The assembly is then magnetised by inserting a closefitting metal conductor through the motor frame and introducing a high-current shock pulse through the assembly by discharging a capacitor bank through the conductor. The current pulse polarises the elements of the sintered magnets and saturates them to one end of their magnetisation characteristic.

Rotor Inertia
One of the key elements of a D.C. PM motor is the rotor or armature design. It is a normal requirement of a servo system to be able to accelerate and decelerate the motor (and load) very rapidly. To facilitate rapid changes in speed, and in order to reduce the input power required to achieve these changes, the inertia of the rotor is kept to a minimum. Since inertia increases by the square of the diameter, D.C. PM servomotor rotors tend to be long in relation to their diameter. Care should be taken not to apply very-low-inertia servomotors to high-inertia loads as the inertia mismatch can lead to very troublesome resonances. A prime reason for choosing a servo system as opposed to a conventional variable-speed D.C. system is the ability to operate the motor at or very near to zero speed, while still offering full-rated torque.

Permanent-magnet Materials
Choice of magnetic material for the D.C. PM motor is a major deciding factor in the operational characteristic. Conventional D.C. PM motors use a low-cost ferrite magnetic material which, although more than adequate for the majority of applications, does have some inherent disadvantages. Mechanically, ferrite magnets are fragile, and so are usually bonded to the motor body, which is very secure. However, because ferrite magnets are also brittle, they are susceptible to mechanical shock and can be fractured unless the motor is handled with great care. One of the more important disadvantages of low-cost magnets, however, is the ease with which they can become demagnetised. Ferrite magnets have a relatively low demagnetising level. Under normal circumstances, and with a correctly-determined servo system, demagnetisation of the motor is not possible. However, if the system is not correctly


The A.C. squirrel-cage induction motor is the basic, universal work horse of industry, converting some 70 to 80 per cent

of all electrical power into mechanical energy. This type of motor nonetheless exhibits some quite unattractive performance characteristics in spite of intensive development, notably instability and a nonlinear load-current

Chapter 1.2

characteristic. It is invariably designed for fixed-speed operation, larger ratings having such features as deep rotor bars to limit direct-on-line (DOL) starting currents. Electronic variable-speed converter technology is able to provide the necessary variable-voltage/current, variable-frequency supply which the three-phase A.C. machine requires for efficient, dynamic and stable variable-speed control. Modern electronic control technology is able not only to render the A.C. induction motor satisfactory for many modern drive applications but also to greatly extend its application and enable advantage to be taken of its low capital and maintenance costs. More striking still, microelectronic developments have made possible the highly dynamic operation of induction motors by the application of flux-vector control. The practical effect is that it is now possible to drive an A.C. induction motor in such a way as to obtain a dynamic performance in all respects better than could be obtained with a phase-controlled D.C. drive combination. The various forms of A.C. induction motor control are fully described in Chapter 4.

connections for the winding coils are not shown, but R and R1 represent the start and finish of the red phase winding, and similarly for the Y and B phase conductors. The R, Y and B phase windings are displaced 120 in space relative to one another. Assuming that when stator current is positive it is flowing inwards in conductors R, Y and B, and therefore outwards in R1, Y1 and B1, that the current in phase R in Figure 1.13a is at its maximum positive value and that in phase windings Y and B the currents at the same instant are negative and each equal to half maximum value, then these currents produce the magnetic fluxes represented by the arrowed lines in Figure 1.13a and the flux axis is horizontal. Thirty degrees later in the supply cycle, the currents in phases R and B are 0.866 (x/~/2) of their maximum, and zero in the Y phase. The pattern of the flux due to this current is shown in Figure 1.13b. It will be noted that the axis of this field is now in line with coil Y-Y1 and therefore has turned clockwise through 30 from that of Figure 1.13a. After a further 30 in the supply cycle the current in phase winding B has reached maximum negative value, and the currents in R and Y are both positive, at half their maximum value. These currents produce the magnetic flux shown in Figure 1.13c, the flux axis being displaced clockwise by a further 30 compared with that of Figure 1.13b. Thus, for every time interval corresponding to 30 in the supply cycle, the axis of the flux in a two-pole A.C. stator rotates 30 in space. With a two-pole stator (one pair of poles) the flux rotates through one revolution in space in one cycle of the power supply. The magnetic flux is said to rotate

FUNDAMENTAL EQUATIONS AND PERFORMANCE Electrical Characteristics of Induction Motors

Consider the stator winding of a simple three-phase two-pole A.C. cage induction motor, each phase winding having only one slot per pole per phase as shown in Figure 1.13. End



Figure 1.13 Three-phase rotating field


A.C. INDUCTION MOTORS: Fundamental Equations and Performance

at synchronous speed, and the rotational speed of the flux is:


revolutions in one second

where f = supply frequency in Hz, and p = number of pole pairs. Note: a two-pole motor has one pole pair! It is more usual to express speed in revolutions per minute:

windage, the rotor could not continue to rotate at synchronous speed. The rotor speed must therefore fall below the synchronous speed, and as it does so rotor e.m.f, and current, and therefore torque, will increase until the speed matches that required by the losses and by any load on the motor shaft. The difference in rotor speed relative to that of the rotating stator flux is known as the slip. In Figure 1.15, for a given torque OA, rotor speed is AC and the slip is AD, and: AD = A B - AC = CB It is usual to express slip as a percentage of the synchronous speed, i.e. AD/AB x 100. Slip is closely proportional to torque from zero to full load. Slip at full load varies from about 7 per cent for small motors to about 2 per cent for large, and is a good indication of the efficiency of a machine; the lower the slip the higher the efficiency. The fact that the A.C. squirrel-cage induction motor is neither a true constant-speed machine nor inherently capable of providing variable-speed operation constitutes a major limitation. Analysis shows that the rotor flux ampere turns travel at the same speed as those of the resultant of the stator flux ampere turns and that both are stationary relative to each other. The three-phase induction motor can be regarded as a transformer having an air gap between the magnetic circuits of the primary (stator) and secondary (rotor) windings. This air gap, although designed to be as small as practicable, has an important effect on the characteristics of A.C. squirrel-cage motors. The air gap requires larger magnetisation and gives larger magnetic leakage than a transformer of the same kVA rating. However, stator and rotor ampere turns have to balance, as well as supplying magnetisation and no-load losses, just as in transformer design. So an increase in percentage slip due to load is accompanied by an increase in rotor current (secondary current) and therefore by a corresponding increase in stator current (primary current) above the no-load current. This no-load current for a typical squirrel-cage motor lies between 25 and 40 per cent of the full-load current and is largely due to the magnetic excitation current required by the air gap. Thus the cage induction machine over its working load range is self regulating for input power, this regulation being controlled by the percentage slip with load. The most popular squirrel-cage induction motor in sizes up to about 5 kW is of four-pole design. Its synchronous speed with a 50 Hz supply is therefore:

60f/p revolutions per

minute (min -1)

The e.m.f, generated in a rotor conductor by transformer action is at a maximum in the region of maximum flux density. With the flux rotation assumed to be clockwise, the direction of the e.m.f.s generated in stationary rotor conductors (determined by Fleming's fight-hand rule) are indicated by conventional crosses and dots in Figure 1.14a. The e.m.f, generated in the single rotor conductor shown in Figure 1.14b produces a current, the effect of which is to reinforce the flux density on the left-hand side of the conductor and to weaken it on the right-hand side. In consequence, a force is exerted on the rotor tending to move it in the direction of the flux rotation. The higher the speed of the rotor, the lower the speed of the rotating stator flux field relative to the rotor winding, and therefore the smaller the e.m.f, generated in the rotor winding. If the speed of the rotor became the same as that of the rotating field, i.e. synchronous speed, the rotor conductors would be stationary in relation to the rotating flux. This would produce no e.m.f, and no rotor current and therefore no torque on the rotor. Because of friction and

60f/p =

60.50/2 = 1500 min -1

Slip accounts for about 5 per cent, and a typical nameplate speed is 1425 min -1.

Torque Characteristics
Figure 1.14 Torque on the rotor

a currents induced in rotor bars producing field b distortion of field

A disadvantage of the squirrel-cage machine is its fixed rotor characteristic. The starting torque is directly related to the rotor circuit impedance, as is the percentage slip when running at load and speed.

C h a p t e r 1.2


synchronous speed


rotor speed
i i i i i i i

full-load torque

Figure 1.16 Typical rotor bar profiles


rotor slip 600 ent a 500 ~400



Figure 1.15 Speed curve of an induction motor

. current b

Ideally, a relatively high rotor impedance is required for good starting performance (torque against current) and low rotor impedance for low full-load speed slip (and high efficiency). This problem can be overcome to a useful extent, for directon-line (DOL) application, by designing the rotor bars with special cross sections, Figure 1.16, rather than round or square, so that eddy-current losses, which these special sections cause, increase the impedance at starting when the rotor flux (slip) frequency is high, and reduce it at normal running speed when the flux frequency is low. Altematively, for special high-starting-torque motors, two or even three concentric sets of rotor bars are used. Relatively costly in construction but capable of a substantial improvement in starting performance, this form of design produces an increase in full-load slip. Since machine losses are closely proportional to working speed slip, increased losses may require such a high starting torque machine to be derated. The curves in Figure 1.17 show indicative squirrel-cage motor characteristics. In the general case, the higher the starting torque the greater the full-load slip. This is one of the important parameters of squirrel-cage design as it is tied to working efficiency and therefore working losses.

o 300 200



~o speed, %



Figure 1.17 Typical torque~speed and current~speed curves

a standard m o t o r b h i g h - t o r q u e m o t o r (6 per cent slip)

the voltage impressed upon the winding by the supply. This is because the resistance of the winding results in only a small voltage drop, even at full-load current, and therefore in the steady state the supply voltage must be balanced by the e.m.f, induced by the rotating field. This e.m.f, depends on the product of three factors: 1 2 3 The total flux per pole (which is usually determined by the machine designer). The total number of turns per phase of the stator winding. The rate of field rotation or frequency.



Nearly all commercially-available industrial induction motors are wound for direct connection and starting on the supply voltage and frequency which prevail in the country where they will be used. It is a relatively simple matter for the motor manufacturer to select the number of tums per coil and the wire size to match any voltage within a wide range. If it is desired to convert a constant-speed motor operating direct-on-line to a variable-speed drive using an inverter it is necessary to consider the effect of frequency on flux and torque. An induction motor on a normal supply operates with a rotating field set up by three-phase currents in the stator winding. The magnitude of the field is controlled broadly by

Exactly the same factors are valid for transformer design, except that the field is pulsating instead of rotating. In a transformer it can be shown that the e.m.f, induced in a winding is given by:

where V is the e.m.f, induced in a coil, K is a constant, ~; is the flux through the coil andfis the frequency of the supply. This can be rearranged to give:

For constant flux, it can be seen that the ratio held constant.

V/f must be


A.C. INDUCTION MOTORS" Voltage-Frequency Relationship

For inverter operation the speed of field rotation for which maximum voltage is appropriate is known as the base speed.
100 %

Increased Voltage
If, again in the steady state, the voltage applied to the stator terminals is increased without a corresponding increase in the frequency, only the flux can vary to regain the balance between applied voltage and e.m.f. If the flux is forced to increase by applying excessive voltages, the iron core of the machine is driven progressively into saturation. This not only increases iron losses due to hysteresis and eddy currents, but can lead to a very marked increase in stator current, with corresponding resistive losses. Since most machines are designed to work with the minimum of material, their magnetic circuits are normally very close to saturation and excessive stator voltage is a condition which must be carefully avoided.
C~ 0 > 0 0

100 %



Reduced Frequency
The consequence of reducing the supply frequency can readily be deduced from the relationship described above. For the same flux the induced e.m.f, in the stator winding will be proportional to frequency, hence the voltage supplied to the machine windings must be correspondingly reduced in order to avoid heavy saturation of the core. This is valid for changes in frequency over a wide range. The voltagefrequency relationship should therefore be linear if a constant flux is to be maintained within the machine, as the designer intended. If flux is constant so is the motor torque for a given stator current, hence the drive has a constant torque characteristic. Although constant voltage-frequency (V/f) control is an important underlying principle, it is appropriate to point out departures from it which are essential if a wide speed range is to be covered. First, operation above base speed is easily achieved by increasing the output frequency of the inverter above the normal mains frequency; two or three times base speed is easily obtained. The output voltage of an inverter cannot usually be made higher than its input voltage and therefore the V/f characteristic is typically as shown in Figure 1.18a. Since V is constant above base speed, the flux will fall as the frequency is increased after the output voltage limit is reached. The machine flux falls (Figure 1.18b) in direct proportion to the actual V/f ratio. Although this greatly reduces the core losses, the ability of the machine to produce torque is impaired and less mechanical load is needed to draw full-load current from the inverter. The drive is said to have a constant power characteristic above base speed. Many applications not requiring full torque at high speeds can make use of this extended speed range. The second operating condition where departure from a constant V/fis beneficial is at very low speeds, whereby the voltage drop arising from the stator resistance becomes significantly large. This voltage drop is at the expense of flux, as shown in Figure 1.18b. To maintain a truly constant flux within the machine the terminal voltage must be increased above the constant V/f value to compensate for

wi,_ 0 0

100 %


base speed
m 0 >

J .e'"

. ~t ' " " "




frequency and speed

Figure 1.18 Voltage-frequency characteristics

a linear VIf below base speed b typical motor flux with linear V/f (showing fall in flux at low frequency as well as above base speed) c modified V/f characteristic with lowfrequency boost (to compensate for stator resistance effects in steady state)

the stator resistance effect. Indeed, as output frequency approaches zero, the optimum voltage becomes the voltage equal to the stator IR drop. Compensation for stator resistance is normally referred to as voltage boost and almost all inverters offer some form of adjustment so that the degree of voltage boost can be matched to the actual winding resistance. It is normal for the boost to be gradually tapered to zero as the frequency progresses towards base

Chapter 1.2


speed. Figure 1.18c shows a typical scheme for tapered boost. It is important to appreciate that the level of voltage boost should increase if a high starting torque is required, since the IR drop will be greater by virtue of the increased stator current. In this case automatic load-dependent boost control is useful in obtaining the desired lowspeed characteristics. Such a strategy is referred to as constant V/f (or V/Hz) control and is a feature of most commercially available A.C. drives, although more advanced open-loop strategies are now becoming available (refer to Chapter 4). So far the techniques described have been based on achieving constant flux within the air gap of the machine or, if that is not possible, then the maximum flux. Constant flux is the ideal condition if the largest capability of torque is required because the load cannot be predicted with certainty, or if the most rapid possible acceleration time is desired. A large number of inverters are used, however, for variable air volume applications where control of airflow is obtained by variable-speed fans. The torque required by a fan follows a square-law characteristic, Figure 1.19, with respect to speed, and reducing the speed of a fan by 50 per cent will reduce its torque requirement to only 25 per cent of its rated torque. As the load is entirely predictable there is no need for full torque capability and hence flux to be maintained, and higher motor efficiency can be obtained by operating at a reduced flux level. A further benefit is that acoustic noise, a major concern in air-conditioning equipment, is significantly reduced. It is therefore common for inverters to have an alternative square-law V/f characteristic or, ideally, a self-optimising economy feature so that rapid acceleration to meet a new speed demand is followed by settling to a highly efficient operating point.

With the correct value of (usually) resistance inserted in the rotor circuit, a near-unity relationship between torque and supply current at starting can be achieved, i.e. 100 per cent full-load torque, with 100 per cent full-load current, 200 per cent FLT with 200 per cent FLC etc., (i.e. comparable with the starting capability of the D.C. machine). Not only the high starting efficiency but also its smooth controlled acceleration historically gave the slip-ring motor great popularity for lift, hoist and crane applications. It has had similar popularity with fan engineers, to provide a limited range of air volume control, either 2:1 or 3 : 1 reduction, at constant load, by the use of speed regulating variable resistance in the rotor circuit. Although a fan possesses a square-law torquespeed characteristic, so that motor currents fall considerably with speed, losses in the rotor regulator at lower motor speeds are still relatively high, severely limiting the useful speed range. Rotor slip-ring systems, used with this type of motor, offer a similar service life to that of the D.C. motor commutator system. Efficient variable-speed control of slip-ring motors can be achieved by converters using the slip energy recovery principle first proposed by Kramer in 1904. Such schemes are based upon converting the slip frequency on the rotor to supply frequency. (These schemes are described in Chapter 2.2.) It is also possible to retrofit variable-frequency inverters to existing slip-ring motors. This can be done simply by shorting out the slip-ring terminations (ideally on the rotor thereby eliminating the brushes) and treating the motor as a cage machine. Variable voltage control of slip-ring motors has been used extensively, notably in crane and lift applications, although it is now largely being replaced by flux-vector drives and will therefore not be considered further.


The wound rotor or slip-ring A.C. machine, although introducing the negative aspect of brushes, does address some of the disadvantages of the cage induction motor but with the handicap of cost compared to the equivalent rated D.C. shunt-wound machine.

The cage rotor already described in outline is constructed with copper rotor bars brazed to shorting end tings, or is aluminium die cast in a single operation. The construction is simple, cheap and robust. It has the further advantage that with various stator winding pole combinations, whereby the sequence of current reversals in the rotor bars is altered, the rotor end tings provide a free path for the current to flow, adapting to the differing number of stator poles. This allows stator windings of more than one pole combination to be wound on the same stator, and such multispeed A.C. squirrel-cage motors can take one of two forms. The simplest is to combine two quite separate stator windings in the one machine, for example four and eight poles, providing a 2:1 speed range with a constant torque relationship between the two speeds. This would suit many applications in materials handling and in lift and crane drives, but would be wasteful of torque and therefore of cost for fan and pump drives, where the load torque requirement for low-speed operation falls as the square of the speed reduction.

100 %


~o s

"10 0


base speed "'~'

frequency and speed

Figure 1.19 Typical square-law characteristic of a fan or pump load





Alternatively, and often more commercially attractive, there is the design of a single stator winding which allows different pole combinations by switching external connections. This is termed consequence pole switching and is restricted to 2:1 speed range combinations, i.e. 2/4 pole or 4/8 pole, few other 2:1 combinations being practicable. Since this uses a single winding only, a constant torque relationship between the two speeds is available. More than two speeds can be made available from A.C. squirrel-cage motors by combining separately-wound and consequence pole switching in the same machine. Pole amplitude modulated (PAM) speed change, as this method is known, is also available for machines of larger kW ratings and is somewhat similar in principle to consequence pole switching.

A die-cast aluminium rotor cage induction motor construction is generally used up to about 50 kW and allows best economics in manufacture. Rotor losses with aluminium bars are bigger than those with copper and copper tends to be substituted at 50 kW or thereabouts. The use of aluminium in the frame construction of smaller machines up to 37 kW rating is becoming standard practice on account of the better thermal conductivity when compared with cast iron. It offers benefits in the areas of resistance to corrosion, ease of machining, reduced weight and generally improved appearance. The better heat conduction of aluminium and the much improved force-ventilated cooling of the standard IP54 enclosed machine of current design show significant benefits, although such cooling still falls short of the effectiveness of screen-protected IP22 enclosures, which exchange the cooling air around the stator and rotor windings directly with the surrounding ambient air. Above 37 kW, aluminium construction appears to lose its advantages, particularly in terms of the mechanical strength required for the higher torque loading involved. Cast iron and steel construction becomes the standard as does detailed attention to effective cooling of the IP54 enclosed machine. At these higher kW ratings the inferior efficiency of the A.C. cage machine compared with the equivalent rating of D.C. motor and brushless synchronous A.C. motor gives the larger cage machine additional kW losses to dissipate. However, a major advantage of the A.C. cage machine is that it is readily available in enclosures to match the requirements of difficult and hazardous environments and can be specified to accept much higher levels of external vibration and shock loading than is possible for a comparable D.C. machine. Standard cage induction machines are available to comply with a wide variety of international standards.


The A.C. cage induction motor stator coils are wound on a laminated iron core formed into a pack usually by seam welding. Semiclosed slots accommodate the windings, which are normally wound as a concentric three-phase, typically four-pole, arrangement suitably insulated per coil set, per phase and from earth. A laminated cylindrical rotor core is mounted centrally in the stator with the smallest practical radial air gap and with semiclosed or closed rotor slots around the periphery to accommodate the rotor winding. This winding comprises bare or lightly-insulated aluminium or copper bars, occupying the slots and connected to shorting end rings at each end face of the rotor pack. A typical die-cast aluminium rotor construction is shown (with stator) in Figure 1.20, for comparison with a copper bar cage rotor, Figure 1.21. For lower ratings, the rotor laminated core and cage construction is replaced by a cylinder of mild steel, which works quite well but with greatly reduced output power relative to size. i~i!i: ~!~: i!i: !~!: .... ? ::~ ~ ::%::: ~:i:i~.... : : i . . .


~:~~ ii~!i~i:i~

ili i i!i il i ii! i

! !iil i !~i


::~ i: !i(~~:iiiii)/!i:i:iiii~ii~!:iii!~ iii! : i ~ !i i~ ! i : .... !i:~i!:iiil:i:i~i Iil~I!~:'Ii:iii~ii~iiiii~i~ i:i:iiii ii ~ i ~i: i ~ !i i !

.... !

Figure 1.20 Typical squirrel-cage induction motor die-cast rotor with its laminated stator (courtesy Leroy-somer)

Chapter 1.3


Figure 1.21 Typical squirrel-cage induction motor rotor with inserted copper rotor bars (courtesy Leroy-Somer)


Considering a rotor of the solid pattern it is possible to visualise a series of flats machined around the rotor block periphery, the number of flats corresponding to the number of poles with which the three-phase stator winding is wound, and as shown in Figure 1.22 for a four-pole (1500 m i n - 1 50 Hz) design. A motor fitted with such a rotor would run at synchronous speed with no load on the shaft and, for a typical low kW design, would pull out of, or lose, synchronism at a torque loading in excess of about 10 per cent of that of the equivalent frame size asynchronous design. Having pulled out of the synchronous condition the percentage slip would be high, the motor essentially reduced to a crawl if the load is not removed, causing significant losses and excessive temperature rise. The performance of such a simple synchronous cage machine would be poor but could be greatly improved by designing

permanent magnets into the rotor system, Figure 1.23, or by borrowing A.C. alternator technology and making the rotor of wound multipole design, either with a separate D.C. supply for excitation or of brushless alternator design with shaft-mounted exciter. A.C. synchronous motor application engineering needs particular care to ensure that the machine will accelerate its load to synchronous speed and be capable of holding synchronous speed under transient load conditions. Such a motor can prove commercially attractive, particularly where precise speed holding or speed matching between drives is essential and where, by borrowing alternator technology, advantage can be taken of volume manufacture and the availability of an IP23 protection enclosure, which allows improved cooling. Small salient-pole synchronous motors possess limitations for general and for variable-speed drive use. For this reason, further comment refers to the larger salient-pole and to cylindrical-rotor synchronous motors, which are likely to



frequency related to the flux speed difference, the mean torque being zero. Such conditions can occur due to load transients and during starting. Wound-rotor synchronous motors have a stator design similar to that of a standard A.C. cage induction motor with a three-phase balanced winding of either two or four-pole configuration. It is the rotor design, which can take alternative forms, that affects the performance of a given size of synchronous machine. Cylindrical-rotor synchronous motors have not only a uniform air gap but also a rotor flux distributed sinusoidally in space. This combines with the three-phase stator flux, which in the air gap produces a sinusoidal pattern of mutual flux, resulting in a sinusoidal torque pattern. Such a rotor has a suitably-distributed three-phase winding excited by D.C., usually one rotor phase carrying the full D.C. current value, the other two phases each carrying half current, in a series/parallel connection. An advantage of the synchronous motor is that it does not require reactive current to magnetise the air gap so there is little restriction on the radial air gap when compared with the induction cage machine, where air gap is of great importance. A common alternative to the cylindrical rotor is the salientpole machine. Generally, a three-phase four-pole winding is employed, except on some small machines where a two-pole winding is more convenient. Salient-pole rotors provide a large space for the windings but generally give a flux wave pattern rich in lower-order harmonics, although these can be somewhat reduced by correct design of the pole tips. The general symmetry per pole, with differing reluctances between stator and rotor, results in a tendency for stator and rotor poles to align. This reluctance torque allows the rotor to accelerate and run synchronously without rotor excitation at small to moderate loads. Normal excitation supplements the torque produced by reluctance and allows the machine to hold synchronism with transient load changes. To improve starting, rotor bars are often built into the salient-pole faces with similar design considerations as to the profiles of the slots and bars as in a standard cage machine. With the salient-pole machine running at synchronous speed, for reasons already considered, little current will circulate in the rotor bars. That which does tends to reduce harmonic currents. In the same way as the design and arrangement in modern A.C. generators provide rotor excitation without the use of either slip rings or commutators, such self-excited designs have now become general for A.C. synchronous motors in view of the poor maintenance record of slip-ring systems. Self excitation is additionally attractive, since rotor D.C. excitation involves relatively little power, perhaps only up to about 2 per cent of the machine rating, derived from a common shaft-mounted A.C. exciter with rectifiers. The excitation is capable of being regulated externally to the machine. This is necessary where advantage is to be taken of the power factor control capability of this type of machine, and where inverter-based variable-speed control is required. The so-called V curves for a typical synchronous motor are shown in Figure 1.24. The operating power factor is

Figure 1.22 Synchronous motor rotor with flats machined on a solid core

permanent magnets

Figure 1.23 Synchronous motor rotor with permanent magnets

have wide application in A.C. variable-speed drives above about 40kW. Permanent-magnet motors are considered here, but only their more common application in the specialised field of servomotors. Switched reluctance motors are considered entirely separately in Chapter 5.2. A.C. asynchronous induction motors produce shaft torque, which is proportional to percentage slip, implying that with zero slip the machine produces zero torque. Synchronous torque can be produced at speed:

where ns =synchronous rotational speed in min-1, f = frequency in Hz and p = number of field pole pairs. This is achieved by a field winding, for convenience of design generally wound on the rotor and D.C. excited so that it produces a rotor flux which is stationary relative to the rotor. Torque is produced when the rotating three-phase field and the rotor field are stationary relative to each other, hence there must be physical rotation of the rotor at speed ns in order that its field travels in step with the stator field axis. At any other speed a rotor pole flux would approach alternately a stator north-pole flux, then a south-pole flux, changing the resulting torque from a positive to a negative value at a

Chapter 1.4


teD L_ 0 C~ c.m


full load

increases the excitation field voltage proportionately, to compensate for the effects of flux neutralisation. The A.C. synchronous motor appears to have some attractive features for inverter variable-speed drive applications, particularly at ratings of 40 kW and above. Not least is overall cost when compared with an A.C. cage motor plus inverter, or D.C. shunt-wound motor and converter alternatives. In applications requiring a synchronous speed relationship between multiple drives or precise speed control of single large drives the A.C. synchronous motor plus inverter control system appears attractive: freedom from brush gear maintenance, good working efficiency and power factor are the main considerations.

full load no load


E lag .~ power"factor -,! I

,~ lead

field excitation current

Figure 1.24 V curves of a typical synchronous motor

determined by the relationship between load current and excitation current. The dotted line indicates the minimum stator current at the various loads, that is with excitation adjusted to give unity power factor. For steady-state operation, adjustment of the excitation current will change the working power factor. With variable loading, and where variable-speed inverter control is involved and operating power factor is important, conventional power-factor correction techniques are employed. Motor excitation level is set to give optimum power factor for the working conditions. In common with D.C. motors, increased line current partially neutralises the excitation field m.m.f. Since shaft torque is approximately related to the product of line and field current fluxes, increased line current can give a reducing torque increase. A typical inverter for variable-speed control automatically regulates the main stator voltage to be in proportion to motor frequency. It is possible to arrange an excitation control loop, which monitors the main stator voltage and


With the wound-rotor A.C. synchronous motor the stator remains generally the same as for the A.C. cage motor, with the rotor slots now accommodating a fully insulated wire or bar-wound rotor winding of the same pole number as the associated stator with connections brought out to external terminals via slip rings and brushes. Synchronous A.C. motors above the low kW rating sizes are similar in construction to A.C. alternators. Again, the stator winding is similar to that of the A.C. cage motor, the rotor carrying the field excitation winding, typically of brushless design, wound either upon a cylindrical rotor (similar to the wound rotor arrangement mentioned above) or upon salient poles. It is interesting to note that slip-ring induction motors can also be made to operate synchronously by supplying the rotor with D.C. current through the slip rings.


In Chapter 1.2 we saw that a three-phase stator winding of an induction motor produces a sinusoidal rotating magnetic field in the air gap. The speed of rotation of the magnetic field is directly proportional to the supply frequency. In Chapter 1.3 we saw that the same is true for a synchronous motor. In this case the rotor has a D.C. excited winding or permanent magnets designed to lock on or synchronise with the rotating magnetic field. The synchronous machine with permanent magnets on the rotor is the heart of the modem brushless servomotor.

The synchronous motor stays in synchronism with the supply, although there is a limit to the maximum torque which can be developed before the rotor is forced out of synchronism. Pull-out torque will be typically between one and a half and four times the continuously-rated torque. The torque speed curve is therefore simply a vertical line, which indicates that if we try to force the machine above the synchronous speed it will become a generator. The industrial application of brushless servomotors has grown significantly for several reasons: reduction of price of power conversion products; establishment of advanced




o" i-



b S W-


1.25 Steady-state torque speed curve for a synchronous motor supplied at constant frequency

control of PWM inverters; development of new, more powerful and easier to use permanent-magnet materials; development of highly accurate position controllers; the manufacture of all these components in a very compact form. They are, in principle, easy to control because the torque is generated in proportion to the current. In addition, they have high efficiency and high dynamic responses can be achieved. Brushless servomotors are often called brushless D.C. servomotors because their structure is different from that of ordinary D.C. servomotors. Brushless servomotors switch current by means of transistor switching within the associated drive/amplifier, instead of a commutator as used in D.C. servomotors. In order to confuse, brushless servomotors are also called A.C. servomotors because brushless servomotors of synchronous type with a permanent magnet rotor detect the position of the rotational magnetic field to control the three-phase current of the armature. It is now widely recognised that brushless A.C. refers to a motor with a sinusoidal stator winding distribution which is designed for use on a sinusoidal or PWM inverter supply voltage. Brushless D.C. refers to a motor with a trapezoidal stator winding distribution, which is designed for use on a squarewave or block commutation inverter supply voltage.

! i

Figure 1.26 Principle of a rotating field

at point A, and phases V and W are both negative. Therefore, the direction of the current of each coil is as shown in Figure 1.26a and the composite vector of the magnetic flux induced by the current is generated in the direction from N towards S. If there is a rotor field intersecting the magnetic flux at right angles at that time, torque is generated to rotate the rotor clockwise owing to the repulsive and attractive forces between the magnets. At point B, magnetic flux is generated 60 degrees further clockwise. It follows from the above that a continuously rotating field can be obtained by making three-phase currents flow in the stator coil. If the sine wave phase and rotational position can be made to be always at right angles, it becomes possible to make a highly efficient motor producing smooth torque without using brushes.


The brushless servomotor lacks the commutator of the D.C. motor, and has a device (the drive sometimes referred to as the amplifier) for making the current flow according to the rotor position. In the D.C. motor, torque variation is reduced by increasing the number of commutator segments. In the brushless motor, torque variation is reduced by making the coil three phase and, in the steady state, by controlling the current of each phase into a sine wave. Figures 1.26a and b are cross-sectional views of a threephase synchronous motor, with U +, U - , V +, V - , W + and W - indicating the beginning or the end of the coil of each phase. When a motor is energised by three-phase alternating currents as shown in Figure 1.26c, only phase U is positive

Torque Constant
In the armature of the motor of Figure 1.27, the current distribution is as illustrated. If the current flowing in the conductors to the right of the symmetrical axis OO t is in the direction of (away from the reader), then current in the conductors to the left flows in the opposite direction of (towards the reader). Assume that all the conductors in the fight-hand half are under the north pole and all the conductors in the left-hand half are under the south pole of the permanent magnet and the magnetic flux density has an average value of B to simplify the discussion. Then the torque RBIL should work on every conductor and the whole

Chapter 1.4


force (back e.m.f.), and the direction of this force is opposite to the terminal voltage applied. This value is directly pro, portional to the rotational speed 9t (in radians per second) and is given by the following equation:

The proportional constant Ke in this equation is the back e.m.f, constant. Note: the back e.m.f, constant is usually expressed in units of V/kr.p.m. (where the voltage is the r.m.s, voltage and kmin-1 is in thousands of revolutions per minute). The back e.m.f, constant Ke can be expressed in terms of other parameters. If the rotor is revolving at a speed of f~ radians per second, the speed of the conductor v is:

S I~

!~1 N

v = f~R Therefore, the back electromotive force e generated in a conductor is:

Figure 1.27 Field flux and current distribution in a rotor

e = ~ RBL
If the total number of conductors is Z, then the number of conductors in series connection is Z/2 and the total back e.m.f. E at the motor terminals is:

torque T around the axis will be:

T = ZRBLI = (ZRBLIa) / 2

E = ~ RBLZ/2
By substitution we can express E in terms of the flux q) as:

where Z is the total number of conductors, R is the radius of the rotor, B is the flux density linking the stator windings, L is the inductance of the winding and Ia is the current from the motor terminal, which is equal to 2L In this model, the magnetic flux is equal to: -- 7rRLB Therefore, by substitution, we get:

E = (q,Z/ZTr)f~ We can therefore obtain for the back e.m.f, constant Ke:
Ke = (Z/ZTr)ff

T = (Z/Tr)~Ia/2
Now let us consider this equation. The number of conductors Z never changes in a finished motor. Because the magnetic flux ~I, is determined by the motor dimensions and state of magnetisation, (Z/Tr)~b is a constant. Therefore, we can conclude that the torque T is proportional to the armature current Ia. We can therefore define the torque constant Kt as"

It should be noted that Kt and Ke are only equal when a selfconsistent unit system is used. The international system of units (SI) is one such system. For example, if Kt is equal to 0.05NmA -1, then Ke is equal to 0.05 Vsrad -1. As stated earlier, although it is normal to express the torque constant in terms of NmA-1 it is more usual to express the back e.m.f. constant in terms of V/kmin-1 (volts per thousand revolutions per minute).

Stationary Torque Characteristics

A motor which uses permanent magnets to supply the field flux is represented by the simple equivalent circuit of Figure 1.28. This is a series circuit of the armature resistance, Ra, and back e.m.f., E. If we ignore the voltage drop across the transistors, the equation for the voltage is"
V -- gala + ge[~

KtTherefore, we obtain


T = Ktla
The torque constant is usually expressed in units of NmA(where the current is the r.m.s, current). It should be noted that the above torque equation is identical to that of a D.C. motor with constant field (equation 1.4).

The armature current Ia is:

Ia-- (V-ge~)/Ra

Therefore, from above, the torque T is:

Relationships between Torque and Back E.M.F. Constant

As each conductor passes through the north and south magnetic poles, the electromotive force changes successively. The total electromotive force on each coil merges to the motor terminals. This voltage is the back electromotive

T - Kda
-- ( K t / R a ) ( V - Ke[~)

Figure 1.29 shows the relationship between T (torque) and f~ (rotational speed) at two different voltages. The torque decreases linearly as the speed increases. The slope of this function is a constant KtKe/Ra and is independent of the


BRUSHLESSSERVOMOTORS:Principles of Operation of Brushless Servomotors

terminal voltage and the speed. Such characteristics make the speed or position control of a D.C. motor relatively easy. The starting torque and the no-load speed (assuming no beating friction and windage loss) are given by:
r, = K, z/1o ~o = V/Ke

the stator of a D.C. motor. In other words, the magnetic field for generating torque is stationary in D.C. motors, but rotates in brushless servomotors; conversely, the armature revolves in D.C. motors and is stationary as a stator in brushless servomotors. Servomotors require rapid acceleration and deceleration, and the maximum torque therefore has to be several times larger than the rated torque. As brushless servomotors, unlike D.C. motors, do not have a commutation limit, they can be operated up to the boundary of high-speed rotation without decreasing the maximum torque. Further, in a brushless servomotor, the primary area of heat dissipation occurs not on the rotating part but on the armature in the stator, since the permanent magnets are mounted on the axis of rotation. The heat dissipated in the stator diffuses into the air through the flame. It is therefore relatively easy to cool brushless servomotors. Moreover, brushless servomotors provide more precise overload protection, because the temperature of the hottest part can be detected directly.


Figure 1.30 shows the structure of a typical brushless servomotor. The windings are in the static armature, which is part of the stator, and therefore the rotor of a brushless servomotor can be considered to be the equivalent of

R~ 0 i | [

Stator Structure

Figure 1.28 Simplified equivalent circuit

E= KeD

slope= Kt Ke


A typical stator consists of an armature core and armature windings. The armature core is made of laminating punched silicon steel sheet of 0.35 to 0.5 mm thickness (laminated core). In many cases, the armature core is slotted and skewed to reduce torque tipple, which results in speed tipple. The armature windings are similar to that in an A.C. motor and are usually of the distributed three-phase type. The windings are usually designed according to the drive specification, which requires either a sinusoidal or trapezoidal back e.m.f. waveform. The factors governing the design are stator slot number, pole shape, windings coil pitch, rotor pole number and magnet shape.


Rotor Structure
The structure of a typical rotor is characterised by permanent magnets fixed on the axis of rotation. The shapes of the permanent magnets vary according to design and can be classified into two principal types: cylindrical and salient pole.


Figure 1.29 Torque speed characteristic

Fixing of the permanent magnets to the rotor is critical in the design of brushless servomotors. Various techniques have been applied to the adhesion of magnets in order to

Figure 1.30 Structure of a brushless servomotor

C h a p t e r 1.5


prevent the destruction of motors from centrifugal force in high-speed rotation or that caused by repetitive rapid acceleration and deceleration. Common methods used to prevent separation of permanent magnet from the rotor surface are: binding the outer surface of the permanent magnet with glass-fibre tape or yarn; using a thin stainlesssteel cylinder as a sleeve to cover the outer surface of the permanent magnet. Adhesives are used in combination with either of these methods and are chosen with a linear expansion coefficient which is comparatively close to that of the permanent magnet and that of the axis of rotation. At

the same time they need to be stable against any thermal change. Recent development in rare-earth magnets has also contributed greatly to improve brushless servomotor performance. Rare-earth magnets have nearly the same residual magnetic flux density as that of an Alnico magnet and two to three times higher coercive force than that of a ferrite magnet. These features greatly help in making modem permanent magnet brushless servomotors of light weight and high performance.


The reluctance motor is arguably the simplest synchronous motor of all, the rotor consisting of a set of iron laminations shaped so that it tends to align itself with the field produced by the stator. The stator winding is identical to that of a three-phase induction motor. However, the rotor is different in that it contains saliency (a preferred path for the flux); this is the feature which tends to align the rotor with the rotating magnetic field, making it a synchronous machine. The practical need to start the motor means that a form of starting cage also needs to be incorporated into the rotor design, and the motor is started as an induction motor, the reluctance torque then pulling in the rotor to run

synchronously in much the same way as for a permanent magnet rotor. Reluctance motors may be used on both fixed-frequency (mains) supplies and inverter supplies. These motors tend to be one frame size larger than a similarly rated induction motor and have low power factor (perhaps as low as 0.4) and poor pull in performance. As a result of these limitations their industrial use has not been widespread except for some special applications such as textile machines where large numbers of reluctance motors may be connected to a single bulk inverter and maintain synchronism. Even in this application, as the cost of inverters has reduced, bulk inverters are infrequently used and the reluctance motor is now rarely seen.

saturable bridge Q-axis teeth fhJX barri~



filled with aluminium

structurally reinforced for mechanical strength

Figure 1.31 Rotor punching for a four-pole reluctance motor




Single-phase A.C. commutator motors of small kW ratings are manufactured in large quantities, particularly for domestic applications and power tools. Development generally has been in the direction of improved commutation and reduced cost. The series A.C. machine is basically similar to the series D.C. motor and remains the more important. Although not capable of commutation to the standards required for industrial drives, the type can provide approximate speed control simply by crude voltage regulation often facilitated by a single semiconductor switch. An alternative design of some interest is the single-phase A.C. commutator repulsion motor in which the armature brushes are shorted together, with the A.C. supply taken to the field windings only. Limited additional speed control can be achieved by angular brush shift - an inconvenient method, particularly in machines of low rating and small size. Starting torque is high for the repulsion motor and its design survives nowadays in the form of the repulsion-start, induction-run single-phase machine, where a centrifugal device shorts out the commutator segments when the motor is up to speed. More important is the A.C. three-phase commutator motor, which over past years has provided outstanding service where variable speed is required.

Again, although many ingenious designs have been produced in the search for improved variable-speed performance, the most successful for ratings between 5 kW and 150 kW has been the Schrage or rotor-fed machine, followed by the stator-fed or induction regulator motor, which has been manufactured in ratings in excess of 2000 kW. Both designs combine commutator frequency changing and variable-speed motor in the same frame, generating the required slip frequency voltage in a primary winding, for injection into a secondary winding. The kW rating of the Schrage motor is restricted by the slip rings, carrying as they do the total power of the motor. The stator-fed or induction regulator motor is a somewhat similar arrangement which converts the mains frequency injected voltage to slip frequency voltage for the rotor coils. A step-down variable ratio transformer is generally connected between the line supply and the commutator connection to give voltage/phase control. The parallel transformer connection gives shunt characteristics and the series transformer connection, series characteristics. The stator power feed of this design allows, as has been mentioned, ratings over 2000 kW to be designed successfully although commutators now appear to be an increasing disadvantage.



Internationally agreed coding applies to a range of standard mountings for electric motors, D.C. and A.C. which covers all the commonly required commercial arrangements. Incorrect mounting of a motor can cause premature failure and loss of production. All motor manufacturers will provide advice on the suitability of a particular build for a specific application. Within IEC 60034-7 (EN 60034-7) there are examples of all practical methods of mounting motors. NEMA publishes alternative standards within NEMA standards publication no. MG 1 - Motors and generators.

IEC 60034-7 standard enclosures

Electrical machines have been categorised within this standard by the prefix IM (international mounting), a letter and one or two subsequent digits. It is unusual for the prefix IM to be used and it is more usual to see only a letter followed by one or two digits e.g. B 3 - foot mounting. The most usual types of construction for small and mediumsized motors are shown in Table 1.1.

NEMA standard enclosures

Motor mounting and location of the terminal box location is designated in accordance with the arrangements shown in Table 1.2.

Chapter 1.7
Table 1.1 Usual types of construction for small and medium-sized motors foot mounted two bearing plates


/. .

a , l



I iI
V5 M...


11 !
| | ,I i a I I

B8 flange mounted (with through holes on the beating plate) two beating plates


! [


{ Ii

B5 attachment to front end of casting (as B5, V1 or V3 with beating plate removed) only one bearing plate B9 two beating plates flange with holes on the machine casing




V8 V9

B10 two beating plates flange with threaded holes on the beating plate



I I 1



B14 two beating plates feet plus flange with through holes on bearing plate t.....,--....,.2.a B3/B5 two beating plates feet plus flange with threaded holes on bearing plate




MECHANICAL AND ENVIRONMENTAL: M o u n t i n g o f t h e m o t o r

Table 1.2 NEMA standard enclosures

floor mountings


F-2 ~('~

wall mountings






k 1

( '"~



W-7 ceiling mountings

/ m

N m




All types of electric motor are classified in accordance with the provisions which specify a standard coding to indicate the degree of protection, afforded by any design, against mechanical contact and against various degrees of ambient contamination. The designation used as defined in IEC 60034-5, (EN 60034-5) consists of the letters IP (international protection) followed by two numerals signifying conformance with specific conditions. Additional information may be included by a supplementary letter following the second numeral. Interestingly, this system is contained within NEMA MG1 but is not widely adopted by the industry in the United States.

IEC 60034-5
It has already been stated that the designation used as defined in IEC 60034-5 consists of the letters IP followed by two numerals signifying conformance with specific conditions.
IP characteristic letters 1st characteristic numeral 2nd characteristic numeral 4 4

Example of designation

The first characteristic numeral indicates the degree of protection provided by the enclosure with respect to persons and also to the parts of the machine inside the enclosure.

Chapter 1.7
Table 1.3 First characteristic numeral
First characteristic numeral Brief description nonprotected machine machine protected against solid objects greater than 50 mm diameter Definition


no special protection no accidental or inadvertent contact with or approach to live and moving parts inside the enclosure by a large surface of the human body, such as a hand (but no protection against deliberate access); no ingress of solid objects exceeding 50 mm in diameter machine protected against solid no contact by fingers or similar objects not exceeding 80 mm in objects greater than 12 mm in diameter length with or approaching live or moving parts inside the enclosure; no ingress of solid objects exceeding 12 mm in diameter machine protected against solid objects no contact with or approaching live or moving parts inside the greater than 2.5 mm in diameter enclosure by tools or wires exceeding 2.5 mm in diameter; no ingress of solid objects exceeding 2.5 mm in diameter machine protected against solid objects no contact with or approaching live or moving parts inside the greater than 1 mm in diameter enclosure by wires or strips of thickness greater than 1 mm in diameter dust-protected machine no contact with or approaching live or moving parts within the machine; ingress of dust is not totally prevented but dust does not enter in sufficient quantity to interfere with the satisfactory operation of the machine dust-tight machine no contact with or approach to live or moving parts inside the enclosure; no ingress of dust

Table 1.4 Second characteristic numeral

Second characteristic numeral Brief description nonprotected machine machine protected against dripping water machine protected against dripping water when tilted up to 15 from the vertical machine protected against spraying water machine protected against splashing water machine protected against water jets machine protected against heavy seas Definition no special protection dripping water (vertically falling drops) shall have no harmful effects vertically dripping water shall have no harmful effect when the machine is tilted at any angle up to 15 from the vertical, from its normal position water falling as a spray at an angle up to 60 from the vertical shall have no harmful effect water splashing against the machine from any direction shall have no harmful effect water projected by a nozzle against the machine from any direction shall have no harmful effect water from heavy seas or water projected in powerful jets shall not enter the machine in harmful quantities

machine protected against the effects of immersion in water to depths of between 0.15 m and 1 m protected against the effects of prolonged immersion at depth

The second characteristic numeral indicates the degree of protection provided by the enclosure with respect to harmful effects due to ingress o f water.

The most frequently used degrees o f protection for electrical machines are as given in Table 1.5. Brushless servomotors are normally IP65 - this single feature is often the key reason for a user to select brushless servomotors for specific applications such as in the food industry where w a s h d o w n is a c o m m o n requirement.

For open internally air-cooled machines suitable for use under specific weather conditions and provided with additional protective features or processes, the letter W m a y be used. This additional letter is placed immediately after the IP e.g. IPW 54. Similarly, the letter R is used to indicate duct-ventilated machines (in such cases the air discharge must be located outside the r o o m where the motor is installed).

US Practice
It is c o m m o n practice for manufacturers o f electrical machines in the United States to adopt less formal designations, as given in Table 1.6.



Table 1.5 Frequently used degrees of protection for electrical machines

First numeral Second numeral 0 IPO0
IPl l

IP02 IP12


7 IP17

8 IP18



IP23 IP44 IP54

IP55 IP65


Table 1.6 Protection categories in the US

open drip proof (ODP) a machine in which the ventilating openings are so constructed that successful operation is not interfered with when drops of liquid or solid particles strike or enter the enclosure at any angle from 0 to 15 downward from the vertical; these are motors with ventilating openings, which permit passage of external cooling air over and around the windings the TEFC-type enclosure prevents free air exchange but still breathes air; a fan is attached to the shaft, which pushes air over the flame during operation to help the cooling process the TEAO enclosure does not utilise a fan for cooling, but is used in situations where air is being blown over the motor shell for cooling such as in a fan application; in such cases the external air characteristics must be specified a TENV-type enclosure does not have a fan an enclosure designed for use in the food processing industry and other applications that are routinely exposed to washdown, chemicals, humidity and other severe environments

totally enclosed fan cooled (TEFC)

totally enclosed air over (TEAO)

totally enclosed nonventilated (TENV) washdown duty (W)

Very much related to the enclosure of the machine, but not synonymous with it, is the method of cooling. All rotating electrical machines designed for economy of materials and dimensions require an effective form of cooling to ensure that internal losses are dissipated within the limits of the maximum temperature rise for the class of winding insulation employed, and so that bearing and surface temperature rise figures are kept within safe limits. The different historical periods of development between the D.C. and the A.C. motor resulted in the two having different enclosures and cooling arrangements, the D.C. motor as standard being drip proof (IP23) and force ventilated, the A.C. cage machine being totally enclosed (IP44), fan cooled (TEFC), with a shaft-mounted fan at the nondrive end running within a cowl to duct the cooling air over a finned motor body. The result is that the standard D.C. motor is readily force ventilated internally through its winding space, Figure 1.32, to allow a wide speed range, whereas the standard A.C. cage motor is not. It should be noted that for variable speed A.C. applications, the shaft-driven internal or external fans of the standard IP23 or IP54 A.C. cage machine can produce a cooling problem, as their cooling performance varies inversely as the square of the shaft speed. At half speed they provide only 25 per cent of the full-speed cooling effect. On a typical

constant-torque load requiring constant motor current, as discussed under A.C. motor characteristics, the motor has approximately constant losses over its speed range. The motor will require significant derating of its output, or cooling with constant-velocity cooling air provided by a fan driven at constant speed independently of the motor shaft speed. D.C. motors and synchronous motors are generally of IP23 enclosure and therefore suitable for constant-velocity forced ventilation cooling, either from a frame-mounted fan as already mentioned or, where the working environmental conditions are so difficult as to require an IP54 enclosure, from a remote fan mounted in a clean-air position, with ventilating ductwork between motor and cooling fan unit. The D.C. IP23 machine is restricted to use in a clean-air environment or to one in which an air filter on the vent fan inlet gives sufficient protection to the winding and commutator. Where the working environment is so difficult as to require a totally enclosed machine, an attractive form of cooling is single-pipe motor ventilation. This employs a remote fan, drawing air from a clean source and delivering it through a pressurised duct system to an adapter on the commutator-end end shield (IC17 of IEC 60034-6/EN 60034-6). Alternatively, double-pipe ventilation supplies cooling air from a remote fan as described, and in addition a discharge duct takes the used and warmed air away from the motor (IC37).

Chapter 1.7


: ~i


Figure 1.32 Typical arrangement of a forced ventilating fan on a D.C. motor (courtesy Leroy-Somer)

Ducted ventilation possesses the advantage that injurious gases and contaminants are unlikely to invade the motor winding space from the working environment. Additionally, the use of double-pipe ventilation and the introduction of (say) a ten-minute purging period before main motor starting may sometimes allow the IP55 enclosed D.C. motor to be used in an atmosphere where there is an explosive gas risk, for example in the printing industry, where the volatiles from some inks constitute a hazard. Although ductwork costs have to be taken into account, such single or double-pipe ventilated D.C. motors can prove an attractive alternative to totally enclosed D.C. motors, with either closed-air circuit-air cooling (CACA) or closed-air circuit-water cooling (CACW), IC0161, IC3666 and IC3166. For the screen-protected IP23 D.C. machine, it is relatively easy to arrange the machine with a frame-mounted fan, blowing constant velocity air through the winding space. For the much less popular IP54 totally enclosed fan-cooled machine, the fan being motor-shaft mounted, the mechanical arrangement is less satisfactory, requiring a separate cooling fan motor to be mounted externally to the motor but inside the cooling fan cowl. This in turn can pose a problem when the application requires a tachogenerator feedback signal, since the tachogenerator is normally mounted coaxially at the rear of the motor. With close-coupled bearingless tachogenerators of short axial depth, however, it is possible to arrange satisfactory mounting which is compatible with the forced cooling arrangement. CACW and CACA machines are available as both A.C. and D.C. designs. Although a relatively expensive solution, this is worth investigating where difficult ambient conditions preclude lower cost alternative enclosures. The differing patterns of development mentioned give the A.C. cage motor the enormous advantage of being

available both for the difficult and for the explosion-hazard environment in the flameproof enclosure form. Flameproof D.C. motors are of restricted availability, particularly above 20kW, and on account of small demand are relatively expensive. The more common forms of motor cooling are specified by IEC 60034-6 (EN 60034-6).

Air Filters
The fitting of an air filter to a forced ventilation fan can provide useful protection against internal motor contamination. However, heavy contamination of the filtering element can reduce cooling airflow markedly. Although thermal devices protect against this circumstance, closer, more direct protection is possible by using an airflow, or air-proving switch. This is generally arranged to monitor the air pressure driving the air through the motor winding space. As filter contamination gradually builds up, pressure falls and the air switch will indicate an alarm condition when the preset limit is detected. The drive system can be arranged either to shut down immediately or after a preset time interval to allow the driven machine to be cleared of its product or to give a warning to the machine or process operator of imminent shutdown. By providing accurate and reliable protection, particularly for A.C. cage motors (which on account of their nonlinear electrical characteristics require special care in protection), the facilities inherent in modem electronic motor controls and associated sensors have enhanced the standard of reliability of modem industrial drive systems.

The capacity of an electrical machine is very often temperature dependent, and therefore the duty cycle of the



where PN and n N are the rated power and speed, respectively, of the motor.

Short-time D u t y - $2




With a short-time duty cycle the on-load period is too short for the motor to reach a steady thermal condition and the subsequent off-load period is long enough for the temperature of the motor to drop to that of the cooling medium (even with the motor at rest). Starting and braking are not taken into consideration. With $2 duty the load torque may be greater than the rated torque, but only for an appropriately short period. When specifying short-time duty $2 it is also necessary to state the on period e.g. $2-30 min. The standard specifications recognise the following on periods: 10min, 30min, 60 min and 90 min.

Figure 1.33 Load, lossesand temperature for duty type S1

application may significantly affect the rating IEC 60034-1 defines duty cycles as follows.

Intermittent D u t y - $3
An intermittent duty rating refers to a sequence of identical duty cycles. Each cycle consists of an on-load and off-load period with the motor coming to rest during the latter The on-load period during one cycle is too short for the motor to reach a steady thermal condition, and the off-load period is likewise too short for the motor to cool to the temperature of the cooling medium. Starting and braking are not taken into account on the assumption that the times taken up by these events are too short in comparison with the on-load period, and therefore do not appreciably affect the heating of the motor. The load torque during one cycle may be greater than the rated torque of the motor. When stating the motor power for this form of duty, it is also necessary to state the cyclic duration factor: cyclic duration factor = (on time/cycle time) 100% The standards specify that the duration of one cycle must be shorter than l Omin. Cases where the duty cycle is longer than 10 min must be brought to the attention of the motor manufacturer.

Continuous D u t y - $1
Continuous duty rating S 1 relates to a duty where the onload period is long enough for the motor to attain a steady thermal condition. With rated load, this refers to its maximum permitted temperature. Starting and braking are not taken into consideration on the assumption that the duration of these events is too short to have any effect on the temperature rise of the motor. Short time overloads are acceptable. An off-load dwell period is of no significance if it is followed by a load run. The load torque must not exceed rated torque.

Example o f motor selection for continuous duty S1

The application calls for a power P1 over a speed range of nal ~_ rta ~_ na2 (min- 1). The motor is rated according to the maximum load torque and the maximum speed na2. The selected motor must comply with the following requirements: rated power: rated speed: rated torque:
PN > P1 x (na2/nal) nN > na2
PN/nN >_P1/nal

kW mm - - 1 kW/min- 1








Figure 1.34 Load, lossesand temperature for duty type $2

Figure1.35 Load, lossesand temperature for duty type $3

Chapter 1.7


load load





/ TL


Figure 1.36 Load, lossesand temperature for duty type $4

Figure 1.38 Load, lossesand temperature for duty type $6

With short cycle times starting and braking must be taken into account and the motor temperature rise must be checked (see $5). The standards specify certain preferred cyclic duration factors: 15, 25, 40 and 60 per cent. The motor rating for intermittent duty $3 may be increased over that applied to S 1 duty by a factor ks3. This factor may be of the order of magnitude 1.4. Motor manufacturers can advise on this.

If the motor had a ks3 of 1.4 then we could choose a motor with an equivalent continuous duty rating of 56.52/ 1 . 4 - 41 kW at 1200 min-1.

Intermittent Duty with Starting- S4

$4 is similar to $3 but taking starting into account.

Example of motor selection for continuous duty 53

0.5 kNm torque M 2 - 0.4 kNm torque M 3 - 0.6 kNm cycle time T - 120s;
torque M1-

Intermittent Duty with Starting and Electric Braking- S5

$5 is similar to $3 but taking starting and electric braking into account.

for a period for a period for a period motor speed

tl -- 10 s

t2--30 s t3 = 5 s = 1200min -1

Continuous Operation Periodic Duty- S6

$6 is similar to $3 except that the duty cycle is such that the motor has not returned to the temperature of the cooling medium by the end of the off period. The cycle should be repeated until the temperature at the same point on the duty cycle has a gradient of less than 2C per hour.

cyclic duration factor = [(10 + 30 + 5) / 120] x 100 % = 37.5% select cyclic duration time of 40 per cent: r.m.s, motor torque - v/{[(0.52 x 10) + (0.42 x 30) + (0.62 5)]/[10 + 30 + 5]} = 0.45 kNm motor rating PN = (27r/60) x 0.45 x 1200 = 56.52 kW @ 40 % cyclic duration time

Continuous Operation Periodic Duty with Electric Braking- S7

$7 is as $6 but taking electric braking into account.







Figure 1.37 Load, lossesand temperature for duty type 55

Figure 1.39 Load, lossesand temperature for duty type $7



Duty Cycles



Duty with Nonperiodic Load and Speed Variations- $9

$9 duty cycles should be discussed with the motor manufacturer as the effects of this duty cycle are heavily dependent upon specific motor design philosophies.




~) rnax


Duty with Discrete Constant Loads - Sl0




As with the $9 duty cycle, $10 duty cycles should be discussed with the motor manufacturer.

Figure 1.40 Load, lossesand temperature for duty type $8




Terminal markings and directions of rotation are set out in the international standard IEC 60034-8 (EN 60034-8). NEMA also defines the terminal markings in NEMA standards publication no. MG1. For clarity the two standards will be discussed separately.



~..A ~



IEC 60034-8
IEC 60034-8 describes the terminal windings and direction of rotation of rotating machines. A number of broad conventions are followed: windings are distinguished by a CAPITAL letter (e.g. U, V, W) end points or intermediate points of a winding are distinguished by adding a numeral to the winding (e.g. U 1, U2, U3)

Figure 1.41 Load, lossesand temperature for duty type $9


0 Pv

I- "1" I

1 t


IEC 60034-8 defines the direction of rotation as an observer facing the shaft of the motor (viewing from the drive end). In A.C. polyphase machines (without a commutator) the direction of rotation will be clockwise when the alphabetical sequence of the terminal letters of a phase group corresponds with the time sequence of the supply/terminal voltages. For D.C. machines see below.


01 $A02

1 t






1 ff

Figure 1.42 Load, lossesand temperature for duty type SI0

The terminal markings for three-phase stator windings of synchronous and asynchronous machines are marked as shown in Figures 1.43 to 1.46.

Continuous Operation Periodic Duty with Related Load Speed Changes- $8

$8 is as $7 but with defined and cyclic load speed changes.


The terminal markings for D.C. commutator machines are as shown in Figures 1.47 to 1.51.

Chapter 1.7






Figure 1.51 Separately excited field winding with two terminals

Compensated motor with compensating and commutating windings for clockwise rotation U2 V2 W2 E1

Figure 1.43 Single winding with six terminals

I E2
A1 ~ A 2

Figure 1.44 Delta connection with three terminals

AlE1 or AE or A or A1
Figure 1.52 Compensated motor



C2E2 or CE or C or C2


N ()

Compound motor with commutating windings for clockwise rotation


Figure 1.45 Star connection with four terminals



A1 ( ~ A


B2 D2E2 ou/or DE ou/or D ou/or D2

Figure 1.46 D.C. excitation winding of a synchronous

AlE1 ou/or AE ou/or A ou/or A1 A10 O OA2

Figure 1.53 Compound motor

Figure 1.47 Armature winding with two terminals

Shunt-wound D.C. m o t o r - connections for clockwise rotation B10 / r ' Y ' ~ ' - " ~ O B2

Figure 1.48 Commutating winding with two terminals

A1 C10 ~ C2 E1


Figure 1.49 Compensating winding with two terminals


Figure 1.54 Shunt-wound motor


O D2

Figure 1.50 Series excitation winding with two terminals

Note: the direction of rotation will be clockwise, regardless of voltage polarity, if connections are made as in Figure 1.54.


MECHANICAL AND ENVIRONMENTAL: T e r m i n a l M a r k i n g s and D i r e c t i o n o f R o t a t i o n

Series D.C. m o t o r - connections for clockwise rotation


Table 1.7 General guidance for terminal markings

armature brake alternating-current rotor windings (slip rings) capacitor control signal lead attached to commutating winding dynamic braking resistor field (series) field (shunt) line magnetising winding (for initial and maintenance magnetisation and demagnetisation of permanent magnet fields) A1, A2, A3, A4 etc. B1, B2, B3, B4 etc. M1, M2, M3, M4 etc. J1, J2, J3, J4 etc. C

A2 d

Figure 1.55 Series motor

Note: the direction of rotation will be clockwise, regardless of voltage polarity, if connections are made as in Figure 1.55.

Series D.C. generator- connections for clockwise rotation


resistance (armature and miscellaneous) resistance (shunt field adjusting) shunt braking resistor space (anticondensation) heaters stator starting switch thermal protector (e.g. thermistor) equalising lead neutral connection

BR1, BR2, BR3, BR4 etc. S1, $2, $3, $4 etc. F1, F2, F3, F4 etc. L1, L2, L3, L4 etc. El, E2, E3, E4 etc. note: El,E3 and other odd-numbered terminals should be attached to the positive terminal of the magnetising supply for magnetisation R1, R2, R3, R4 etc. V1, V2, V3, V4 etc. DR1, DR2, DR3, DR4 etc. H1, H2, H3, H4 etc. T1, T2, T3, T4 etc. K P1, P2, P3, P4 etc. = (equals sign) terminal letter with numeral 0

O A2 J

Figure 1.56 Series generator

Note: the direction of rotation will be clockwise, regardless of voltage polarity, if connections are made as in Figure 1.56.
O' ~D
e-i i. i

shunt field


comp field

comm field

NEMA MG1 provides for general guidance for terminal markings for motors, generators and their auxiliary devices as given in Table 1.7.




Figure 1.57 Shunt-wound motor anticlockwise rotation facing nondrive end

The standard direction of rotation of the shaft for D.C. motors is anticlockwise (counter clockwise) as viewed from the end opposite to the motor shaft. The direction of rotation depends upon the relative polarities of the field and armature, therefore if the polarities are both reversed then the direction of rotation will be unchanged. Reversal can be obtained by transposing the two armature leads or the two field leads. The standard direction of rotation for generators will of course be the opposite of the case for motors. Connections for the three main types of D.C. motor are shown in Figures 1.57 to 1.62.

t" :) _ | | |

shunt field

comm field




Figure 1.58 Shunt-wound motor clockwise rotation facing nondrive end




comp field

comm field

series field

The first of these categories includes temperature, altitude and the effects of the weather (electric storms, for example). The second includes effects of the electrical supply system including system faults, voltage surges, voltage dips and power switching effects.




Figure 1.59 Series-wound motor anticlockwise rotation facing nondrive end

Again, there are different standards. Countries using a 60 Hz supply system tend to follow the ANSI or NEMA standards and the remaining countries tend to follow the IEC standards.

comm field comp field series





One of the key factors which determines the rating of an electric motor is the temperature rise at full load of its active materials. The permissible temperature rises for various classes of insulation material are specified in the standards. Both NEMA and IEC use 40C as the base ambient temperature. The life of a motor is equal to the life of its winding insulation (disregarding the wear of bearings, brushes, slip rings or commutator which can be replaced at relatively low cost). Any service conditions influencing temperature rise and thus the condition of the insulation must be given particular attention. If a motor is installed in an ambient temperature above its rated value, the permissible temperature rise will need to be reduced to keep the absolute value of the maximum temperature at its design level. This is a key consideration for all users and high ambient temperatures must be discussed with the supplier in order to ensure high availability and long life. The temperature rise in the motor results from the losses caused by the conversion of energy (electrical to mechanical) which can be expressed in the following equation: Ploss - Pelec -- Pshaft

Figure 1.60 Series-wound motor clockwise rotation facing nondrive end

F 1.[ .

o ''
"~ |
| | | r

shunt field comp


. . . . . . .

series field

. . . .







1.61 Compound-wound motor anticlockwise rotation facing nondrive end

F 1_[ .
0 I i ( D i l l I

shunt comm


1.1.1 i i i i i i i


series field






In practice, it is not the losses of a motor but its efficiency 07) which is quoted; this is calculated as follows: Z] -- (Pshaft 100)/Pelec The resulting energy losses are stored in the motor and the greater part is dissipated to the surrounding atmosphere by ventilation, the condition depending upon the heat storage capacity of the motor and the temperature rise. With constant load, the steady-state condition is reached when the amount of heat produced by the losses in the machine is equal to the heat dissipated. In continuous duty, this state of equilibrium is typically reached for industrial motors after about three to five hours. The resulting temperature rise of the winding and other parts of the motor is the difference between the temperature of the particular motor part and the coolant temperature. It may be determined from the increase in resistance of the winding:
0 -- [(Rw - RK)/RI,:] x

Figure 1.62 Compound-wound motor clockwise rotation facing nondrive end


NEMA MG1 states quite clearly that terminal markings of polyphase induction machines are not related to the direction of rotation. For synchronous machines, numerals 1, 2, 3 etc. indicate the order in which the terminals reach the maximum positive values (phase sequence) with clockwise shaft rotation. Again, it is best to consult with the manufacturer.


There are two main categories of ambient condition: 1 those due to geographic conditions 2 those which are man made

(235 +




where O is the temperature rise of the winding (C), t c o l d is the temperature of the cold winding (C), tcoolant is the temperature of the coolant (C), Rw is the motor winding resistance at operating temperature (~) and Rx is the motor winding resistance when cold (f~).



As a rule of thumb the life of typical winding insulation decreases by about 50 per cent for each 10C. It should be noted that the flame temperature is neither a criterion for the quality of the motor nor for the temperature rise of the winding. An extremely cold motor may have higher losses and higher winding temperatures than an extremely warm motor.

0 -~ 0.9
0'} c'-

~ o.8

Owing to the fact that air density reduces with increasing altitude, it is necessary to allow for the resulting reduction in cooling capacity of the air when motors are operated at altitudes in excess of their rating. It is normal for motors to be rated for a maximum altitude of
1000 m.



voltage unbalance, %
Figure 1.63 Derating factor due to unbalanced supply


Voltage unbalance in per cent may be defined as: percentage voltage unbalance

Some manufacturers rate their machines for combinations of ambient temperature and altitude without derating, for example 40C/1000 m or 30C/2000 m or 20C/3000 m. It should be remembered that, although the outdoor temperature at higher altitudes is usually low, the motors will probably be installed indoors at higher ambient temperatures. The IEC recommendation is to reduce the permissible temperature rise by 1 per cent per 100 m above 1000 m. The operating altitude is important and should be specified when purchasing a motor.

= 100x

maximum voltage deviation from average average voltage

For example, with voltages of 400 V, 408 V and 392 V, the average is 400 V, the maximum deviation from the average is 8 V and the percentage unbalance is [100 x (8/400)] = 2 per cent. Voltage unbalance can produce serious overheating in A.C. motors due to the high negative sequence currents which flow with a relatively small out of balance voltage component. This negative sequence voltage produces in the air gap of the motor a flux rotating against the rotation of the rotor, tending to produce high currents. It is important, on all supplies where voltage unbalance may be a problem, to provide protective devices to trip the motor if the sustained unbalance exceeds 3 per cent. A derating curve is given by NEMA in MG1 which should be applied to motors operated on an unbalanced supply. The standard also recommends that motors should not be operated on supplies with a voltage unbalance in excess of 5 per cent. The derating curve is given in Figure 1.63.

Power Supply System

The windings of any electrical machine must be designed to operate on the supply to which it is to be connected. Further, it is necessary to coordinate the protection of the motor and its cables within the system. The constraint frequently imposed on any user of a supply system is the maximum current or kVA which can be drawn during starting. This constraint may lead the user to consider altematives to direct-on-line starting (DOL) such as, in order of increasing flexibility, star-delta starting, an electronic soft start or a variable-frequency inverter. When considering the supply constraints at start, it is important to take into account the supply impedance to ensure that there is sufficient voltage at the machine terminals to provide sufficient torque capability to overcome the load torque.


Noise and vibration are both unwanted cyclic oscillations. Vibration can be considered as structure-borne noise as opposed to airborne noise.


It is laid down in IEC 60034-1 that motors must be capable of delivering their rated output at supply voltages between 95 per cent and 105 per cent of the rated value. NEMA MG1 describes conditions of supply variation between 90 per cent and 110 per cent of the rated value. It states that operation under such conditions for extended periods of time may accelerate the deterioration of the insulation system of the motor.

Magnetic, mechanical and airflow inaccuracies due to construction lead to sinusoidal and pseudosinusoidal vibrations over a wide range of frequencies. Other sources of vibration can also affect motor operation, such as poor mounting, incorrect drive coupling, end-shield misalignment etc.

Chapter 1.7
Consider the vibrations emitted at the operating frequency, corresponding to an unbalanced load the amplitude of which swamps all other frequencies and on which the dynamic balancing of the mass in rotation has a decisive effect. In accordance with ISO 8821, rotating machines can be balanced with or without a key or a half key on the shaft extension. ISO 8821 requires the balancing method to be marked on the shaft extension as follows: H half-key balancing F full-key balancing N no-key balancing The testing of the vibration levels is undertaken with either the motor suspended, Figure 1.64, or mounted on flexible mountings, Figure 1.65. The vibration speed can be chosen as the variable to be measured (in mms-1). This is the speed at which the machine moves either side of its static position. As the vibratory movements are complex and nonharmonic, it is the r.m.s, value of the speed of vibration that is used to express the vibration level. Other variables, which could be measured, are the vibratory displacement amplitude (in microns) or vibratory acceleration (in m s-Z).


If the vibratory displacement is measured against frequency, the measured value decreases with the frequency. Highfrequency vibrations are not taken into account. If the vibratory acceleration is measured against frequency, the measured value increases with frequency. Low-frequency vibrations (unbalanced loads) cannot be measured. The maximum value of r.m.s, speed of vibration is the variable chosen by the standards and is generally classified as in Table 1.8 for medium-sized machines.

It is inevitable that even an economically designed and efficient electric motor will produce audible noise, due for example to magnetic torsions and distortions, bearings or airflow. The latter type of noise is most predominant in twopole and four-pole machines with shaft-mounted fans and most D.C. machines where a separately-mounted fan is most common. Procedures for testing for motor noise are clearly laid out in the standards; although there are detailed differences between them, the principles are the same. Referring to Figure 1.66, a series of background soundpressure readings is taken at the prescribed points. The motor will then be run on no load and at full speed. A.C. motors will be supplied at rated voltage and frequency. Synchronous machines will be run at unity power factor. Certain correction factors may be applied where the test reading is close to the background reading. In most industrial applications the motor is not the predominant source of noise and the overall situation must be taken into account when making a noise analysis. Many steps can be taken to reduce the noise of electrical machines including: the use of oil-lubricated sleeve bearings which are much quieter than most other bearing types careful choice of bearing lubricant which can affect the noise of the machine careful design of the machine air circuits to minimise ventilation noise

Figure 1.64 System for suspended machines (measuring points as indicated)

f ~



] ~'~

Although the D.C. motor is generally less troublesome than the induction motor, there are applications where special action is needed. The following measures can be made to reduce noise: use a reduced magnetic loading (lower flux densities) increase the number of armature slots skew the armature slots (or, less commonly, the pole shoes) use continuously graded main pole gaps or flare the gaps at the edges of the main pole increase the air gap brace the commutating poles against the main poles use semiclosed or closed slots for the compensating winding select the pitch of the compensating winding slots to give minimum variation in air-gap permeance

Figure 1.65 System for machines with flexible mountings (measuring points as indicated)



Table 1.8 Maximum r.m.s, speed o f vibration Class Speed n (min- ]) Frame size H (mm) 8 0 < H < 132 N (normal) R (reduced) S (special) 600 < N < 600 < N < 1800 < N < 600 < N < 1800 < N <

132 < H < 225 2.83 1.13 1.76 0.70 1.13

225 < H < 315 4.45 1.76 2.83 1.13 1.76

3600 1800 3600 1800 3600

1.76 0.70 1.13 0.44 0.70

i i


~-- - -O- . . . . . . O- . . . . . . .
i i i i i




i i i I I | | i | i


.... h
i i i

| | | i | I |


- ~

, ....

4 ........

Figure 1.66 Location of measuring points for horizontal machines a vertical plane b horizontal plane

use a twelve-pulse rather than a six-pulse D.C. drive or fit a choke/reactor in series with the machine to reduce the current ripple After the designer has taken whatever steps to minimise the noise generation at source, it may still be higher than is acceptable. To achieve the specification it may now be necessary to apply external silencing. This may be in the form of inlet or outlet air duct silencers or even the fitting of a complete enclosure. The use of acoustic partitioning requires a good knowledge of at least the octave band sound pressure levels present in order that good silencing can be achieved. Airborne noise striking an acoustic partition will, like other forms of energy, be dissipated in various ways as shown in Figure 1.67.

mechanical vibration absorbed energy(heat)

incident e n ~ (airbornen o ~ reflected ~ energy

transmitted energy (airbornenoise)

Figure 1.67 Energy flow in an acoustically excited partition

Chapter 1.7



Standard industrial motors are often unsuitable as direct drives for low-speed applications. Moreover, the use of motors with low and medium ratings is uneconomic at low speeds. Geared motors are available for such applications. These units consist of a high-speed motor and a gear reducer assembled to form an integral unit. The hardened teeth of the gear wheel resist high stressing and ensure long life of the assembly. Geared motors are widely used on single machines such as tower cranes, lifts/elevators, construction machinery, in agriculture etc., as well as in industrial plants.
O" L


o~ 60

N 40
20 0

I 20


60 speed, %



Figure 1.68 Torque~speed characteristic of a torque motor

Brake Motors
Mechanical brakes are often used in conjunction with motors instead of, or as well as, electrical braking circuits. These units consist of a motor and a brake assembled to form an integral unit. It is important to note that the brake may be rated to brake the motor and its load or may be rated to provide a holding duty only. A holding only brake will be quickly destroyed if it is used to brake a load from speed. It is common practice particularly on brushless servomotors that the brake be rated for holding duty only. In order to rate a brake motor correctly the following information about the application is necessary:

environment, where explosive gas-air mixtures may occur in dangerous concentrations. The decision as to whether an outdoor area or an enclosed location should be considered subject to an explosion hazard as defined by the relevant regulations and specifications rests entirely with the user or, in case of doubt, the competent inspecting authority. It should be realised that the degree of hazard is variable and this has led to the concept of area classification and the development of design techniques to ensure that electrical equipment will operate safely in the specified hazardous area zones. Harmonised standards exist in Europe under the guidance of CENELEC. In North America, emphasis is also placed on test and accreditation, under the guidance of Underwriters' Laboratories (UL) and CSA; however the nomenclature is different from European practice.

type of load and the type of duty of the unit frequency of braking cycles per hour total inertia (motor, brake, gearbox (if fitted) and load) referred to motor speed load torque as a function of speed, referred to the motor shaft whether the load torque has a braking or accelerating effect braking time and braking torque required

The seven commonly recognised methods of protection, as published by CENELEC are as follows: EN 50 014 EN50015 EN50016 EN50017 EN50018 EN50019 EN50014 general (Ex)o (Ex)p (Ex)q (Ex)d (Ex)e (Ex)i

Brake motors can be designed to be fail safe - if the power is lost than the brake will automatically be applied.

Torque Motors
Torque motors have been developed from the basic designs of three-phase squirrel-cage induction motors. They are not designed for a definite output but for a maximum torque, which they are capable of delivering at standstill and/or at low speed (when supplied from a fixed-frequency supply). They have a torque speed curve of the form shown in Figure 1.68.

Although comprehensive guidance on the selection of explosion-protected equipment is contained within the standards, the information to be considered falls into four categories. (i) The type of protection of the apparatus in relation to the zonal classification of the hazardous area. The degree of protection required is dependent upon the presence of ignitable contaminations of inflatable gas or vapour in relation to the length of time that the explosive atmosphere may exist, and this is defined as in Table 1.9. (ii) The temperature classification of the apparatus in relation to the ignition temperature of the gases and vapours involved.


Manufacturing processes in many sectors of industry can be described as hazardous by the nature of the operating


MECHANICAL ENVIRONMENTAL:Motors for Hazardous Locations AND

Table 1.12 Equipment suitable for different zones Zone 0

Table 1.9 Protection in hazardous locations Zone 0 Zone 1 Zone 2 a zone in which an explosive gas-air mixture is continuously present, or present for long periods a zone in which an explosive gas-air mixture is likely to occur in normal operation a zone in which an explosive gas-air mixture is not likely to occur in normal operation, and if it occurs will only exist for a short time

Type of protection (Ex)ia (Ex)s - specifically certified any of the above plus (Ex)d; (Ex)ib; (Ex)p; (Ex)e; (Ex)s any of the above plus (Ex)N or (Ex)n; (Ex)o; (Ex)q

Table 1.10 Temperature classification Maximum surface temperature (C) T1 T2 T3 T4 T5 T6 450 300 200 135 100 85

These recognised types of protection are as follows: (Ex)o - oil immersed All or part of the apparatus is immersed in oil to prevent ignition. (Ex)p - pressurised Since it is not practical to manufacture explosion-proof motors in large sizes, it is common practice to employ pressurised motors for zone 1 applications. These motors tend to be manufactured to normal industrial standards except that special attention is paid to the sealing of all removable covers and to shaft seals. The motors must be totally enclosed. Cooling must be by air-to-air or air-to-water heat exchangers. Before the motor is energised it must be purged with at least five times its own volume of clean air to remove any flammable gases which may be present. (Ex)q - p o w d e r filled

Table I. 11 Ignition temperatures


Example of compound Acetone Butane Hydrogen sulphide Diethyl ether Carbon disulphide

Ignition Suitable equipment regarding temperature temperature classification


535 365 270 170 100


T2 T2

T3 T3 T3

T4 T4 T4 T4

T5 T5 T5 T5 T5

T6 T6 T6 T6 T6

This form of protection is very unusual in rotating machines. ( E x ) d - explosion p r o o f

Electrical apparatus must be selected to ensure that the maximum surface temperature is below the ignition temperature of the specified gas. EN50 014 gives temperature classifications for equipment as in Table 1.10. These classifications can be related to the ignition temperatures given in Table 1.11. (iii) The apparatus subgroup (if applicable) in relation to the relevant properties of the gases and vapours involved. Explosion-protected electrical apparatus is divided into two main groups: group I group II mining applications all nonmining applications

All parts of the motor where igniting arcs or sparks may be produced are housed in a flameproof enclosure. The sealing faces, cable entries, shaft glands etc. are made with comparatively large gap length and limited gap clearances to prevent ignition of the surrounding explosive atmosphere. During operation explosive mixtures penetrate only seldom into the enclosure. Should an internal explosion occur, it is prevented from spreading to the external atmosphere. (Ex)e - increased safety This type of construction is used for motors without commutators or slip rings, which do not produce sparks during normal operation. This type of motor may be used in zone 1, with some qualifications, and in zone 2 areas. It is required that all surface temperatures are kept within the ignition temperature of the specified gas under all conditions of operation or fault. To avoid the danger of ignition in the event of a fault, suitable protective devices such as circuit breakers with matched thermal characteristics should be used to protect the motor against overheating. The worst abnormal condition that can occur without permanently damaging the motor is a stalled condition. Most

For some types of protection, notably flameproof enclosures, it is necessary to subdivide group II according to the properties of the gases, vapours or liquids, since apparatus certified and tested for, say, a pentane-air mixture will not be safe in a more easily ignitable hydrogen-air mixture. This has led to apparatus subgrouping i.e. IIA, liB and IIC. (iv) The suitability of the apparatus for the proposed environment. Explosion-protected apparatus appropriate to a particular zone is readily identified by reference, Table 1.12.

Chapter 1.7


motor designs are rotor critical, the rotor temperature increasing more rapidly than the stator under stalled conditions. The surface temperature of the rotor conductors is the critical and limiting factor in this type of motor. The te characteristic of (Ex)e motors is important and must be quoted on the nameplate. It is defined as the time taken for a winding, when carrying the worst case current, to be heated up from the temperature under rated operating conditions to the limiting temperature, te must never be less than 5 s.

Testing authorities
The main US and EEC testing authorities are shown in Tables 1.15 and 1.16.

Table 1.13 US and European temperature classification

European classification (EN50 014) T1 T2 North American classification (NEC NFPA 70) T1 T2 T2A T2B T2C T2D T3 T3A T3B T3C T4 T4A T5 T6 Maximum surface temperature (C) 450 300 280 260 230 215 200 180 165 160 135 120 100 85

(Ex)i - intrinsically safe

The concept of intrinsic safety is based upon restricting the electrical energy within the apparatus and its associated wiring to prevent the occurrence in normal operation of incendive arcs, sparks or hot surfaces. It is necessary to ensure that high voltages cannot be induced into the intrinsically safe circuit. Shunt diodes are usually employed as barriers between the intrinsically safe and the hazardous areas. This method is used in signalling, measuring and control circuits but is not practical for motors.


T4 T5 T6

(Ex)s- special protection

This protection concept allows for certification of equipment which does not comply with the specific requirements of the established forms of protection.

Table 1.14 US hazardous area classification

Class I gas or vapour group C group D ethyl-ether, ethylene, cycle propane gasoline, hexane, naphtha, benzene, butane, propane, alcohol, lacquer vapours, and natural gas carbon black, coal or coke dust flour, starch or grain dust fibres easily ignitable but not able to be suspended in air to produce ignitable mixtures, such as rayon, nylon, cotton, saw dust and wood chips

(Ex) N a n d (Ex)n
In the designation of type-N apparatus, the upper case N is used in the UK, but the lower case n has been proposed for the European standard having a similar concept. Nonsparking motors, which are suitable for use in zone 2 areas, are supplied in the UK to BS5000:Part 16. No part of the motor may exceed 200C during normal operation but it may do so during starting. Class II hazardous dusts easily ignitable fibres group F group G Class III

North American Standards

The principles applied in North America are broadly similar to those in Europe. The key differences are as follows.

Table 1.15 European Economic Community approved testing authorities

Country Belgium Denmark France Germany Italy UK Name 1NIEX DEMKO Cerchar LCIE BVS PTB CESI BASEEFA Location Paturages Herlev Verneuil Paris Dortmund-Derne Braunschweig Milan Buxton

Temperature classification
Although the basic temperature classifications of the european standards are retained, interpolation has occurred between some T classifications giving greater resolution. Table 1.13 shows the result (with cross reference to the European classification).

The apparatus subgroup

In the US, the system of area classification and gas grouping is again different to European practice. Here the hazardous area is divided into flammable gases or vapours and combustible dusts. The key classifications are given in Table 1.14. Class I, group D is approximately the equivalent of the European Group IIA with temperature classification to suit the specific explosion hazard.

Table 1.16 Main North American testing authorities

Country Canada USA Name CSA Factory Mutual Underwriters' Laboratory Location Toronto Norwood Northbrook




The control of electric motors by means of power electronic converters has a number of significant effects. These are primarily due to the introduction of harmonic components into the voltage and/or current waveforms applied to the motor. In the case of A.C. machines which are normally considered to be of fixed speed there are additional implications which need to be considered including mechanical speed limits and the possible presence of critical speeds within the operating speed range.

Machine Rating-Thermal Effects

Operation of A.C. machines on a nonsinusoidal supply inevitably results in additional losses in the machine. These losses fall into three main categories: (i) Stator copper loss - this is proportional to the square of the r.m.s, current. Additional losses due to skin effect must also be considered. Rotor copper loss - the rotor resistance is different for each harmonic current present in the rotor. This is due to skin effect and is particularly pronounced in deep bar rotors. Since the rotor resistance is a function of frequency, the rotor copper loss must be calculated independently for each harmonic. The increase in rotor copper loss caused by harmonic currents is very often a significant component of the total losses, particularly with PWM inverters which have significant higher harmonics for which slip and rotor resistance are high. Iron loss - this is increased by the harmonic components in the supply voltage. The increase in iron loss owing to the main fluxes is usually negligible, but there is a significant increase in losses due to end winding leakage and slew leakage fluxes at the harmonic frequencies.



The effects due to deviation from a smooth D.C. supply are, in general, well understood by drive and motor manufacturers. The impact of ripple in the D.C. current clearly increases the r.m.s, current, which leads to increased losses and hence reduced torque capacity. The harmonics associated with the current ripple lead to the now universal practice of using laminated magnetic circuits, which are designed to minimise eddy currents. With chopper converters, which are used in servo amplifiers and traction drives, frequencies in excess of 2 kHz can be impressed on the motor. Special care is needed to select a motor with sufficiently thin laminations. The ripple content of the D.C. currents significantly affects commutation within a D.C. machine. The provision of a smoothing choke can be extremely important in this respect, and recommendation should be made by the motor manufacturer depending upon the supply converter used. Apart from the thermal and commutation impacts, the ripple current also results in pulsating torque, which can cause resonance in the drive train. Laminating the armature not only improves the thermal characteristic of the motor but also its dynamic behaviour by decreasing the motor time constant.


The total increase in losses does not directly relate to a derating factor for standard machines since the harmonic losses are not evenly distributed through the machine. The nonfundamental/harmonic losses mostly occur in the rotor and have the effect of raising the rotor temperature. Whether or not the machine was designed to be stator critical (stator temperature defining the thermal limit) or rotor critical temperature clearly has a significant impact on the need for, or magnitude of, any derating. Many fixed-speed motors have shaft-mounted cooling fans. Operation below the rated speed of the motor therefore results in reduced cooling. Operation above the rated speed results in increased cooling. This needs to be taken into account by the motor manufacturer when specifying a motor for variable-speed duty.


It is often stated that standard off-the-shelf A.C. motors can be used without problem on modem PWM inverters. Although such claims may be largely justified switching converters do have an impact and certain limitations do exist. NEMA MG1-1987, Part 17A gives guidance on operation of constant-speed squirrel-cage induction motors for use on a sinusoidal bus with harmonic content and general purpose motors used with variable-voltage or variable-frequency controls or both.

Machine Insulation
Current-source inverters feeding induction motors have motor terminal voltages characterised as a sine wave with the superposition of voltage spikes caused by the rise or fall of the machine current at each commutation. The rate of rise and fall of these voltage spikes is relatively slow and only the peak magnitude of the voltage is of practical importance in considering the impact on machine insulation. The supply voltage never exceeds twice the crest voltage of the sinusoidal waveform, and is consequently below almost all recognised insulation test levels for standard machines.

Chapter 1.8
Current-source inverters feeding synchronous machines are even gentler on insulation systems, as the sinusoidal terminal voltage is reduced during commutation producing the same effect as notching on the supply associated with supply converters.


For supply voltages less than 500 V A. C. Check that the motor has the capability to operate with a PWM drive. Most reputable motor manufacturers have assessed their products for drive applications and can give an assurance of compatibility.
Alternatively, Figure 1.69 shows the peak voltage/rise-time withstand profile, which is required for reliable operation. The motor supplier should be asked to confirm this capability. Figure 1.69 also shows the capability of a typical good quality motor, which comfortably exceeds the requirement. However, note that conformity of the motor with IEC 60034-17 alone is not sufficient.

PWM inverter drives are used with standard induction motors in very large numbers throughout the world, and their advantages are well known in terms of improved energy efficiency and flexibility of control. Occasionally drive users are advised to take special precautions over the motor terminal voltage because of an effect sometimes referred to as spikes or dv/dt which could possibly damage the motor insulation. This section explains the effect and prescribes the steps which should be taken to ensure that the motor insulation system gives a long reliable life when used with a PWM drive.

The main effects of PWM drive waveforms on motor insulation are as follows: Motor winding insulation experiences higher voltages when used with a PWM inverter drive than when driven directly from the A.C. mains supply. This effect is caused by the fast-rising PWM voltage pulses which result in a transiently uneven voltage distribution across the winding, as well as short duration voltage overshoots because of reflection effects in the motor cable. It is a system effect that is caused by the behaviour of the drive, cable and motor together. For supply voltages up to 500V A.C., the voltage imposed by a correctly designed inverter is well within the capability of a standard motor of reputable manufacture. For supply voltages over 500V A.C., an improved winding insulation system is generally required to ensure that the intended working life of the motor is achieved. When the motor used is of uncertain quality or capability, additional circuit components can be added to protect it.

For supply voltages in the range 500 V-690 V A. C. Select an inverter-rated motor. An enhanced insulation system is required. The permitted voltage/rise-time curve should equal or exceed that shown in Figure 1.70. Figure 1.70 also shows the capability of a typical inverter-rated motor for use up to 690 V, which comfortably exceeds the requirement. Note, however, that conformity of the motor with NEMA MG31 alone is not sufficient.
(b) Alternative a p p r o a c h - use additional preventative methods It may not be possible to follow the above recommendations, for example because the drive is to be retrofitted with an existing motor or data is not available for the motor concerned. In this case, additional preventative measures are recommended. The most cost-effective measures are usually drive output line chokes, for lower power systems, and motor cable termination networks, for higher powers. More details are given later. (c) Factors affecting motor selection: star windings are preferable to delta windings windings with single conductors are preferable to those containing parallel paths motor loading and duty should be carefully assessed to ensure that the motor does not over heat the insulation system is degraded by excessive temperature.

Guidance for avoiding problems and explanation of the phenomena involved

1 The voltage at the drive terminals is limited within tight bounds by the drive circuit. The motor cable increases the peak motor voltage. In applications with short motor cables (i.e. 10 m or less) no special considerations of any kind are required. Output inductors (chokes) or output filters are sometimes used with drives for reasons such as long-cable driving capability or radio frequency suppression. In such cases no further precautions are required because these devices also reduce the peak motor voltage and increase its rise time. In all other cases the following guidance should be followed: (a) Preferred approach - select a suitable motor

Special cases 1 High braking duty Where the drive spends a large part of its operating time in braking mode, the effect is similar to increasing the supply voltage by up to 20 per cent and the relevant precautions must be taken for the higher voltage. Active front end (regenerative/sinusoidal/unity power factor input drives) For drives with active front ends (regenerative and/or unity power factor) the effective supply voltage is increased by up to 15 per cent and the relevant precautions must be taken for the higher voltage. Special control schemes Some drive designs using flux vector control with fast acting flux orientation can generate continuous double pulses where the output voltage changes by twice the D.C. link voltage in a single step. This can result in four


EFFECTS OF SEMICONDUCTOR POWER CONVERTERS: Drive C o n v e r t e r Effects u p o n A.C. M a c h i n e s

2.6 2.4 2.2 2.0 > 1.8 ~11"~ l/ typical standard motor for up to 500 V



0.6 0.4 02
"0 O



0.6 0.8 1.0 1.2 1.4 1.6 1.8 voltage pulse rise time, ps (IEC definition)


i t2.2

Figure 1.69 Peak voltage~rise-time profile requirements for supplies up to 500 V A.C.

2.4 2.2

2.6 I 2.0 minimum requirement up to 690 V

> 1.8

; 1.2-I II
0.8 0.6 0.4 0.2
0 l l l l




0.6 0.8 1.0 1.2 1.4 1.6 1.8 voltage pulse rise time, Its (IEC definition)



Figure 1.70 Peak voltage~rise-time profile requirements for supplies up to 690 V A.C.

times the D.C. link voltage appearing at the motor terminals, causing increased stress and possibly premature motor failure. The stress is so extreme that a combination of inverter-rated motor and additional measures such as line chokes may be required to prevent motor damage. The drive supplier should be consulted for detailed guidance in this case. Please note that Control Techniques drives do not use this form of control.
Additional preventive measures The two most cost-effective techniques are:

The use of an inductor with 2 per cent impedance at the maximum output frequency is sufficient to lengthen the rise time to a point where it is no longer a consideration 5 ps is easily attainable. The natural high-frequency loss in a standard iron-cored inductor gives sufficient damping, and this is more cost effective than using a low-loss inductor with separate damping components. Commercially available dv/dt and sinusoidal filters should not normally be considered purely for motor protection, since their cost is excessive. They may, however, be specified for other reasons such as EMC or motor acoustic noise.
Choice of inductor The inductance should be chosen so that the impedance does not exceed 3 per cent pu at maximum frequency, otherwise the voltage drop will cause significant loss of torque at high speed.

Output inductors (chokes) and output filters These are all connected at the drive in series with its output. They all work by forming a low-pass filter in conjunction with the motor cable impedance, thus reducing the rate of rise of the drive output voltage. Some overshoot still occurs, which is controlled by damping or clamping. This results in some power loss, which must be allowed for in sizing the inductors or selecting the filter. The loss is roughly proportional to the motor cable length and the drive switching frequency.

Conventional iron-cored inductors are suitable. Allowance should be made for additional core loss because of the presence of high frequencies. Special low-loss

Chapter 1.8
.. .... ,. .... , .... ,. .... ! .... ... ~ ~ '. ,.


high-frequency inductors should not be used because severe resonance problems can occur. Individual phase inductors or three-phase inductors are equally effective.

Other benefits
Reduced loading effect on the drive from the cable capacitance Reduced radiofrequency emission from the motor cable (EMC)

i ........
.......... : ......... l .........



Voltage drop Power loss Cost is modest at low current ratings but increases rapidly with increasing rating For power levels above 100 kW the inductance at 3 per cent pu may be insufficient for the purpose 2 Motor cable terminating unit With increasing drive rating the above methods become increasingly expensive since they have to pass the entire drive output current. For powers exceeding about 70 kW it may be more cost effective to use a terminating unit. This is a resistor-capacitor network, which is connected at the motor terminals in parallel with the power connections and presents an impedance approximately matching that of the cable during the pulse edges. This suppresses the reflection. It does not change the rise time but it virtually eliminates the overshoot. It has the advantage of not carrying the drive output current, but power loss tends to be greater than that for an inductor and mounting at the motor terminals may be inconvenient and require a special sealed construction to match that of the motor. Figures 1.71 and 1.72 show typical waveforms produced by these methods.

: ......... i ........ ~ ........

": .........

i ........
"t ........

i ......
": .........



. . . . i . . . . : . . . . . . . . .


. . . . . . . . . . . . . . . . . . ;

. . . . . . . . .

Figure 1.71 Motor terminal voltage with inductor









. . . .





. . . .













. . . .

. . . .

A. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ~I T t + .
. . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . " " "

. . .

. . .

. . . . . . . . . . ,-I . . . . . .

. . . . . . .

. . . . . .

. . . . . . . .

. . . . . . . . . . . .

. . .

+ :I:
. ~.
. . . . .

. .
. . . . . . . .

. .
.. . . .

. . . . . . . . . . .

MI max

. . .

. . .

800 v
. .

.l.~. .

. . . . .

-~++-~ + + + ~ - ~ - ~ - + ~ ~

'-O-~H-~-H ++-,-+i + + + ~


i .... . .

: .... . .

i ....

i ....

~- ....

i ....

: ....

i ....

: ......



i .... i .... i .... i ......

/ ~ , , , i . . . . i . . . . i . . . . i . . . . "t . . . . i . . . . ' . . . . " . . . . " . . . .

Figure 1.72 Motor terminal voltage with terminating unit (note different time scale)

The unit must have an enclosure rating (e.g. IP number or NEMA category) suitable for the motor application.

Table 1.17 Relative costs of alternative techniques

Motor rating

Drive Motor Output dv/dt Sinusoidal Terminator inductor filter filter 443% 334% 99% 146% 65% not practical 170% 9% 3%

Other benefits

(400 V supply) 2.2kW 350% 100% 74% 75kW 220% 100% 14% 250kW 120% 100% 5%

May affect the control of some kinds of flux vector or other closed-loop controllers Power loss Additional cost and inconvenience of motor terminal mounting 3 Output filters More advanced output devices are available, in the form of dv/dt filters and sinusoidal filters. They have similar benefits for motor terminal voltage, but since they are relatively expensive they are unlikely to be cost effective unless they are also needed for other reasons.

Technical explanation of the phenomena

Review of PWM principles The output voltage of the drive is a series of pulses with magnitude either + VD.C. or --VD.C., where VD.C. is the drive D.C. link voltage, with pulse-width modulation (PWM). Because the motor load has inductance, the current flowing and the magnetic flux in the motor comprise mainly the underlying low frequency of the pulse-width modulation with a small ripple component at the switching frequency. Figure 1.73 illustrates in simplified form a part of the output voltage waveform, with the associated motor magnetic flux. VD.C. is typically about 1.35 times the r.m.s, supply voltage, for example 540 V with a 400 V supply.

Table 1.17 shows some relative costs of typical examples of these alternative techniques. From this it may be concluded that output inductors are the most economic measure for systems rated up to about 70 kW, beyond which terminators become more attractive.


EFFECTS SEMICONDUCTOR OF POWER CONVERTERS: Converter Effects upon A.C. Machines Drive voltage i1~~[~ I
overshoot lasts for about twice the time of flight in the cable. If the rise time of the pulse is longer than twice the time of flight in the cable, then the overshoot is cancelled before it reaches 100 per cent. For a single pulse of magnitude VD.C., regardless of the motor cable length, the overshoot can never exceed 100 per cent of VD.C.. However, the duration of the overshoot does increase with increasing cable length. For an ideal lossless cable, the rise time of the pulse is maintained along the cable so that the rate of change of voltage at the motor terminals (dv/dt) approaches twice that at the drive. However, in practice, the cable exhibits high-frequency loss, which causes an increase in the rise time. This also means that the rise time at the motor terminals is fixed mainly by the high-frequency behaviour of the cable, so that contrary to statements sometimes made it is not the case that the introduction of new fasterswitching power devices increases the stress on the motor. Note that bipolar pulses, which have pulse edge magnitudes of 2 VD.C., are also increased by 100 per cent so that the total voltage during the reflection is then 4 VD.C.. These can be generated by some kinds of drive with special vector control schemes. Figure 1.74 shows some typical measured voltage waveforms, which illustrate the effect in practice. Even with 4 m of cable some overshoot is apparent. With 42 m the overshoot is virtually 100 per cent. Winding voltage The voltage overshoot has little effect on the main motor insulation systems between phases and from phase to earth, which are designed to withstand large overvoltage pulses. Typical dielectric strengths for motors of reputable origin are about 10 kV. However, some small low-cost motors may have had economies made in the interphase insulation, which can lead to premature insulation failure. Because of its short rise time the pulse also affects the insulation between turns, and especially between coil ends. The voltage pulse travels around the motor winding as it does along the motor cable. Figure 1.75 illustrates how this results in a large part of the pulse appearing across the ends of a coil during the time between it entering one end and leaving the other. In practice even in the largest low-voltage motors the voltage between electrically adjacent turns is insignificant, but between the ends of the coil it may briefly reach a substantial part of the pulse magnitude. In this simplified illustration the entire pulse voltage appears across the coil. In practice magnetic coupling between turns reduces this. Figure 1.76 gives a summary of the results ofmeasurements made with a range of rise times on a variety of motors. With a sinusoidal supply voltage the coil ends only experience a fraction of the phase voltage, as determined by the number of series coils. With a drive, therefore, there is a considerable increase in the voltage stress between the coil ends. The effect of this depends on the motor construction. Large motors using form winding are constructed so that the coil ends are not in contact. The interturn insulation



Frequency (Hz) Period/time 20 ms 333 #s 100 ns

Figure 1.73 PWM inverter output voltage and current

Table 1.18 Typical frequencies and times

Power output Switching Pulse rise time

50 3000 -

Typical frequencies and times are given in Table 1.18. Note the timescales. The rise time is five orders of magnitude shorter than the output period. Drive designers generally aim to use the highest practical switching frequency, since this has a variety of benefits including reducing the audible noise from the motor. This means they are constantly seeking to use faster power switching devices, which give lower switching losses through shorter rise times. All of the pulse edges in Figure 1.73 have amplitude equal to the D.C. link voltage. Standard PWM controllers only generate these unipolar pulses. Some special control schemes without PWM modulators can generate bipolar pulses, which change from + VD.C. to -- VD.C. in one transition. Motor voltage The PWM pulse rise times are so short that the time for the pulse to travel down the motor cable can easily exceed the rise time. For example, the velocity of the pulse is typically 1.7 x 10s m s - 1, SOin 100 ns it has travelled only 17m. When this happens, analysis needs to rely on transmission-line theory. Full details are beyond the scope of this guide, but the essential mechanism is as follows: At each pulse edge the drive has to charge the inductance and capacitance of the cable, so a pulse of energy is delivered into the cable. The pulse travels at a velocity which is characteristic of the cable and is typically 1.7 x 108 ms-1. When the edge reaches the motor terminals, a reflection occurs because the motor surge impedance is higher than that of the cable (this is true for most low-voltage motors although the impedance does fall as the motor rating increases). The voltage tends towards double the step magnitude; i.e. there is an overshoot approaching 100 per cent. The reflection returns to the drive where it is again reflected, but in the negative sense because of the low impedance of the drive. When this second reflection returns to the motor terminals, it cancels the overvoltage. Therefore the

C h a p t e r 1.8


.................. !....M2;max...........................................................

order of 1400 V, as could be generated by a drive with a 500 V A.C. supply. In the USA where supplies of this level are common, many motor manufacturers routinely use an inverter grade wire with further enhanced insulation withstanding at least 1600 V. There is a possibility of a low-energy electrical discharge effect called partial discharge, which can occur in voids between wires. This is because of the electric field concentration in such voids where the permittivity of the gas or air is lower than that of the insulation material. At every pulse edge a small discharge of energy occurs, which may gradually degrade the insulation system. If the effect is excessive, the motor fails prematurely with an interturn fault. Resin impregnation suppresses this effect, as well as contributing to the physical stability of the winding under high mechanical stress or vibration. For supply voltages higher than 500 V further measures are required to prevent partial discharge. Inverter rated motors use inverter-grade winding wire, which is resistant to partial discharge, as well as multiple impregnation regimes to minimise voids, and enhanced interphase insulation.

'i'i'i'i'i'i'i'i'i ...................................................
, , , , . i . . | . i

. . . .

"....'.'."''. "....".".'..' ......" .....

i :

20d V
. . . . . . . . . . . . . . . . . . . : . . . . . . . .

. . . . . . . . . . . . . . . . . . .

. . . .

. . . .

! . . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

i ........ i......... i......... i ......... i ................. i ......... i ......... i ......... i ........ t

i i i M~


........ ; ......... !A .... ~ 4 v i

........ + ......... i ......... i ......... i ......... i .........

: : : : : : : : : : : : : : : : : : : : : : : :


; ....

; ....

; ....

i ....

Motor standards
.................. i ..............................................................................

. . . . . .

. . . . .

200 V






~ .........



i .....

: :

~ :


" ...................

: :

.: .........


" .........

i .

i :

" .........

: i


. .........

M2m.~ ~ . . . . . . . . . . . . . . . . . . .

i .........

~ .........

i .........

~ .........

! 13t~:d)


i2.00.; i i i [

The international standard IEC60034-17 gives a profile for the withstand capability of a minimum standard motor in the form of a graph of peak terminal voltage against voltage rise time. This replaces the older IEC3417 standard, which gave a rather arbitrary 500 V/~ts limit without a clear rationale. The new standard is based on research into the behaviour of motors constructed with the minimum acceptable level of insulation within the IEC motors standard family. There is a great deal of published technical information on this subject. The best description is contained in a paper written by workers at Dresden University who carried out a major research exercise on the subject: 'Failure mechanism of the interturn insulation of low-voltage electric machines fed by pulse-controlled inverters', M. Kaufhold et al., IEEE Electrical insulation magazine, vol. 12, no. 5, 1996. Tests show that standard PWM drives with cable lengths of 20 m or more produce voltages outside the IEC6003417 profile. However, most motor manufacturers produce as standard motors the capability of which substantially exceeds the requirements of IEC60034-17. Figures 1.69 and 1.70 give the actual requirements for supply voltages up to and exceeding 500 V, respectively. Standard motors are widely available to meet the requirements of Figure 1.69. Usually, a special inverter-rated motor is needed to meet the requirements of Figure 1.70. Such motors carry a price premium of between 3 and 10 per cent depending on the rating. Figure 1.77 gives some measured voltages for a typical system, showing that they exceed the IEC60034-17 limits but they do not exceed the capability profile of a typical standard motor from a well known manufacturer. This graph illustrates clearly the effect of lengthening the motor cable. The rise time increases steadily with increasing length, and the overshoot falls off after a peak

! 06
Figure 1.74 Motor terminal voltage waveforms for varying cable lengths (note scale changes) a cable length = 0.5 m b cable length = 4.0 m c cable length = 42.0 m then does not experience the high-voltage pulse. Smaller random-wound motors may however have coil end wires in contact, so then special attention is required to the quality of the interturn insulation. Motor interturn insulation design Modern motors of good quality manufacture use advanced winding wire, which has a multilayer insulation system and is easily capable of withstanding peak voltages of the



motor termin~l=J ,






Voltage between coil ends: i/--~ (7=propagation time around coil) ~ ' L
i i



Figure 1.75 Propagation of pulse through motor windings

1.0 0.8 from one end of the rotor shaft to the other. If the bearing breakover voltage is exceeded this will result in a current flowing through both bearings. In some large machines it is common practice to fit an insulated bearing, usually on the nondrive end, to stop such currents flowing. This mains frequency issue is well understood and with modem motors such problems are rare.

~ 0.6

0.4 0.2 0

s eady s a e

.~ rise time, gs

Supply asymmetry
An ideal power supply is balanced and symmetrical. Further, the neutral is at zero potential with respect to the system earth. With all modem PWM inverter supplies, although it can be assumed that the supply feeding the motor is indeed balanced and symmetrical in peak and r.m.s. amplitudes, it is impossible to achieve perfect balance between the phases instantaneously, when pulses of different widths are produced. The resulting neutral voltage is not zero with respect to earth, and its presence equates to that of a common-mode voltage source. This is sometimes referred to as a zero sequence voltage. It is proportional in magnitude to the D.C. link voltage (itself proportional to the supply voltage), and a frequency equal to the switching frequency of the inverter. This common-mode voltage will lead to the flow of currents through stray impedances between the inverter phase connections and earth. This includes motor cables. Considerable research into this complex subject has shown that the common-mode currents can be usefully considered in three distinct frequency ranges:

Figure 1.76 First coil voltage distribution against incident voltage rise time
at about 50 m. The voltage stress on the motor therefore falls above quite moderate cable lengths. NEMA publishes similar limits in the USA in MG1 part 31, shown in Figure 1.78. The measurements suggest that these limits are insufficient for drives operating much above 500 V. However, inverter-rated motors are readily available with much improved capability, as shown.

Bearing Currents
In theory the sum of the three stator currents in an A.C. motor is zero and there is no further path of current flow outside the motor. In practice, however, there are conditions which will result in currents flowing in or rather through the bearings of A.C. motors even when fed with a sinusoidal 50 or 60 Hz supply.

It is well understood that an asymmetric flux distribution within an electrical machine can result in an induced voltage

Supply frequency (typically up to 100 Hz) This type of current flow is usually related to motor cable asymmetry and not power supply characteristics.

b) The switching frequency of the inverter (typically 1 to

20 kHz)

Chapter 1.8
2,4 -








typical standard motor for up to 500 V ..... lEG 34-17 (new version, up to 690 V)
............... 7

m.,0 m o 1.2 ," .[20 ,rrff"..:. . . . . . . . . . . . . . "0" 1.0 100 m


0.2_0.4_0.6 5Om i.................................................................................................... - ..........

up(lversion,to d 690 V)


I "~



0.6 0.8 1.0 1.2 1.4 1.6 voltage pulse rise time, its (IEC definition)




Figure 1.77 IEC limits, manufacturers" limits and measurements Test results at voltages with SWA cable lengths as indicated: 415 V A. C. supply; II 480 V A. C. supply

2.6 2.4 2.2

$ $ $
s s s s

typical inverter-rated motor for 500V-690 V

2.0 1.8 >




NEMA (MG1 part 31) inverter-fed motor for up to 690 V


1.6 1.4 1.0 0.8 typical standard motor for up to 500 V I i "c"a'4 i T iX : m30 m50.m . . . . . . . ....................

o 1.2 >

; ; "u'p" "gg"o" i "


II 5m

100 m

.............(EC 34-17


(Oldup toVersion'690 V)

0.6 0.4

I ~"



0.6 0.8 1.0 1.2 1.4 1.6 voltage pulse rise time, #s (IEC definition)




Figure 1.78 Limits and measurements for motors rated over 500 V, and NEMA MG1 Test results at voltages with SWA cable lengths as indicated: II 480 V A. C. supply; A 690 V A. C. supply

As described above, this is the fundamental frequency of the common-mode voltage. Owing to the relatively low frequency, most of the currents which flow at this frequency return to the inverter without passing through motor bearings due to their high impedance.

Rather than consider the phenomena in further detail, it is more helpful to consider what can be done to protect against the risk of bearing currents.

c) Common-mode resonant frequencies stimulated by

inverter switching (typically 50 kHz to 5 MHz) This is considered to be the most critical and is responsible for nearly all nonmechanically induced bearing problems in inverter-fed motors. Frequency range can be limited by limiting the switching time of dv/dt of the PWM pulses.


The first thing to remember is that although a great deal has been discussed and published on the subject of bearing currents associated with inverter-fed motors, it is in practice a rare event where the particular combination of motor construction, installation and inverter has caused a problem.



Table 1.19 Maximum motor speeds and balancing for L-S MV (2/4/6 pole) motors Motor type 80 90 100 112 132 160 160 LU 180 200 225 ST/MT/MR 225M/MK 250 280 SP 315 Maximum speed (min-1) 15 000 12 000 10 000 10 000 7500 6000 5600 5600 4500 4100 4100 4100 3600 3000 Balancing

That said, any motor may be subject to beating currents if its shaft is connected to machinery at a different ground potential than that of the motor frame. In order to eliminate motor flame voltage it is necessary that a grounding strategy is adopted to keep all system components grounded at the same potential. This needs to be achieved for all frequencies, not just the 50/60Hz which many grounding practices were based upon. This means avoiding high-inductance paths - keeping cable runs as short as possible. Define a low-impedance path for the common-mode currents to flow back to the inverter. As the common-mode current flows through the three motor conductors (cable), the best return path would be through a shield around that cable. This could be in the form of a screen. Obviously, it is necessary to connect the screen at both the motor and the inverter, although this is in conflict with conventional practice on screening. Such measures are well defined by most reputable manufacturers in their EMC guidance. A conduit would act in the same way, but it is important to ensure that the conduit is capable of providing a reliably continuous high-frequency path. Conduit is designed to provide mechanical protection and may not be electrically continuous. Further, care needs to be taken to ensure that the cable lies within conduit for the entire distance between motor and inverter. In all cases take great care with the terminations of the screen. All terminations must be of low resistance and low inductance or the benefit may be greatly attenuated. An obvious action is to use symmetrical motor cables. Take care to ensure that the ground cores in the cable are symmetrically arranged to avoid asymmetrical induced currents in the motor cable.


The following issues should be considered by the motor manufacturer when sanctioning use in the overspeed range: mechanical stress at the rotor bore and assurance that the shaft to core fit is secure beating life, which is a function of the speed for antifriction beatings; each beating has overspeed and temperature limits which need to be reviewed beating lubrication, which is also a function of speed and operating temperature; grease may not adhere properly and oil may churn or froth vibration, which is a function of the square of the angular velocity; care must be given to ensure no operation near system critical or natural frequencies airborne noise can be dramatically increased at higher speeds winding stress caused by vibration of the windings at high frequency may require additional winding bracing and treatment proper attachment of balancing weights affixed to the rotor or fan assembly; at high speeds shear stress levels may be exceeded methods of shaft coupling should be reviewed; this would apply to any other auxiliary devices attached to the motor shaft, notably including speed and position transducers the speed rating and energy absorption capability of brakes maintaining acceptable internal stress levels and fits of cooling fans decreased motor efficiency caused by increased losses motor torque capability at increased speed

Bearing currents in inverter-fed motors is a complex subject area. It has received wide publicity and is a practical problem but only in very limited situations. The number of motor bearing failures due to beating currents is very small compared with mostly mechanical reasons for failure. The higher the supply voltage the greater the potential risk. Good system grounding and cabling practice are critical in defining the risk of beating current flow. When a beating failure has occurred and beating currents are suspected, detailed analysis at an experienced tribology laboratory is necessary to identify the cause.

Most standard industrial motors may be capable of operating at speeds above their 50/60 Hz rating; however it is important to have the manufacturer's assurance on the suitability of any motor for operation above base speed. The bearings and type of balancing of the standard rotor dictate a maximum mechanical speed, which cannot be exceeded without endangering the motor or its expected life.

As an example, Table 1.19 gives the maximum speeds which can be tolerated by Leroy-Somer MV motors in horizontal and vertical operation, directly coupled to the load and with no radial or axial loading.


The effects of converter supplies described above are, of course, applicable to motors designed for use in hazardous locations. Traditionally, it has been the responsibility of the

Chapter 1.8


user to ensure that the motor does not overheat as a result of misuse. This has been achieved through the use of devices such as current-sensing protection relays. With converter supplies, the situation is somewhat more complex and it is necessary for the motor manufacturer to assume the responsibility for rating the motor correctly for variablefrequency inverter supplies. Thermistor sensors are mounted at critical points in the motor, and used to monitor motor temperature during operation. Trip relays are used to remove the supply from the motor if any one thermistor reaches the tripping temperature. In North America, NEMA MG 1-1987 17A.04.10 states that motors operated from variable-frequency or variable-voltage

power supply or both, should not be used in division 1 hazardous (classified) locations unless: The motor is identified on the nameplate as acceptable for variable-speed operation when used in division 1 hazardous (classified) locations. The actual operating speed range is not outside the permissible operating speed range marked on the motor nameplate. The actual power supply is consistent with the type of power supply identified in information, which is supplied by the motor manufacturer.


Power Electronics

1 2


51 72



All A.C. and D.C. drives use power semiconductor devices to convert and control electrical power. The devices operate in the switching mode (either on or off) which causes the losses to be reduced and conversion efficiency to be improved compared to operation in linear mode. The practically important power semiconductor devices in relation to motor drives can be considered as follows:

many of them are, in general, complex with integration of many protection features such as overcurrent. For this reason, details of these circuits have been, for the most part, limited to a description of the requirements to gate the devices.

The PN junction diode, Figure 2.1, is the simplest of all semiconductor devices. It may be considered as an electronic switch the conduction state of which depends on the polarity of an externally applied voltage. When a sufficiently high positive voltage is applied to the anode with respect to the cathode, current will flow in a forward direction, the device acting as a closed switch. The forward voltage drop across the device is typically one to two volts. Conversely, when a negative voltage is applied, current flow is prevented and the diode is able to block voltages up to a certain level, VRRM, which is the maximum reverse voltage that can be applied repetitively if breakdown of the PN junction is to be

diode rectifier thyristor (includes phase control, fast and asymmetric types) gate turn-off thyristor (GTO) bipolar junction transistor MOSFET insulated-gate bipolar transistor (IGBT) integrated-gate commutated thyristor (IGCT) other devices

This section reviews the important characteristics of these devices. The electronic gate drive circuits for operating



b anode

forward current
forward volt



applied to turn
d i o d e off

C forward volt drop

diode on

diode off

Figure 2.2 Diode switch o f f and reverse recovery

,i{--- forward conduction

leakage current



~_ reve rse breakdown
reverse blocking

optimised for speed they tend to have higher forward voltage drops which restricts their current rating for a given chip size. Fast-recovery diodes find their main use in free-wheel functions (where they must quickly commutate current from and to primary switching devices) and high-frequency rectification.

The thyristor, or silicon-controlled rectifier (SCR), is a fourlayer PNPN device shown in Figure 2.3. In the off state the device can be considered as three diodes in series, so that current conduction is prevented in either direction. Figure 2.3c shows that the reverse characteristic (cathode positive with respect to anode) is similar to that of the diode; however, the forward characteristic exhibits no current flow other than leakage current until the central control junction, J2, breaks over. Anode current/14 is then able to flow, limited solely by the external load and supply capacity. The forward breakover voltage is equal in magnitude to the reverse voltage, because in the blocking state J1 supports almost all the voltage, junction J3 breaking over at about 10 V. Once breakover in the forward direction occurs, the thyristor behaves rather like a diode which has two junctions (J1 and J3) because the gate P region is neutralised by forward current flow. The overall forward voltage drop is therefore between 1.5 and 2 V. With the thyristor forward biased it is normally tumed on by injecting a positive pulse of current, Io, into the gate, causing J2 to break down (assuming that the device has been forward blocking). Once the anode current has exceeded the latching current, the gate pulse can be removed. Typical waveforms of gate current, anode current and anode cathode voltage during turn on are shown in Figure 2.4. Typical turn-on time for a thyristor is several microseconds, depending on anode current. For the thyristor to remain in the conducting state, the anode current must reach the latching current level, It, and not fall below the holding current, 1/4, IL being greater than 1/4. The thyristor is normally turned off by forcing the anode current to zero by applying a reverse voltage for a minimum period of time before it can regain its forward-blocking state, as shown in Figure 2.4. During the first stage of tum off a reverse current flows because of stored charges for a time, trr, while junctions J1 and J3 recover. This process is similar

Figure 2.1 Diode

a symbol b PN junction c VII curve

prevented. The voltage-current characteristic of the diode is shown in Figure 2.1 c, illustrating the two modes of operation. Unfortunately, the diode does not behave as a perfect switch when it is forced from a conducting state to a blocking state. A reverse current, with peak value Irrm, flOWS during the reverse recovery time t,~, see Figure 2.2. During this time, stored charges, responsible for forward current flow, are removed from the PN junction. The total charge recovered, Qrr, along with lrrm and t,~, forms part of the diode specification; these are important parameters in many applications as they determine the energy loss in the diode each time it switches off. The reverse recovery time may be reduced by careful design of the doping profile of the PN junction and by measures such as doping with particular elements or irradiating the junction with an electron beam. These features are designed to reduce the number of charge carriers in the diode and also to reduce their lifetime so that I , ~ and trr are both reduced. A side effect of this is that the forward voltage drop increases so there is a trade off between it and speed. Generally, a diode used for low-frequency rectification of A.C. to D.C. power has a long &r, and high Qrr as it has been optimised for minimum forward voltage drop. Diodes of this type are available in ratings as high as 8000 V and 6000 A. Diodes with fast characteristics, that is short trr and low Q~r, are referred to as fast-recovery diodes. As these have been

Chapter 2.1
b anode A J1 gate G J2 J3




l 1


forward on-state volt drop


values of V when thyristor gated

holding current reverse leakage current


latching current

_ )~)
forward leakage current

) J
forward breakdown


reverse breakdown
Figure 2.3

a thyristor symbol b thyristor structure c thyristor characteristic

stffte J



r e c ~ t q

_r "v''"
J. off

/ reapplied

off turn I on j.
on 4 Figure 2.4 Thyristor switching waveforms a gate current


turn off

b anode current c anode voltage



to a diode turning off. The control junction J2 needs an additional time to recover, called the recombination time, tqr, and only then can a forward voltage be reapplied at a maximum specified rate. The total turn-off time, tq is an important parameter for thyristors in fast switching applications.

minimum levels, dependent on junction temperature, which lie between the upper and lower resistance limits shown in Figure 2.5a. It is also necessary to ensure that the peak gate power (Vc Ic) is not exceeded. Figure 2.5b shows a typical gating characteristic, illustrating the boundary conditions. A simplified example of a pulse-transformer-based firing circuit is shown in Figure 2.6. Resistor R1 limits the gate current while R2 provides a low impedance across the gate to attenuate any gate voltage when the thyristor is in the off state. To achieve short turn-on times, the gate current is required to rise at a minimum of 1 A/Its. A succession of gate pulses, Figure 2.6b, supplied by the gate drive circuit, causes firing to occur when external conditions are suitable for conduction.

Thyristor Gating Requirements

The gate cathode characteristic of a thyristor resembles that of a poor PN junction and will vary between production batches for a given type. To be certain of turning on the thyristor, the gate current and voltage must attain

VG~ /

gate power boundary

characteristicsof individual .,. ~ \ thyristors ~ ~,\~,"~ \ of same type

Power Losses and Current Ratings

In normal operation the thyristor dissipates power in the form of heat resulting mainly from: forward conduction loss, which is a function of on-state voltage and forward current switching loss, which is energy dissipated during turn on and turn off blocking leakage loss, which is a function of the offstate forward or reverse leakage current and blocking voltage


+25C I / / / J I ~ minimumlimit to ensure firingat given Tj



The heat generated must be removed by a cooling system in order that the maximum junction temperature of the device is not exceeded (usually 125C for a thyristor). The following equation describes the relationship between junction temperature Tj and power dissipation PD for any semiconductor:

Tj = PD Rthjc + Tcase
where Rth/c is the thermal resistance, junction-to-case, and Tcase is the case temperature of the device. Clearly, the more efficient the cooling system, the greater the power that can be dissipated for a given case temperature, which leads to greater current capability.


0.1 10-3 10-2 10-1 100 101 A 102



Gate characteristic and peak gate power dissipation

Parameter A B C D Pulse duration tp ms 10 1 0.5 0.1 Maximum allowable peak gate power W 40 80 100 150
Figure 2.5 Thyristor gate curves

b 1A

Figure 2.6 Thyristor gate circuit

a range of characteristics and limits b example characteristic

a simplified firing circuit b 20 kHz gate current pulse train

C h a p t e r 2.1


There is a finite limit, however, expressed by the r.m.s. current rating. This ensures that excessive heating of internal joints and bonding wires is prevented. Thyristors are therefore given an average current rating based on practical case temperatures for a defined waveform, and an r.m.s, rating, both of which must not be exceeded. Manufacturers also provide graphs of average power versus average current, Figure 2.7a, and allowable case temperature versus average current for various waveforms, Figure 2.7b. From the graph in Figure 2.7a it can be seen that average current decreases with duty cycle, the lines terminating at the point when the maximum r.m.s, current is reached. The graph in Figure 2.7b shows that the maximum junction temperature is 130C and all the curves converge to this point as the current reduces. This type of information is essential for the selection of thyristor size and for the design of the heat sink.

Surge Current Ratings

It is possible for the junction temperature to be exceeded for short periods of time under fault or overload conditions. The thyristor becomes predominantly resistive and it can be shown that the temperature rise is proportional to:

IZ x tp
where tp is the pulse duration and I is the r.m.s, value of the pulse current calculated over time tp. The equation assumes that all the heat generated is not dissipated but stored in the mass of silicon. An IZt rating can be specified and is a useful parameter in determining the size of fuse for overload protection. The thyristor is also given a nonrepetitive surge rating, half sine wave in shape, for 10 ms and is typically equal to about ten times the r.m.s. current rating.

400 D.C. 180sin 120 J-I. 60 J-l. 30 Yl.



E < ~ lOO







temperature, C (Tc)

b l

400 RthKAK/W 0.1

~ d.c.

0.2 0.3 0.4 0.6 0.8 1.0 in






A 0







current, A(ITAvMIIFAvM)

temperature, C (TA)

Figure 2.7 Thyristor on state current and loss a average on-state p o w e r dissipation for rectangular current w a v e f o r m b m a x i m u m allowable case temperature Tcasefor rectangular current w a v e f o r m



The conventional thyristor, which is turned off by application of reverse voltage, has a structure which can be altered to have characteristics to suit specific applications. Three types of thyristor can be distinguished: (i) Thyristor for A.C. line commutation (phase control thyristor) - thyristors used in A.C. applications are turned off, or commutated, naturally by the existence of the A.C. supply, which changes polarity in alternate half cycles. The thyristor is designed to have a low on-state voltage, thereby maximising current rating at the expense of relatively long turn-off times (typically 100 to 200 Its). This does not matter because the thyristor switching frequency is low. Equal forward and reverse voltage up to 12 000 V are possible for large phase-control thyristors. For applications onA.C, supplies up to 500 V A.C., it is usual to specify 1400 V types, to allow for an overload factor of two. It is common practice to use RC networks and varistors across the thyristor to give additional protection. (ii) Fast thyristors - these devices are generally used in D.C. circuits such as choppers or inverters, although their use is now less frequent as more modern devices such as IGBTs have replaced them in many applications. Within a D.C. circuit there is no natural reversal of the supply for thyristor commutation, therefore it must be derived by external circuits. The process oftum off under these conditions is called forced commutation in contrast to line commutation. Typical commutating circuits are expensive because they consist of inductors, capacitors and auxiliary thyristors; however their size can be reduced if tq, the total turn-off time, is kept to a minimum. The design of the thyristor is therefore optimised for low tq (typical values, 15 to 30 Its) but unfortunately this has the undesirable effect of increasing the on-state voltage drop, which consequently lowers the current rating. (iii) Asymmetric thyristor - in many fast-switching applications the reverse blocking capability of the thyristor is not required because an antiparallel diode must be connected across the device for reactive current conduction. Manufacturers have exploited this relaxation by offering the asymmetric thyristor or ASCR, which has even lower tq times than the fast thyristor but at the expense of very limited reverse blocking. Turn-off times as low as 8 Its are possible while still retaining an acceptable 15 V reverse blocking. Another technique for enhancing the performance of both ASCRs and fast thyristors is to use an interdigitated gate structure which considerably increases the device di/dt rating at turn on. This technique effectively enlarges the turn-on area of silicon available at the start of gate firing, thus preventing excessive current density near the gate which could lead to device failure.

forced commutation. High switching frequencies coupled with high rate of rise of on-state current cause the turn-on losses to reach very significant levels: in some situations the device does not fully turn on owing to the limited current spreading over the chip, demanding stringent current derating. Resonant load circuits, such as those used in induction heating, overcome the problems of excessive switching loss by switching on at the point where the load passes through either zero current or zero voltage. This technique allows fast thyristors to be used at up to 50 kHz. Furthermore, the load resonant circuit forms part of the forced commutation circuit and in some circuit topologies additional auxiliary thyristors are not required for commutation as this function is performed by the main thyristors.


The gate turn-off thyristor (GTO), like the conventional thyristor, can be latched into conduction by a short positive gate signal but, unlike the thyristor, the GTO can revert to the forward blocking state by the application of a negative gate signal. The GTO can therefore replace the fast thyristor and its associated commutation circuits in D.C. switching applications. The circuit symbol and the more complex GTO structure are shown in Figure 2.8. It is still a four-layer device with similar voltage blocking capabilities to those of the thyristor. The gate region is highly interdigitated with the cathode, producing a patterned structure which is designed to give high current turn-off capability by preventing conduction continuing in the cathode islands between the gate contacts. This also ensures even current density across the die during turn off. Although the physical operation of the GTO is very complex, it is helpful to refer to the twotransistor model of the GTO, Figure 2.8c, to understand how turn off is achieved. The devices may be considered as two interconnected transistors which have regenerative action: the collector current of one feeds the base current for the other transistor. It can be shown that (neglecting leakage currents) the anode current IA is given by:


I G OLNpN 1 -- (OgupN -Jr- Olpup)

where Ia is the gate current, OlNp N and OLpNP are the common base gains, where a =/3/(1 +/3) and/3 is transistor current gain, Ic/I8. The current gains are dependent on the collector current and increase as the current increases from zero. Conduction of the thyristor is initiated by a gate current pulse which raises the loop gain (O~NPN--[-OZpNp) unity so to that from the above equation IA is infinite. In practice, the anode current is limited by the load. Tum-off action is produced by extracting sufficient current from the gate to cause the loop gain to fall to a point where regenerative action ceases. The turn-off gain ~ o f f is the ratio of anode current being controlled to negative gate current required to produce turn off, and is an important parameter. Typical values of ~off lie between three and five. To reduce the loop gain, and hence increase the tum-off gain, the gate is often connected to the cathode with a low resistance within the package, emitter shorts, which has the side effect of making the gate less sensitive for tum on. Unfortunately, there is a limit to the maximum anode current that can be

High-Frequency Current Operation

Although the ASCR and fast thyristor have been designed for high-speed operation, the maximum frequency that can be switched in practice is limited to approximately 1-2 kHz with

Chapter 2.1


anode A

gate cathode C interdigitated gate

current flowing at the time. The anode current then falls rapidly, in several microseconds, as the device recovers its blocking capability. After the bulk of anode current has been commutated, a small tail current flows due to trapped charges in the base N region, adding to the turn-off loss. After turn off it is normal to reverse bias the gate cathode with a 15 V supply to prevent retriggering of the device. During switch on and switch off considerable power is dissipated as heat, limiting practical GTO switching frequencies to around 2 kHz.

st ru ~//u,l~/lilll,sct / re

0 A

Snubber Design
(base region)
~lV p



The GTO must have a snubber circuit connected across it to limit the rate of rise of voltage at turn off. An example is shown in Figure 2.10. Unlike the thyristor, forward voltage is reapplied immediately after turn off (compare Figure 2.4c) which means that the dv/dt limit usually has to be lower, necessitating a larger and more lossy snubber. It is also important that the voltage spike, marked Vs in Figure 2.9a, which is due to snubber circuit inductance, is minimised through good layout. In fact, manufacturers stipulate that the maximum controllable anode current cannot be guaranteed unless a certain size of snubber capacitor is used and Vs does not exceed a specific level.

Voltage and Current Ratings

Like the thyristor, the GTO can be designed to have equal forward and reverse blocking capabilities, or limited reverse blocking with the advantage of improved turn-off times. The latter type is sometimes referred to as an anode-short GTO and is designed to be used with an antiparallel diode. The forward on-state voltage drop is of the order of 2 V, giving current ratings similar to those for the fast thyristor.

I~T 2
c2 ~

~PNP =

O~NPN=~E 1

,4 1 ` ~IE

Figure 2.8 The gate turn-off thyristor

a circuit symbol

b structure
c two-transistor model of the GTO

The most common type of power bipolar transistor, sometimes referred to as the bipolar junction transistor or BJT, is a three-layer NPN device as shown in Figure 2.11. PNP types are available but they tend to have inferior voltage and current ratings. To conduct a collector current Ic, the transistor must be supplied with a continuous base current 18, depending upon the voltage level between collector and emitter. This relationship is shown in Figure 2.12. The ratio Ic/18 is called the current gain hFE and may be less than l0 for a 1000 V transistor. The gain can be greatly improved if the base current is obtained from another transistor using the Darlington connection shown in Figure 2.13. The three transistor stages are integrated on the same silicon chip giving an overall gain of several hundred. A typical curve of gain versus Ic for a 50 A 1000 V threestage (Mitsubishi QM50DY-2H) Darlington transistor is shown in Figure 2.14. This illustrates the dependence of gain on collector current. In practice, for motor drive applications the power transistor is always operated as a switch. When closed sufficient base current is provided to ensure that the transistor operates in the saturated or quasisaturated mode.

switched off; generally this is about four times the average current.

Switching Characteristics and Gate Drive

Example anode-current, anode-voltage and gate-current waveforms are shown in Figure 2.9. Turn on is initiated by a relatively high amplitude gate pulse and takes about 3 to 5 las. Despite the highly interdigitated gate structure there is a practical limit to the rate of rise of anode current: therefore external inductors are sometimes used to limit di/dt and turnon losses. After turn on, the gate current may fall to a lower level, the 'back-porch' current, sufficient to minimise the on-state voltage drop. If a continuous gate drive were not provided, the volt drop would tend to increase at low anode currents owing to the fall in loop gain. To turn the GTO off, the forward gate current is removed and a negative voltage source ( - 1 2 to - 1 5 V) is applied. The resultant negative gate current increases rapidly until large enough to stop GTO regenerative action. This may take 20 rts (the storage time ts) depending on the amount of anode



anode current anode voltage



IG ...J=.~-,,,.~~ gatecurrent

diGQ/dt IGQ Figure 2.9 GTO gate waveforms

a collector C



base B


emitter E C

Figure 2.10 Typical GTO snubber
Typically, VcE is 2 to 3 V for a Darlington operating in the saturated region. When open a small negative bias is applied to the base to ensure minimum leakage current.

Ic Figure 2.11 Bipolar transistor
a structure b circuit symbol with current directions

Voltage Ratings
The transistor will break down if a sufficiently high collector emitter voltage is applied, causing permanent damage. To prevent this, various voltage ratings are defined: the two most important are:

1 VcEv- the maximum voltage between collector and

emitter with the base reverse biased (usually - 2 V). This is the highest rating and is used to classify the device. VcEv ratings up to 1400 V are available.

Current Ratings
The continuous or D.C. current rating is the principal rating used for classification. Although quoted at 25C case temperature it is usually possible to achieve this rating at 100C without exceeding the maximum junction temperature, tjmax, assuming adequate cooling. The transistor can also be pulsed with twice the D.C. rating or ICM for 1 ms at 50 per cent duty cycle provided that peak collector power and tj'ma x a r e not

VcEo- the maximum collector-emitter voltage with the base open circuited. This rating tends to be the lowest rating, being 50 to 90 per cent of VcEv. It is more important than VcEv and indicates how rugged the switching capability of the transistor is.

Chapter 2.1
continuous collector power dissipation


ex_ 0 0 0

s SI
I! s S

saturation voltage

I %%%

-at fixed values of Ie

I Ii

leakage current ~

c o l l e c t o r emitter v o l t a g e


breakdown ,,-lOV __~ ~,l~ saturated region linear region

lforward breakdown voltage

Figure 2.12 Characteristic VCE curve


VcE=5.0 v

E ~

s s

I s

4 3


Vce- 2.8 V




d 102

i I I i I



7101 2 3 45

~. =125C , 2 3 45 7102

Figure 2.13 Three-stage Darlington arrangement


collector current I o A

exceeded. Ratings up to 1000 A with VcEv equal to 1200 V are available in a single module.

Figure 2.14 D.C. current gain curve


C h a r a c t e r i s t i c a n d Base D r i v e
reverse current IB2 the level of which is controlled by the base drive circuit flows out of the base for a period called the storage time during which stored charges in the base region are removed. Once this is complete, the collector voltage rises and then the collector current falls in 0.5 to 2.0 Hs. Typical storage and fall times are 10 ps and 2 ItS, respectively. The storage time may be reduced by operating the transistor in a quasisaturated state. This is achieved by controlling the base current so that only just enough current is supplied to hold VcE at a value just above the saturated value. This may be achieved using a relatively simple diode clamp circuit known as a Baker clamp. As the collector current falls any stray inductance present in the power circuit

Typical operating waveforms are shown in Figure 2.15 for the power transistor switching an inductive load clamped by a free-wheel diode - a situation encountered in chopper circuits and voltage source PWM inverters. When the transistor turns on (ton typically 0.5 to 2 ps) the collector current will see a peak current higher than the load current due to the reverse recovery current of the free-wheel diode. The peak current must be less than IcM. Forward base current I81 is then maintained to keep the transistor at or near saturation level. Turn off is initiated by quickly removing the forward base current and applying a reverse voltage to the base emitter. A



will cause the collector emitter voltage to overshoot as shown in Figure 2.15, making snubber circuits necessary. The total power loss in the transistor will comprise switchon, switch-off and conduction losses. Switching losses are almost double at 125C junction temperatures compared with those at 25C because of the temperature dependence of the switching times, but the conduction loss remains nearly constant. Typical calculations for power and switching losses are shown in Table 2.1. Generally, switching frequency is limited to between 2 and 5 kHz, switching loss and storage time being the determining factors.

Safe Operating Areas

During turn on and turn off, the instantaneous collector current and voltage are both very high for short periods of time, hence the switching losses. This is permissible provided that the locus of Ic and Vce lies within specified safe operating curves defined by the manufacturer. When the transistor is forward biased, for example during turn on, the forward-biased safe operating area (FBSOA) curve is applicable. For very short pulses it has a current boundary limited by lcu and a voltage boundary limited by the VcEo rating as shown in Figure. 2.16a. For longer pulses the

1 I

.L -C

+~,~ T Vcc

o peak current due to reverse recovery of freewheel diode ve~h;ot

ts - - ~ %~ ~--




turn-on A
loss " ~ ' ~ A conduction loss

turn-off loss

Figure 2.15 Bipolar transistor half-bridge switching an inductive load a test circuit b collector current and voltage waveforms c base drive d power loss over one switching cycle




Table 2.1 Calculation of switching and conduction losses for a bipolar transistor

forbidden area


Energy dissipated during turn on Energy dissipated during turn off Energy dissipated during conduction period Switching loss (watts) Conduction loss (watts) Total average power loss

Wo, ~ 0.5 x t~c x Vcc x Ic Woj:f ~ 0.5 x tic x Vcc x Ic Wcond ~ tcond X VCEsat X I c

(2 xlc) forward bias safe operating area


esw :fsw

(Won "JI-moff)

E)3 (-

PcoM = f~w x Wco,,a Ptotal = P~w + PcoM

maximum power is limited by the junction temperature thermal limit and a complex phenomenon called secondary breakdown. Secondary breakdown can be broadly described as excessive localised heating in small areas of the chip where current density is high due to uneven current distribution. At turn off the transistor is reverse biased, therefore the reverse-biased safe operating area (RBSOA) curve is applicable, Figure 2.16b. This curve is defined by the maximum voltage and current ratings and secondary breakdown effects due to uneven current distribution during turn off.

0 (cO 0

= Ic

Tj ~< ~2SC
tp defined
(typically <1 ~s)




reverse bias safe operating area repetitive

(/) :3

Short-Circuit P e r f o r m a n c e
If the transistor load is accidentally short circuited the collector current will invariably exceed its allowable peak rating pulling the transistor out of saturation. The device can survive nonrepetitive surges of this nature provided that:

:3 t.m .i.-, tO o v

o tc
Ie2 defined Tj ~< 125 C

a surge is detected and interrupted within a specified time VCE and I c do not exceed specified levels

(guide only) 160 _Tj'- 1:)5~_ 140 I < G 120 E 100
o o


Most manufacturers guarantee that the device will survive a limited number of short circuits under defined test conditions. A typical transistor may be rated for a peak current of four to six times rated current, with VCE of half-rated voltage and a short-circuit duration of 30 ~ts. This type of data is essential for the design of practical protection schemes required by chopper and inverter applications.

I I I I IB2/-~ A/

/82=-3 A

80 60

,'~" A



The metal oxide semiconductor field-effect transistor (MOSFET) was developed into a useful fast-switching power device in the 1980s. There are both N-channel and P-channel devices available, but for power applications N-channel devices predominate due to their lower losses. The device symbol for an N-channel MOSFET is shown in Figure 2.17. When a voltage, VGs, is applied between the gate and source an electric field is set up within the device. The field modulates the resistance between the drain and source, permitting a current to flow in the drain in response to the applied drain circuit voltage. The transfer characteristic is shown in Figure 2.18. With zero VGS applied, a positive voltage will be blocked at the drain until the breakdown limit is reached. Any reverse voltage applied will be clamped by the presence of a parasitic diode in the device structure. If VGs is raised no drain

200 400 600 800 1000 collector-emitter voltage VCE,V

Figure 2.16 Safe operating areas for bipolar transistors

a f o r w a r d - b i a s safe o p e r a t i n g area (FBSOA) b reverse-bias safe o p e r a t i n g area (RBSOA) c typical RBSOA

current will flow until the gate threshold voltage, typically 2 to 3 V, is reached. With Vcs above the threshold voltage the characteristic has two distinct regions: a constant resistance region with channel resistance Ros:o,), and a constant current region where the transconductance of the device is almost constant. RDs:o,) is a key parameter and will determine the forward voltage drop (Ros:o,) I9) and ultimately the current rating of the device. Operation within the constant current region is normally avoided (to minimise conduction losses) by setting Vcs high enough for the load current, a value of 10 V is usually sufficient. With VGs above threshold



,1 I I

'r - - I gate

. - 1, - internal
t I




sourcet - - - J
p o s i t i v e voltage at g a t e produces current in drain



In conventional high-voltage MOSFETs (>200 V) the lower doping density and thicker die required for higher voltages has resulted in devices which have an on resistance pro2.5 portional to V{)ss, where VDSS is the voltage rating. This is why there are relatively few devices rated above 600 V, and devices rated above 1000 V are very rare. A process technology, marketed as CoolMOS by Infineon, uses a new three-dimensional doping profile which allows the doping density to be increased and chip thickness reduced compared to conventional MOSFETs so that the on resistance only increases linearly with voltage rating. This allows a smaller die to be used for a given rating although the more complicated chip fabrication increases costs. For low-voltage MOSFETs (< 50 V) the channel resistance is significant, as opposed to high-voltage devices where the body or drift region dominates. To reduce the channel resistance another process technology has been used: trenchgate structure. A trench-gate device, as its name suggests, has the gate etched down into the chip, rather than being a planar feature near the surface of the chip. This reduces the channel length and so resistance, and also allows the cell size to be reduced thus giving more cells per unit area and hence reduced resistance per unit area.

+ T
Figure 2.17 MOSFET
a circuit symbol b electrical circuit

Switching Performance
VGS=20 V (maximum)
constant current

VGS=10 V

VGS= 6 V VGS=3 V

drain-source voltage, VDS parasitic diode volt drop


The MOSFET has two very important differences to the BJT; first it is a voltage-controlled device rather than being a current-controlled device, and second it is a majority carrier device rather than a minority carrier device. Being voltage controlled with an oxide insulated gate, very little power is needed to control the device, current only being required to charge and discharge the gate capacitance. A majority cartier device only conducts due to intrinsic charge carriers. This enables the device to switch off very quickly because no time is required for the removal and recombination of minority charge carriers, as there is in a bipolar transistor or thyristor. Fast switching allows the MOSFET to be used in applications with switching frequencies of 100 kHz or more. The down side of only having intrinsic charge carriers available for conduction is that the conductivity of the silicon is lower and so a larger chip is required for a given current rating.

Figure 2.18 Power MOSFET characteristic

voltage the channel is able to conduct current in the reverse direction as well as the normal forward direction. This feature can be exploited by using MOSFETs as very-lowvoltage drop diodes.

Safe Operating Area

The MOSFET does not exhibit the phenomenon of secondary breakdown, which means that the safe operating area curve for all operating modes is square. It extends to VDSS along the voltage axis and up to four times ID in the current axis. The only limitation to switching at these levels is that the maximum junction temperature must not be exceeded.

Voltage and Current Ratings

Current ratings are usually given for case temperatures of 25C; at the more practical temperature of 100C the current rating is reduced because of an increase in RDs(on). The variation in RDS(onjwith temperature depends on the voltage rating of the device; lower voltage devices have a lower temperature dependence, typically a factor of 1.2 from 25C to 125C; for high-voltage devices the ratio may be two or more. The positive temperature coefficient of on resistance has the benefit that devices may be connected in parallel and will share current equally. A chip is in fact made up of thousands of MOSFET cells connected in parallel.

Parasitic Diode
The parasitic or body-drain diode that exists within the MOSFET structure has a slow switching characteristic compared with the MOSFET channel itself. The switching frequency of circuits which make use of the diode (for example ultrasonic inverters) may be limited solely by the diode and not the MOSFET. This is mainly due to very significant diode switching losses, which are a function of reverse recovery charge, and operating frequency. In some

Chapter 2.1



low on-state 5 voltage diode ,

. . . . n

gate O
y / ~ - v eor d efast di emitter

-_~"qL I I I .i

__ parasitic diode




Figure 2.19 MOSFETwith antiparallel diode


applications the body diode is actually bypassed by a fast diode as shown in Figure 2.19; the series-connected diode prevents current from flowing in the body diode.


Figure 2.20 IGBT



The insulated-gate bipolar transistor (IGBT) combines the best features of a MOSFET and a BJT to give a voltage-controlled device with low state losses. The circuit symbol and terminal designations are shown in Figure 2.20a.

a circuit symbol b steady-state equivalent circuit

The on resistance of a high-voltage MOSFET is mostly due to the drain drift region as explained in the previous section. In an IGBT the resistivity of this region is substantially reduced by the injection of charge carriers from an additional semiconductor layer in the device. The process is called conductivity modulation, and results in on-state losses comparable to those of the BJT. An IGBT may be modelled as a low-gain PNP transistor and MOSFET connected as shown in Figure 2.20b, the additional semiconductor layer being the emitter of the PNP transistor, which is also the collector of the IGBT. The on-state voltage VCE(saO of the IGBT has three components:

collector doping density than do NPT types. NPT devices generally have faster temperature independent switching and a positive temperature coefficient of VcE(saO, and PT devices have a strong temperature dependence of switching loss and a negative temperature coefficient of VCE(saO. Unlike the MOSFET, the IGBT does not have a parasitic diode inherent within its structure, so that a suitable antiparallel fast diode may be selected to match the speed of the IGBT.

Voltage and Current Ratings

The principal voltage rating of the IGBT is the collectoremitter breakdown voltage VcEs specified with zero gate emitter voltage. Devices are available with ratings from 250 V up to 6500 V. Although devices are available with terminal current ratings up to 2400 A, these devices use many IGBT chips connected in parallel. Single chips typically have a maximum rating of 300 A.

VCE(sat) -- VBE + (Iz) X RDeaFr) + (Iv x RcI-I)

where Vee is the forward voltage drop of the PNP transistor, is the drift region resistance and is much smaller than in an equivalent MOSFET due to the conductivity modulation, and Rci-i is the MOSFET channel resistance. Typical values for VCE(saO a r e 2 to 3 V at rated current and 25C for a 1200 V rated device. Trench-gate IGBTs have been developed recently which with an optimised geometry reduce both RDRIFr and Rci4 giving a lower VCE(saOfor a given current density. There are two basic types of IGBT: punch through (PT) and nonpunch through (NPT). The difference relates to the doping and thickness of the drift region and the collector layer, PT types having a thinner drift region and higher

Switching Behaviour and Gate Drive

Like the MOSFET, the IGBT is voltage controlled and the gate presents a capacitive load to the drive circuit. Turn on takes place when the capacitance has been charged to above the gate-emitter threshold voltage, which is usually 4 to 5 V. Typically a 1200 V 100 A IGBT will turn on in less than 200 ns. During the conduction time, the gate emitter voltage is held at between 13 and 17 V so that VCE(saO is kept as low as possible to minimise conduction losses. Turn off is initiated by discharging the gate emitter capacitance. The MOSFET structure turns off first, allowing Ic to fall rapidly to an intermediate level. A slower fall of current then follows as the PNP structure turns off. Unfortunately, the gate drive circuit can only control the MOSFET turn off



POWER SEMICONDUCTORDEVICES: Insulated-Gate Bipolar Transistor

and has no influence on PNP behaviour. Turn-off delay time and current fall time are much shorter than for equivalent bipolar transistors owing to the low gain of the PNP structure and processing steps taken to reduce the carrier lifetime. Both these measures have the effect of increasing Vce(saOso there is a trade off between low switching loss and low conduction loss. To give optimum performance in low and high-frequency applications IGBTs are available in different families, optimised for either low conduction or low switching loss. Both turn-on and turn-off times can be adjusted by selection of gate resistor value; this controls the rate at which the input capacitance is charged or discharged. This is a very important feature as it allows turn on to be set at a rate which suits the reverse recovery characteristic of the free-wheeling diode, and allows the rate of turn off to be reduced if required to limit inductive voltage overshoot. During the off state the IGBT gate emitter is normally held at a minimum o f - 5 V to ensure that the device cannot be spuriously tumed on.

a short period of very high dissipation without damage. A typical device will be rated for a short-circuit duration of 10 ~ts with test conditions of a current of ten times ID, with a collector-emitter voltage of half rated voltage, and 125C. During this fault condition the junction temperature may exceed 300C. The control and gate drive circuit must detect the overcurrent condition and switch off the IGBT within the rated time to avoid damage. The short-circuit current may be controlled by adjusting the gate voltage. For the Eupec DN2 series devices a Vce between 10 and 17 V will produce short-circuit currents of between two and ten times ID. To reduce the inductive voltage overshoot when turning off the large short-circuit current, the gate drive circuit may slow down the turn off under these conditions by limiting the gate discharge current.

Series and Parallel Operation

Although devices are available up to 6500V, the more common value is 3300 V. For some applications this is not sufficient and so devices must be connected in series. This is easier to achieve with IGBTs than with many other power devices due to the following combination of characteristics: voltage control, fast switching, square SOA. Figure 2.22 shows how two IGBTs may be connected in series. The resistors maintain steady-state voltage balance by compensating for differences in device leakage currents. Very fast transients are balanced by the capacitors, which can be much smaller than for a similar circuit using BJTs or thyristors. During switching, differences in delay from one device to another will tend to lead to unbalanced voltages. The active clamp circuit, formed by the zener string and IGBT, limits VcE to a value a little above the zener breakdown voltage. The voltage-controlled gate means that little current is needed and fast switching ensures a fast response. The IGBT is able to survive this operation because of the square SOA. Devices may be connected in parallel to make up power switches with ratings of many kA. When using devices with a positive temperature coefficient little or no derating is required. When using devices with a negative temperature coefficient it is usual to select devices with the same Vce(saO and apply a derating factor.

Safe Operating Area (SOA)

The FBSOA and RBSOA curves for an IGBT chip are both square, and bounded by the rated breakdown voltage, and the pulsed collector current, IDM, usually twice the rated D.C. current ID. The square SOA makes the IGBT a very robust device and allows operation without snubber circuits, thus reducing system losses and size. Although the chip RBSOA is square the device RBSOA may have some reduction in VcE at high currents. This is due to wiring inductance within the package, which during turn off increases the voltage at the chip above that measured at the terminals. Modern IGBTs do not suffer from secondary breakdown unlike BJTs. The RBSOA for an Eupec BSM100GD120DLC is shown in Figure 2.21.

Short-Circuit Performance
Most IGBTs produced for high-power applications are shortcircuit rated. This means that the device is able to withstand

240 200 160 <120

80 40 0


IC, module IC, chip

6c o

1 oo



Figure 2.21 Reverse-bias safe operating area for Eupec BSM100GD120DLC

C h a p t e r 2.1


2000 ance resistor A.C. balance capacitor 1600

VDM <~ VDR M rj = o - ~ s o c

< v O



Figure 2.22 Series connection IGBT with active clamp

0 1()00 2000 3()00 4000

vo (v)

Figure 2.24 IGCTSOA

oate t I'/1


Figure 2.23 IGCTsymbol


The integrated-gate commutated thyristor (IGCT) is a development of the GTO and has essentially the same device structure. By operating with a turn-off gain of 1, rather than 3 to 5 for a GTO, faster turn off is achieved and the requirements for snubbers much reduced. In order to have the correct conditions for turn off the gate circuit inductance must be very low. To achieve this the gate circuit is integrated into the device - hence the name integrated-gate commutated thyristor. IGCTs are available with or without an antiparallel diode. As with a GTO, the IGCT is a four-layer device with a highly interdigitated gate structure. The device can be triggered into the on state by applying a positive gate current allowing current to flow between the anode and cathode. Once switched on the current is determined only by the external circuit. The on-state voltage is typically 3 V for a 4500 V rated device. Turn off may be achieved by reverse biasing the main power circuit, or more normally by extracting current from the gate. Sufficient current is extracted from the gate so that no current flows across the gate to the cathode junction, and the device behaves as a PNP transistor. This gives fast turn offwith a safe operating area (SOA) similar to that for a bipolar transistor.

This voltage is greater than the maximum voltage which can be sustained during switch off. The on-state average and r.m.s, current ratings, ITAVE and IrnMs, determine the maximum load current. In the on state the device can withstand high surge currents, but the device cannot switch off these currents. The maximum current, which can be turned off under defined conditions, is IrGQM. Devices are available with ratings up to 4000 A. The maximum controllable current is approximately twice the on-state current.

Switching Behaviour and Gate Drive

Turn on is initiated by application of a positive gate current typically up to 100 amps. Lower gate currents may be used but this increases the turn-on time and limits di/dt capability. Once conduction is initiated the gate current may be reduced to the back-porch current of a few amps. As with a GTO this is necessary due to the low device loop gain at low anode currents. At turn on the anode current di/dt must be limited to prevent hot spots as the conduction spreads out from the gate, and also in applications with a free-wheel diode to limit the reverse recovery di/dt. However, once turn on is initiated through the gate, the anode current di/dt cannot be controlled via the gate, and so an inductor and associated snubber must be added. In order to turn the device off sufficient gate current must be extracted from the device for the regenerative action to stop. To allow operation with little or no snubber and minimise switch-off losses a turn-off gain of 1 is used, (i.e. all anode current is diverted out of the gate), and this must be accomplished fast enough to prevent current redistribution in the device which would lead to hot spots and a reduction in device turn-off rating. In a standard GTO package the gate inductance is of the order of 50nil. To divert an anode current of 2000 A out ofthe gate in 1 ~ts would require 100 V. This would lead to very high gate drive power losses, and

Voltage and Current Ratings

The maximum repetitive voltage, which can be blocked, is represented by VDR~, with devices available up to 5500 V.


POWER SEMICONDUCTORDEVICES: Integrated-Gate C o m m u t a t e d Thyristor

also exceeds the gate reverse breakdown voltage of approximately 20 V. By integrating the gate drive unit into the IGCT assembly a gate circuit inductance of 5 nH is achieved. Now only 10 V is required to achieve the required gate current di/dt, gate drive power is reduced and the breakdown voltage is not exceeded. To switch the large gate current required an array of lowvoltage MOSFETs is used coupled to a bank of low-impedance electrolytic capacitors. Gate drive power consumption can be significant, approximately 80 W for a 700 A device switching at 500 Hz. As with a GTO, an IGCT has a minimum on and off time. The minimum times of 10 gs are determined by a combination of time required for current density across the chip to stabilise and for the gate drive circuit to prepare for the next switching event. A typical maximum average switching frequency is 500 Hz, although this may be increased to 2 kHz with a suitable gate unit and some reduction in current rating. In the event of a short-circuit fault the device must be switched off before the anode current rises above the maximum controllable current or control will be lost and the device destroyed.

anode gate %


gate &

on FET
off FET


The MOS-controlled thyristor (MCT) was developed to exploit the low conduction loss of a thyristor with the low gate drive power and fully controlled behaviour of a MOSFET. The symbol and simplified equivalent circuit of an MCT is shown in Figure 2.25. Operation is most easily understood from the equivalent circuit. In the off state the turn-off MOSFET is held on, this keeps the PNP device off and the MCT can block positive anode cathode voltage. To turn on, the turn-off FET is switched off and the turn-on FET switched on. This provides base current to the NPN transistor and regenerative action then takes place as with a standard thyristor. Once in the on state the MCT has a similar surge capability to a thyristor, and low voltage drop of 1 to 2 V. To turn off the MCT the turn-off FET is again turned on. This short circuits the PNP transistor base and so regenerative action stops and the device turns off in a manner very similar to that for a GTO. Although the MCT offers low on-state loss and low gate drive power, there are important limitations which have prevented widespread use of the device; the gate bias must be maintained at all times to ensure that the device remains off, and the switching safe operating area is limited to half rated voltage at rated current. Devices have been made with maximum ratings of approximately 1500 V, 100 A.
on gateC) cathode

Figure 2.25 MCT symbol and circuit


on gate(~


off gate


off FET

MOS Turn-off Thyristor

The MOS tum-off thyristor (MTO) is very similar in many respects to the MCT and IGCT. The circuit symbol and equivalent circuit are shown in Figure 2.26. In contrast to other power devices the MTO is a four-terminal device; in

off gate ( ~


Figure 2.26 MTO symbol and equivalent circuit

Chapter 2.1


addition to the main power terminals there are two gate terminals, one for turn on and another for turn off. As can be seen from the equivalent circuit, the MTO comprises a GTO, which is responsible for turn on and conduction, and a MOSFET, which is only used during turn off. To turn on, a gate current of several tens of amps is injected into the turn-on gate. Once conduction is initiated regenerative action starts and the anode current is limited only by the external circuit. As with GTOs and IGCTs, a back-porch current of several amps is required to ensure that the device stays on with minimum voltage drop. To turn off, the backporch current is switched off and a positive voltage applied to the turn-off gate terminal. This switches on the MOSFET which diverts current out of the gate forcing regenerative action to stop and the GTO to turn off. As with an IGCT, sufficient gate current is removed to give a turn-off gain of near unity, which gives much faster turn off than for a GTO operating at a turn-off gain of 3 to 5. As the turn-off MOSFET is integrated into the MTO, the inductance is low enough to ensure that the gate current can be removed fast enough to prevent current redistribution during turn off. Devices have been made with ratings up to 4500 V, 500 A with plans for ratings up to those of conventional GTOs.

Pressure Contact Packages

Pressure contact packages are in general only used for very high-power applications, and devices rated above 4.5 kV or 2.5 kA are in general only available in these packages. They are well suited to devices with a large single chip such as thyristors, although they are also now being applied to multichip devices such as IGBTs.

As shown in Figure 2.27e, a pressure contact package consists of two large copper pole pieces between which the chip is sandwiched. Externally applied pressure ensures contact between the chip and the contact plates. For the gate connection an internal pressure contact is provided. A ceramic body provides the rest of the enclosure and may be ribbed to increase the creepage distance over the surface. After assembly the joints are welded, the package evacuated and sealed. In use the device is clamped between two plates, usually the heatsinks, and a known pressure applied. As the pole pieces are in direct contact with the chip the heatsinks and mounting plates are all live. Great care must be taken to ensure the correct even contact pressure; this makes the mechanics of pressure contact devices quite complex.

Silicon Carbide
The vast majority of semiconductor devices use silicon (Si) as their base material. Although the process technology is well developed, Si has some limitations for power devices, most notably a maximum junction temperature of 125C to 175C. By using silicon carbide (SIC) as the base material, devices may be operated with junction temperatures of 250C to 350C. Furthermore, SiC has a breakdown field strength five times that of Si and a thermal conductivity three times that of Si. Overall SiC offers lower loss devices, faster switching and much higher operating temperatures. Despite the significant technical gains in performance possible, SiC is much more difficult to process than Si and the fabrication technology is relatively young. This makes devices rated at more than a few amps not commercially viable, although this is expected to change over the next five to ten years.

Pressure contact devices have several features which make them well suited to very high-power, high-reliability applications: Double-sided cooling gives significantly reduced thermal impedance compared to single-sided cooling. This allows a device to operate at a high loss per unit area. The two main causes of power device wear out are the failure of wire bonds and soldered contacts, due to thermally induced mechanical stress. As a pressure contact device has no wire bonds or soldered joints it has a very good thermal cycling capability which is especially important in applications such as railway traction. In the event of a failure the device will go short circuit. This allows redundancy to be built into high-voltage applications which have several devices connected in series. The package has a high rupturing IZt so that with correct fusing it is possible to prevent rupture of the device in the event of failure. It is very important in high-voltage applications (>3 kV) to limit damage to other equipment.


Power devices are available in a very wide range of packages, a selection of which is shown in Figure 2.27. The very smallest packages are the wafer-scale packages, which only have a layer of passivation on the surface of the chip. These must be soldered to a printed circuit board (PCB) or other substrate and switch ratings are limited to a few hundred VA. Packaged devices, normally housing a single chip, for through-hole or surface mounting, are available in industry standard sizes such as D2-Pak, TO220, TO-247 in ratings up to 10 kVA. For larger devices with ratings up to 1 MVA isolated base modules with moulded plastic cases are used. There is a wide variety of packages containing one or more devices but with little standardisation in sizes. For the highest power devices, up to 10MVA, pressure contact packages are used. These only house a single device.

Large Wire-Bonded Packages for Power Modules

Large wire-bonded packages are used for power modules housing single devices rated over 50 A or multiple devices rated over 10 A. Maximum ratings are typically 4.5 kV and 2.5 kA. They are very widely used and are much cheaper than the equivalent device in a pressure pack.




1 to lOmm glass passivation

silicon chip

solder bumps

10 to 40 mm epoxy

bond wire leads

silicon chip copper slug

30 to 150 mm

lead., ~lastic case

)3 DBC substrate

40 to 200 mm leads I"1 bond wire soft-gel fill silicon chip plastic case

DBC substrate

copper base plate

40 to 120 mm

spring contact gate lead

copper pole piece

silicon chip

ceramic body

copper pole piece

Figure 2.27 Package cross-sections types a w a f e r scale

b c d e TO-220 direct b o n d e d c o p p e r (DBC) c o p p e r base pressure


There are two basic types of package, either with or without a copper base plate, as shown in Figure 2.27c and d. The copper base plate gives better transient thermal impedance, aids heat spreading across the heatsink and makes the device less prone to damage due to incorrect mounting. The chip or chips are first soldered to the direct bonded copper (DBC) substrate. This consists of two layers of copper, between which there is an Aluminium Oxide (A1203) or Aluminium Nitride (A1N) insulator. The DBC isolates the chips and the power circuit from the base plate of the device and so the heatsink may be earthed. The top layer of copper is etched to form an interconnect pattern similar to a printed circuit board. Wire bonding is then used to connect the substrate, chips and package terminals. A plastic package supports the power terminals and provides mechanical protection. To provide electrical insulation and environmental protection the chips and bond wires are covered in an insulating gel. In use, the package is fixed to a heatsink using screws with electrical contacts and either soldered or screw terminals.


variety of locations and features may be incorporated to increase the creepage and clearance distances between terminals. The DBC substrate allows several chips to be placed within the package and connected in a variety of ways to form either high-current single switches or complete power circuits, as in the case of a PIM or CIB module. Where additional control or monitoring circuits are required, these may be easily incorporated in the DBC or a separate PCB within the housing. Devices with solder terminals are only designed to be soldered to a PCB. Screw-terminal packages are more flexible as they may be screwed to a PCB for low or medium-current applications or to busbars for higher currents. Compared to pressure packs, wire-bonded packages have some disadvantages for very high power applications: surge current rating may be wire-bond limited, in the event of a failure the device will go open circuit as the wire bonds blow off, and in the event of a major device failure the package may rupture.

Small Wire-Bonded Packages for Discrete Devices

Small wire-bonded packages for discrete devices are made in very large volumes and have the lowest production costs. In general, they only contain a single chip although some packages are available with two chips, for example an IGBT with antiparallel diode. The ratings of these devices are limited by the available chip area and the current rating of the leads. Surface mount (SM) packages are commonly available up to 30A, with through-hole devices up to 70 A.

Single power devices with screw power terminals are available in ratings from approximately 50 A to 2500 A. For currents above 800 to 1000 A several parallel contacts are used. There is a wide variety of package sizes, from 20 mm x 92 mm to 140 mm x 190 mm, some sizes being adopted as a de facto standard. Screw terminal packages are also available with multiple power devices. These may range from a half bridge (two power switches) to a three-phase inverter. Ratings are available up to 450 A, 1200 V as a three-phase inverter. For packages which contain multiple devices at current ratings up to 150 A, all solder terminals tend to be the most cost effective. Two industry standard packages are the Econo2 and Econo3. These are available from a large number of manufacturers and may house a three-phase inverter, three-phase rectifier, or rectifier, inverter and brake chopper. The latter type are often known as power integration modules (PIM) or converter inverter brake (CIB) modules. Packages containing multiple devices are also available with screw terminals for the power connections or pressure contacts. In addition to the main power devices the control, monitoring and protection circuits may also be mounted within the module. These parts are commonly known as intelligent or integrated power modules (IPMs). The additional circuits are normally assembled onto a small PCB, which is wire bonded to the main power devices.

The chip is first soldered to a copper slug before wire bonding to the lead frame. An epoxy resin is then moulded over the chip to provide mechanical support for the leads, electrical isolation and environmental protection.


There is a wide variety of packages, many of which are industry standard. Packages, which are designed for screw mounting, such as TO-220 or TO-247, can also be clip mounted. Using a clip saves parts and labour and gives more even contact pressure to the heatsink. By having a package without a mounting hole, the popularity of clip mounting is being exploited to fit larger chips within a given footprint. In many power applications it is desirable to have isolation for the heatsink. Traditionally this has been achieved by putting an insulating washer between the device and heatsink. By moulding the epoxy resin around the whole copper slug and chip a fully insulated package can be produced. As the epoxy is a poor thermal conductor, device ratings are reduced. To overcome this limitation, devices are also available which have a small DBC substrate rather than a

Large wire-bonded packages with moulded plastic housings are the most flexible of power device packages. The moulded case allows many terminals to be placed in a



copper slug. This provides electrical isolation and much better thermal conductivity than epoxy resin.

Cost and performance requirements are paramount in deciding the most suitable power semiconductor device for a particular application. The following tables summarise device characteristics, packages and applications for the most popular power devices.

Packages are soldered to a printed circuit board, which in the case of surface-mount packages provides the cooling path in addition to electrical connection.

Table 2.2 Comparison of power semiconductor devices

Property Self-commutation ability Maximum r.m.s. current rating Maximum voltage rating Maximum switched VA rating Surge current ability Operating current density at rated device voltage SCR no 5000 A 12 000 V 30 MVA excellent GTO yes 2000 A 6000 V 30 MVA excellent BJT yes 1000 A 1600 V 1 MVA limited (2 x IRMS)
40 A / c m 2

MOSFET yes 300 A 1500 V 30 kVA limited (4 X InMS) 75 A / c m 2 at 200 V 15 A/cm 2 at 800 V

IGBT yes 2400 A 6500 V 4 MVA limited (2 X IRMS)

140 A / c m 2

IGCT yes 1700 A 5500 V 12 MVA excellent

(15 x Imvts)
140 A/cm 2 at 2 kV

(15 x Im4s)
30 A/cm2 at 4.5 kV

(15 X InMS)
30 A / c m 2 at 4.5 kV

at 1000 V

at 1200 V 35 A/cm 2 at 3.3 kV 150C medium low very good very good excellent (100 %)

Maximum junction temperature C On-state losses Switching losses Turn-on ability Turn-off ability Turn-off safe operating area (percentage of rated voltage at rated r.m.s, current) Load short-circuit turn-off ability Snubbers usually required Minimum on or off time Maximum switching frequency Switching time controllable from drive circuit Drive circuit power Drive circuit complexity Series and parallel operation


125C medium very high medium (di/dt limit) poor-slow and lossy poor (50 %)

150C medium high good medium-long storage time medium (75 %)

150C high very low very good very good excellent (100 %)


very high medium (di/dt limit) none via gate na

medium medium (di/dt limit) good medium (70 %)

none yes
10-100 Its

poor (2 x IR.vts) yes 10-50 Its 500 Hz

medium (4 X IRMS) yes <liLts 5000 Hz

medium (4 x It~ts) no < 100 ns

100 000 Hz

excellent (10 x IR~ts) no < 1 Its

10 000 Hz

poor (2 x IRMS)

lOgs 500 Hz

250 Hz








low device selection and passive components required

high high very difficult series or parallel

high medium series difficult; parallel requires device selection



low fairly simple series and parallel

low fairly simple series and parallel, selection may be needed for parallel

high high fairly simple in series, more difficult in parallel

Chapter 2.1


Table 2.3 Typical power semiconductor device application areas as a function of system voltage and equipment VA rating
Application Supply voltage and equipment VA RATING up to 240 V A.C., 400 V D.C. up to 1 kVA a.c. motor drives voltage source inverter current source inverter cycloconverter soft starters d.c. motor drives force commutated line commutated from 240 V A.C., 400 V D.C. up to 690 V A.C., 1200 V D.C. from 1 kVA up to 1 MVA above 690 V A.C., 1200 V D.C. above 1 MVA





Table 2.4 Availability of power device types and ratings in a variety of packages
Property Available devices Maximum voltage rating Maximum current rating Electrical failure mode Power circuit connections Control circuit connections Mounting method Cooling method Discrete SCR, GTO, BJT, MOSFET, IGBT 2000 V 100A open circuit solder solder solder, screw or clip convection to air, conduction to PCB or heatsink only with selected packages low Power module SCR, BJT, MOSFET, IGBT 6500 V 2500 A open circuit solder, screw, or pressure contact solder, screw, or pressure contact screw single-sided conduction to heatsink yes medium Pressure pack SCR, GTO, BJT, IGBT, IGCT 12 000 V 6000 A short circuit pressure contact flying leads pressure plate double-sided conduction to heatsink no high

Isolation from heatsink Package rupture current

Table 2.5 Power semiconductor symbols

Meaning aNPN large signal current gain of a common-base NPN transistor; ratio of collector current to emitter current as NPN, but for a common-base PNP transistor turn-off gain; ratio of anode current being controlled to negative gate current required to produce turn off conduction angle static forward current transfer ratio of a common emitter transistor; ratio of d.c. output current to d.c. input current switching frequency anode current continuous base current forward base current reverse base current continuous collector current peak collector current peak value of pulsed drain current gate current Specific to: GTO Meaning Specific to: GTO GTO


aPNP /3off




0 hfe



18 IB1 182



maximum peak positive gate current maximum negative gate current during turn-off interval holding current latching current drain current of MOSFET in IGBT equivalent circuit collector current of parasitic PNP transistor in IGBT equivalent circuit reverse recovery current direct on-state current the average on-state current rating the r.m.s, on-state current rating nonrepetitive peak controllable on-state current the maximum current which can be turned off under defined conditions power dissipation reverse recovery charge MOSFET channel resistance drift region resistance







Table 2.5 (Contd.)


Specific to:


Specific to:

R thjc case


tl tic tgq t~t toe ton tp tq tq~ trc trr ts




drain source on resistance resistance of epitaxy region in IGBT equivalent circuit thermal resistance, junction to case case temperature junction temperature maximum allowable junction temperature conduction time fall time turn-off crossover time turn-off time turn-on time turn-off time turn-on time pulse duration total turn-off time recombination time turn-on crossover time reverse recovery time storage time base-emitter voltage collector-emitter breakdown voltage, gate-emitter short circuited drain-source breakdown voltage, gate-source short circuited collector-emitter supply voltage collector-emitter voltage

VCE(sat )

VcEo(sus) VcEs






VDS Va VaE Vas



Wco,,d Woff



collector-emitter saturation voltage collector-emitter sustaining voltage, with open base, Ic specified collector-emitter breakdown voltage, specified with zero gate emitter voltage peak collector-emitter voltage during short-circuit current fall time collector-emitter voltage with reverse-biased base-emitter junction direct off-state voltage drain-source voltage gate voltage gate-emitter voltage gate-source voltage reverse voltage repetitive peak reverse voltage voltage spike between anode and cathode during fall time energy dissipation during the conduction time energy dissipation during the turn-off time energy dissipation during the turn-on time






In order to concentrate on the salient characteristics of the many alternative power conversion circuits used in power electronic drives today, it is necessary to make a number of simplifying assumptions. A somewhat idealised theory will therefore be presented. Practical aspects, such as switching delays, will only be discussed where they have practical significance. The diversity of machines which can be used with alternative power converters will be limited to those seen as being of greatest practical importance.

overall A.C. power factor defined as the ratio of mean power (W) to volt-amperes (VA) [note: the power factor thus defined equals the product of the displacement factor (the fundamental power factor or cos4~ i.e. the phase shift of the fundamental current with respect to the A.C. supply) and the distortion factor (the ratio of the r.m.s, of the fundamental current to the r.m.s, of the total current)] maximum attainable A.C. power factor that can be achieved using capacitors only to counter the fundamental VAR consumption of the converter A.C. supply current harmonics for a constant D.C. current Id D.C. voltage as a function of A.C. voltage D.C. power for a constant D.C. current Ia voltage tipple-although form factor or peak-peak values are often used as a measure of voltage ripple, it is more useful in practice to consider a factor M, which is a measure of the volt-second integral of the voltage pulses;


Prior to considering the relative merits of alternative converters it is necessary to establish meaningful performance parameters. Useful to the user and system designer are: r.m.s, value of the A.C. current for a constant D.C. current Id

Chapter 2.2
the peak-peak current ripple A i d is then given by:
A i d -- MVdo


where Vao = the maximum attainable D.C. voltage, f = the A.C. frequency and L = D.C. circuit inductance pulse number p is the number of pulses of D.C. voltage during one complete A.C. cycle (1 cycle = 20ms for 50 Hz supply, 16.67 ms for a 60 Hz supply)

In the half-controlled circuit, two of the thyristors shown in Figure 2.30 are replaced by diodes. A number of diode combinations are possible: (Ap + Bp), (An + Bn), (Ap + An) or (Bp + Bn). Figure 2.32 shows typical waveforms for the case when Ap and Bp are diodes. Note the absence of any negative voltage, and the reduction in A.C. current periods. Only energy flow from the A.C. to D.C. side is possible.


With growing demands to reduce harmonic effects on the supply, and improve the supply power factor, circuits are now coming to the market, notably in switch-mode power supplies (SMPS), which draw near sinusoidal currents from the A.C. supply at unity power factor. A typical circuit is shown in Figure 2.33. It is expected that, as greater regulation is introduced in respect to allowable harmonic content of drive systems, this type of circuit will become more common. However, in such situations it is often more cost effective to have a single

Converters for Connection to a Single-Phase Supply

Only two power circuits need be considered in this category. The half-wave configuration of Figure 2.28 is not particularly useful for power applications, but is included as an introduction to semiconductor behaviour in bridge circuits. The single device is available for conduction from 0 to 180 (the positive half cycle). A freewheeling diode may be added across the load to conduct the load current during the negative half cycle, and prevent it being reduced to zero. For long time-constant loads, the load current can be considered to be continuous: derived from the supply during the positive half cycle and carried by the diode during the negative half cycle. When this circuit is used, a capacitor often replaces the freewheeling diode, maintaining the output during the idle half cycle. A notable application of this particular circuit is for high-frequency SMPS secondaries. The detailed characteristics of this circuit will not be considered further. Figm'e 2.29 shows full-wave rectification from an A.C. supply in its most popular form. The four diodes conduct in diagonal pairs during every alternate half cycle of A.C. line voltage.





Figure 2.29 Single-phase full-wave uncontrolled bridge

Again, only two power circuits are of practical importance. Figure 2.30 shows the power circuit for the fully controlled bridge together with associated A.C./D.C. relationships. Figure 2.31 shows how the D.C. voltage can be varied by adjusting the firing delay angle c~. Since negative D.C. voltages are possible, energy flow from the D.C. side to the A.C. side is possible.



~ Vo.c.


( )




conducting thyristors top row


A.C./D.C. relationships
VD.C 0 0

bottom row
An BR An Bn



Vab Vba

+ [d --]d

Figure 2.28 Single-phase half-wave uncontrolled bridge

Figure 2.30 Single-phase fully controlled bridge


DRIVE CONVERTER CIRCUITS: A.C. t o D.C. P o w e r Conversion


Bp An
~l--ll t:z


conduction periods


power factor controller



Figure 2.33 Single-phase sine wave input converter


"', ..


D.C. side voltage waveforms




I i

A.C. supply current waveforms

Figure 2.31 Single-phase fully controlled bridge, D.C. voltage control



A n


gn I

Figure 2.34 Three-phase full-wave uncontrolled bridge

conduction periods

n o" i . o " " r ~ vba


More complex, and expensive, converter arrangements exist which monitor the supply voltage and control currents in such a way as to minimise supply voltage distortion. Such systems are rare.

Converters for Connection to a Three-Phase Supply


D.C. side voltage waveformi

"'-. .... -'"

i i

A variety of uncontrolled converters is available, however only one is of practical importance in regard to drive systems, namely the full-wave bridge converter. This arrangement is as shown in Figure 2.34. In contrast to the single-phase bridge, where altemate pairs of devices conduct, switching in the three-phase bridge alternates between the upper and lower row of devices. This means that there are six conduction periods per A.C. cycle, each device conducting for a period of 27r/3 (120 elect.).

A.C. supply current waveform

Figure 2.32 Single-phase half-controlled bridge, D.C. voltage control

Two power circuits are of practical importance: supply converter with a common D.C. bus feeding multiple drives. Commercial three-phase sine wave input converters tend to take the form of an IGBT bridge connected to the supply operated as a current controller to draw sinusoidal currents. A practical implementation of such a scheme is shown in Chapter 6.3. The fully controlled circuit is by far the most important practical bridge arrangement. Figure 2.35 shows the power circuit together with associated A.C./D.C. relationships. Figure 2.36 shows how the D.C. voltage can be varied by adjusting the firing delay angle a. The pulse number, p, of this bridge equals 6. Energy flow can be from A.C. to D.C. or D.C. to A.C.

Chapter 2.2
Table 2.6 Salient characteristics of the main single-phase A.C to D.C converters
Bridge Firing angle Fully controlled a Half controlled a ~2N2 Vs



~2N2 Vs


cos a 1 2c 2 ~ No s 7r a

0.5 (1 4- cosa) N { (Tr - c~)/Tv} N{Z/Tr} (1 + cos a ) N (Tr - a ) 1 + cos a

Vd/Vdo Is/Id
Overall power factor Maximum corrected power factor Input Input

cos a N(Tr2/8 - sin 2 a) cosa sina 0 for n even

N{Tr/2. (Tr- a) - sin 2 a }

0.5 (1 + cosa) 0.5 sina 0 for n even 0.64/n- N(1 + cos na) for n odd

power/Pdo VARs/Pdo

Supply current nth harmonic/Id

0.9/n for n odd na

Phase of supply current harmonics



/a I


n I. n2~Cw n

v~= o.75 V~o

Vt,a Vca Vc~ Va~ Vac Vbc Vba Vca Vo~


conducting thyristors
top row

A.C./D.C. relationships

bottom row
An Bn Cn An Bn Cn An Bn Cn

Vab Vac Vba



+ id + id -- id

-- i d 0 + id

0 -- id 0



..'" Vd=0.75 Vdo



+ id

-- i d



-- id

-- id

+ id
4- i d

Vc. Vcb Vab Vac V.c vb.

vo. Vcb


Figure 2.35 Three-phase fully controlled bridge

Figure 2.36 Three-phase fully controlled bridge, D.C voltage control



In the half-controlled circuit either the top three devices of Figure 2.35 (Ap, Bp and Cp) or the bottom three devices (An, Bn and Cn) are replaced by diodes. The pulse number of this bridge equals 3. Only energy flow from A.C. to D.C. is possible. The voltage tipple is much greater than in the case of the fully controlled bridge, but the A.C. current drawn is lower at reduced D.C. voltage. Figures 2.37 and 2.38 show the half-controlled bridge.

current ripple on A.C. supply harmonics is of great practical industrial importance mainly in relation to three-phase bridges (ignoring the single-phase traction requirement). Practical experience has led to the adoption by many of the following values: 15 = 0.2511 (ideal = 0.211) 17 - 0.1311 (ideal = 0.1411) 111 -- 0.0911 (ideal - 0.1111) 113 -- 0.0711 (ideal - 0.0811) In general, the amplitudes of higher harmonics are rarely of significance, in regard to supply distortion. Under conditions of very high D.C. current ripple, the fifth harmonic can assume a considerably higher value than that quoted above. A practical example would be an application with a very capacitive D.C. load (e.g. a voltage source inverter): in such a case where no smoothing choke is used 15 could be as high as 0.511.

Voltage Ripple Characteristics

The voltage ripple characteristics for the most significant bridge configurations are shown in Figure 2.39. It should be noted that these characteristics are for idealised conditions of smooth D.C. current and zero supply impedance. Increases in supply impedance generally tend to result in somewhat lower D.C. voltage tipple levels.

Practical Effects
The characteristics presented above have, for the most part, been based upon idealised conditions of negligible A.C. inductance and constant D.C. current. Although these assumptions provide a convenient means for comparison they are not often valid in practice. It is not practicable to consider all such effects here. The effect of D.C. link

D.C. Motor Drive Systems

In principle, little has changed since 1896 when Harry Ward Leonard presented his historic paper 'Volts versus ohms the speed regulation of electric motors'. In practice, however, many advances have been made from auxiliary machines through mercury-arc rectifiers to thyristors.





_ ~ A - - - - - z~ -1.0



0.0 mean D.C. voltage/Vdo


Figure 2.39 Ideal voltage ripple characteristics

1 phase 1 phase 3 phase 3 phase fully-controlled bridge half-controlled bridge fully-controlled bridge half-controlled bridge x A

Ani Bn.L Cn,, L

v i

Figure 2.37 Three-phase half controlled bridge




I On

I On

- - ~ - an



, ",


,, ,,
















I,, ,I,

Figure 2.38 Three-phase half controlled bridge, D.C. voltage control

Chapter 2.2
Table 2. 7 Salient characteristics of the three-phase A.C. to D. C converters described in the text
Bridge Firing angle Fully controlled a a > 7#3 3N2 --G


Half controlled a a < 7r/3 3N2 --G








Overall power factor


0.5 (1 + c o s a ) N {(Tr- a)/Tr} N3. (1 + cos a)

N(71 ol)

N(3/2) (3/70 cos a

N(3/2) (3/270 - (1 + cos a)

Maximum corrected power factor


N(Tr2/9 - sin2 a)

1 + cosa N{27r(Tr - a) - sin2 a}

1 + cosa N{47r2/9 - sin2a}

Input power/Pao Input VARs/Pao Supply current nth harmonic//d

cos a sin a 0 for n = 3, 6, 9 . . . 0 for n even N6/mr for n odd

0.5 (1 +cosc 0 0.5 sin a 0 for n = 3 , 6, 9 . . . N(3/mr) .N(1 - cos nc0 for n even N(3/mr).N(1 + cos na) for n odd

Phase of supply current harmonics



\ I \I

I\ / \

I 1
Figure 2.40 Single converter D.C. drive
The D.C. motor is still a versatile machine for variable-speed drive systems and is often the preferred choice when considerations such as freedom from maintenance or operation under adverse conditions are not paramount. In Chapter 1.1 it has been shown that complete control o f a D.C. machine can be achieved b y controlling the armature voltage, Va, and the field current, I f Two power converters are employed for this purpose in most variable-speed drives which employ the separately excited D.C. machine. (In referring to the number o f converters in a drive, it is c o m m o n to ignore the field converter - this nomenclature will be adopted below). It is relatively c o m m o n in simple drives for the field converter to be a single-phase uncontrolled bridge thereby applying fixed field voltage.

Figure 2.41 Single converter reversing~regenerative D.C drive

In applications where motor resistance varies with temperature, or on sites with poorly regulated supplies which result in unacceptable variations in field current, a controlled power converter is employed with current control. Such field controllers are further discussed later as applied to field weakening control.

Figure 2.40 shows a single-converter D.C. drive. In its most basic form the motor will drive the load in one direction only without braking or reverse running. This is said to be a single-quadrant drive, only operating in one quadrant o f the torque-speed characteristic. Such drives have wide application from simple machine tools to fans, pumps, extruders, agitators, printing machines etc.



If the drive is required to operate in both the forward and reverse directions and/or provide regenerative braking, a single fully controlled converter can still be used. However, some means of reversing either the field or armature connections, as shown in Figure 2.41, must be added. Reversal of armature current can involve bulky (highcurrent) reversing switches but, due to the low inductance of the armature circuit, can be completed in typically 0.2 seconds. Field current reversal takes longer, typically of the order of 1 second; however lower cost reversing switches may be used. The field reversal time can be reduced by using higher voltage field converters to force the current. Forcing voltages up to four per unit are used but care must be taken not to over stress the machine. Obviously this increased voltage cannot be applied continuously and requires either a switched A.C. supply or a controlled field converter. Armature and field reversal techniques are used where torque reversals are infrequent such as in hoists, presses, lathes and centrifuges.

+1500rev/min can still be achieved in approximately 200 ms. This circulating-current-free dual converter is by far the most common industrial four-quadrant D.C. drive and is used in many demanding applications - paper, plastics and textile machines where rapid control of tension is required are good examples.

The output power of a motor is the product of torque and speed. If torque reduces in proportion to increases in speed, the motor is said to have a constant power characteristic. In applications where material is coiled or uncoiled at constant tension, the torque required to produce that tension varies in proportion to coil diameter, and rotational speed required to maintain a constant peripheral speed (and therefore line speed) is inversely proportional to diameter. A motor having a constant power characteristic is well suited to this type of application, the advantage being that a smaller motor can be used than would otherwise be the case. Machine tool drives also make use of constant power operation, since loads are small at high speeds and heavy work is done at low speed. As explained in Chapter 1.1, the torque produced by a D.C. motor is proportional to the product of armature current and field flux. By weakening the field as speed increases, a constant power characteristic can be achieved. In practice, there are two major techniques for field weakening, both of which rely on a field controller which itself is a simple thyristor converter operating in a current control mode. In the first method, suitable for coiler and uncoiler applications, the field current reference is arranged to be inversely proportional to coil diameter (measured directly, or calculated from the ratio of line speed to motor speed). Since flux is not strictly proportional to field current, this method does not give a true constant power characteristic unless compensation for the nonlinear part of the motor field curve is applied. The second method is to use the field controller with an outer voltage loop having a fixed reference, and to use motor armature voltage as the feedback signal. At low speeds, the voltage loop saturates, providing maximum field current, since armature voltage is below the set value. As speed increases, the armature voltage rises to the point where it matches the preset reference in the field controller. Above that speed an error signal is produced by the voltage loop, which causes the field controller to weaken the motor field current and thereby restore armature voltage to the set-point level. The resulting characteristics are shown in Figure 2.43. As regenerative braking depends on the return of power from the motor to the mains, it cannot work if the mains supply fails due to a blown fuse or a power cut. Dynamic braking of four-quadrant drives is often encountered as a fail-safe means of stopping the motor and its load, and as the only means of (reverse) braking of single-ended drives. This involves switching in a resistor across the D.C. motor.

When a four-quadrant drive is required to change the direction of torque rapidly, the delays associated with reversing switches described above may be unacceptable. A dual converter comprising two fully controlled power converters connected in inverse parallel can be used as shown in Figure 2.42. Bridge 1 conducts when the armature current IA is required to be positive, bridge 2 when it is required to be negative. There are two common forms of dual converter. In the first, both bridges are controlled simultaneously to give the same mean output voltage. However, the instantaneous voltages from the rectifying and inverting bridges cannot be identical, and reactors Lp are included to limit the current circulating between them. The principal advantage of this system is that when the motor torque, and hence current, is required to change direction (polarity), there need be no delay between the conduction of one bridge and the other. This is the dual-converter bridge with circulating current. In the circulating-current-free dual converter, only one bridge at a time is allowed to conduct. The cost and losses associated with the Lp reactors can then be eliminated, and economies can also be made in the drive control circuits. However, the penalty is a short time delay, as the current passes through zero, while the thyristors in one bridge safely turn off before those in the second are fired. This delay introduces a torque-free period of typically 10ms. Speed reversal for a 3 kW drive of this type from - 1 5 0 0 to



Figure 2.42 Single-phase dual-converter D.C. drive

Chapter 2.2


constant field I

field weakening



approximately 100 Hz, which is too low for many servodrive applications. Thyristor-controlled A.C.-D.C. converters have an inherently poor input power factor at low output voltages. (Near unity power factor can be achieved using an uncontrolled rectifier feeding a D.C.-D.C. converter.) Electronic short-circuit protection is not economically possible with thyristor converters. Protection is normally accomplished by fuses. D.C.-D.C. converters are however more complex and somewhat less efficient than A.C.-D.C. converters. They find application mainly in D.C. servodrives, rail traction drives and small fractional kW drives employing permanentmagnet motors. Since step-down converters are of greatest practical importance emphasis shall be placed on their consideration. For the purpose of illustration bipolar transistors will be considered, however MOSFET, IGBT and GTOs are widely used.

constant torque

constant power

"'I 0 0 0

E base speed speed

Figure 2.43 Constant power operation using a field controller

Step Down D.C.-D.C. Converters


Since the kinetic energy of the motor and its load is converted into heat by the braking resistor, it is important to rate it correctly for the duty it is expected to perform, taking account of load inertia and the number of stops per hour. The most basic D.C. to D.C. converter is shown in Figure 2.44. The output voltage is changed by pulse-width modulation (PWM) - that is, by varying the time for which the transistor T is turned on. The voltage applied to the motor is therefore in the form of a square wave of varying periodicity. Because the motor is inductive the current waveform is smoothed, the flywheel diode D carrying the current while the transistor is turned off. The basic formulae relating the variables in this circuit are as follows:
Va = VD.C. " t ' f


D.C.-D.C. power converters (often referred to as choppers) provide the means to change one D.C. voltage to another. It is more usual for the conversion to be to a lower voltage, although step-up converters are available and have significant potential for the future. D.C.-D.C. power converters are fed from a D.C. supply usually comprising an uncontrolled A.C. to D.C. converter or alternatively a battery supply; the controlled D.C. output can then be used to control a D.C. machine as in the case of the controlled A.C. to D.C. converters. D.C. drives employing controlled A.C. to D.C. converters have several important limitations, which are overcome by the D.C.-D.C. converter: The inability of a thyristor to interrupt current means that an alternating supply is necessary to commutate the converter- this precludes operation from a D.C. supply. This is a common requirement on battery vehicles and D.C.-fed rail traction. The D.C. ripple frequency is determined by the A.C. and is, for a 50 Hz supply frequency, 100 Hz for single-phase and 300 Hz for three-phase fully controlled bridges. This means that additional smoothing components are often required when using high-speed machines, permanentmagnet motors or other special motors with low armature inductance. As a result of the delay inherent in thyristor switching (3.3 ms in a 50 Hz three-phase converter) the current control loop band width of the converter is limited to

A Ia = VD.C./(4" La "f)
wherefis the frequency of transistor on pulse (Hz), A/a is the maximum deviation of armature current, La is the motor inductance, t is the on pulse duration (s) and VD.C. is the source D.C. voltage. The circuit is only capable of supplying unidirectional current and voltage to the motor and is therefore not capable of four-quadrant operation, that is reversing or regenerating. Industrial applications for this circuit are normally limited to drives below 5 kW and simple variable-speed applications. In traction applications, however, drives of this fundamental type are designed at ratings of many hundreds of kW.


In order to achieve full four-quadrant operation a converter must be capable of supplying reversible voltage and current to the motor. A circuit that is capable of two-quadrant operation - that is motoring and braking in one direction only - is shown in Figure 2.45. This converter is able to reverse the current flow to the motor but unable to reverse the motor terminal voltage and hence the speed. During motoring, the converter operates as the basic chopper with T1 and D2 carrying the current. During braking (or regeneration) T1 is inoperative and T2 controls the current.

8U[)IEJ q asJaAaJ 9 8u!Jo),ouJ pJEMJO~t q

~!n:)J!:) e

somos se qons S0A!.I "~)'O oouetmojaod p q~!q u.t posn iop!Ax oae Sa0U0AUO0 ~ueapenb-anoj osoq

Ja~a4uo9 0"0-'9"0 .~ue.~penb-.mo_-I 9t,,'rc a.m6t...-I



A o oldd.u 3 o ~u!lqno p ~ pu~ O~EIIOA oldd.u u! uo.uonpoa ~ s! mo33o ~ou oq/, "C[ pue IZ v!,t so~inoa!o ~uoano l o O q ~ l j oq~ leql os 'uo ,i pu~ JJO potum s! kL '~ugonpuoo uooq OAeq lz,L pue IZ :tmje 'oidtuexo aod "pouod j j o oql ~upnp opo!p e pue aols!sue:n e v.~,t pomlnoa!o s! ]uo~no iooqA~ij oq~ onb.mtloo] s.tq] qlt.A -uogelnpom po~po ~ -olqno p se UA~Om onb!uttoo ~ e ~q 0tUO0a0AO 0q U~0 tu0Iq I -oad s!qz "0SaOA~o:toJoaoql s! oldd.u ] oq] pue 'a0U0A -uo0 oldtu!s oql j o ~eq~ O9.IAVISt offmiOA oldd V mdmo oq~ j o oPmiidtue oql ~,eq~,s.t .IOLIOAHO0sIq~,oi0 o~elueApes!p ou 0



".,(IU0A0 ;~.lOttl S09.IA0p :toaxod oq~,u.t sosso I oq~, m,nq.m,s!p o~, :top:to m smm,s~s ie!o:totu -tuoo ~uetu u! pmdope oae soouonbos ~u.tqm,iaxs p00ueApe OaOlAI ", s.t uo.n,e:todo meapenb-anoj iin_J oq~, oouoq qo pue p0S/0AOI St :luoMn9 oql '~u.tlonpuoo pue EL s:to~,s.ts -[II~l1 qlI.l~ "~LII.7~I?,I;~I,I;~;glS.I ;gA.Up~)ql pue p;gSJ;~A0/SI. poads oql oouoq '.uogooa!p otuus oql u! luo~no ql!ax mq ;DS.I;~A;9.I ol 0~Ul[OA mdmo oq~, osneo ii!m ~,uo3 aod 0~; aXoloq oioo ~ n p j o uop, onpoa ~u V "op~o Lmp ~,uoo aod 0g ~,e oaoz ol poonpoa s[ 0~lZlIOA 0ql .,~[luonbosuo9 pul~ p0SlOAOI S.I iddns :tmotu oq~, honpuoo lzO pue E(I uoqA,~ "po.uodjjo oq~,~u!:mp ~,onpuoo lzO pue E(I sopo!p pue 'povod uo oql ~u!:mp uo p0qm!A~S oae 17 pue I s:tms!su~a mdmo OA.II.Isod '~U.UOlOtU ~u.un(I

fiU!)leJ q pJeAAJO~t 9 fiU!JO:~OUJ pJEAAJO~tq 3,!n)J !) e

uoge:touo~oa pue ~UISa0AO:tJO "O'! ;uo~na pue 0~ellOA ~lq!SI~AOI ~Uilddns jo olqedea aOU0AUO0 ,ue:tpenb-anoj a!seq e saxoqs 91z'Z oan~!:I
8:I.LS:IANO~) 9 c ] - 3 a J.NVaavnb-~no-I

.~a.,~aAuo:) 0"0-'.9"0 ~ue.~penb-oNtl ~P'E a.m6t.d

~[ddns plog ao o:mmtme oq] 0SIOAO.I0~l qg].IA, .IOAO0~Ul~q9 XS JO sueotu ~q mo po!~eo s[ ~UtSa0A0:~ oaoqax suopeoildde uo!;oea] ui posn sotugotuos s[ ~! ~;io![dtu!s m! jo osneooq 'aOAOaXOH "suogeoildde ieuo!;vo:qp!un u! posn ~Iuo ~iietmou s! ~! l~S~OAOapoods ao~otu j o olq~d~o ~ou s! ~!no~!o s!q~ oouIS

oql oouoq '.Xlddns oql olu! ~Ioeq I C[ via s! ~uo~no oql :~oj qled Xluo oql 'jjo stum E i uoq/~ ""7 ooumonpu! aolotu Kq pmitu!i 'XIOAII~ffOU dn spi!n q luo:uno aolotu 'spoFmd uo sl! ffupno uo ~Cl

la~aAuo:) D'G-D'G .~uelpenb-al6Ut.S ~t;,'E aJn6t._,-I

uo ~CI



uo ~I~


aslnd uo






UO!SJOAUOZ) JaAAOd ":)'(30~. "D'(] :S.LIn981~) 8:II83ANOZ) 3A18(]

Chapter 2.2


Step-up D.C.-D.C. Converters

As with step-down converters, many alternative configurations exist for step-up converters. Although a full description is not necessary, the principle is of value. Figure 2.47 shows a much simplified arrangement of a stepup converter. When T is turned on, current builds up in inductor L. When T is turned off, the energy stored in L is transferred to capacitor C via D. When the capacitor voltage, which is the same as the motor armature voltage, reaches the desired level, T is turned on once more. C cannot discharge via T as diode D is reverse biased. In this way a stabilised voltage typically twice the input D.C. voltage can be obtained. This circuit is particularly useful when operating on low-voltage supplies and can lead to very cost-effective converter designs.

considered, and consequently are very flexible in their application. Major inherent features include: Multimotor loads can be a p p l i e d - this can be very economical in applications such as roller table drives, spinning machines etc. Inverter operation is not dependent upon the machine - indeed, various machines (induction, synchronous or even reluctance) can be used provided that the current drawn is within the current rating of the inverter. Care should be taken where a low power-factor motor is used (e.g. reluctance) to ensure that the inverter can provide the required VARs. Inherent open-circuit protection, very useful in applications where the cables between the inverter and motor are in some way insecure (e.g. fed via slip-tings, subject to damage etc.). Facility to ride through mains dips can easily be provided by buffering the D.C. voltage link with capacitance or, where necessary, a battery. Motoring operation only in both directions is possible without the addition of resistive dumps for braking energy or expensive regenerative converters to feed energy back to the supply.


This category of A.C. drive, commonly termed a variable frequency inverter, is by far the most important in respect of the majority of industrial applications. It is considered here as a complete converter; however the input stages have been considered earlier in isolation, and their individual characteristics so described are, of course, applicable. Alternative input stages to some of the drives are applicable. Also, some converters may be used with a variety of machine types. Only practically important combinations are described. The concept of these inverter drives is well understood rectification of fixed frequency, smoothing and then inverting to give variable frequency/variable voltage to feed an A.C. machine. Within this broad concept two major categories of drive exist: voltage source in which the converter impresses a voltage on the machine, and the machine impedances determine the current current source in which the converter impresses a current on the machine, and the machine impedances determine the voltage


A typical D.C. link square-wave voltage-fed inverter drive is shown in Figure 2.48. The three-phase A.C. supply is converted to D.C. in the phase-controlled rectifier stage. The rectified D.C. power is then filtered and fed to the inverter. Note that the D.C. link reactor is small compared to those used in current-source designs. Indeed, in drives up to about 4 kW it is not practically necessary. Some manufacturers omit the reactor in designs to 400 kW and above; however this has a significant effect upon supply harmonics and unduly stresses the rectifier and filter capacitor. The inverter switching elements shown as transistors T1 to T6 are gated at 60 intervals in the sequence in which they are numbered in the diagram, and each transistor conducts for 180 . The feedback diodes D1 to D6 are connected in inverse parallel with the transistors, and permit the retum of energy from reactive or regenerative loads through the inverter to the D.C. link. For a star-connected motor, synthesis of inverter output voltage waveforms is shown in Figure 2.49. The phase-toneutral voltage of the inverter has a six-step waveshape, and the corresponding phase-to-phase voltage has 120 conduction angle. The output frequency is controlled by the rate at which the inverter transistors are triggered into conduction by the inverter control circuitry. Reversing the firing sequence of transistors in the inverter changes the direction of rotation of the motor, and no switching of power leads, either on the incoming supply or to the motor itself, is necessary. The phase-controlled rectifier regulates the D.C. link voltage and this, in tum, determines the magnitude of the output voltage from the inverter. Thus, the output voltage/frequency relationship may be controlled to regulate the motor flux in the desired manner.

Voltage Source Inverters

The fixed-frequency mains supply is a voltage source behind an impedance. Voltage source inverters can be similarly


Figure 2.47 Step-up D.C.-D.C. converter


DRIVE CONVERTERCIRCUITS: A.C. t o A.C. P o w e r Converters w i t h Intermediate D.C. Link

D.C. link T3

inverter I I T5

, I ,


Figure 2.48 Square-wave voltage-fed inverter

Very high-speed motor operation is possible by increasing the output frequency. Faster switching devices such as MOS transistors and insulated-gate bipolar transistors (IGBT) can be used to achieve this performance.


It is known that the square-wave inverter gives objectionable torque pulsations at low-frequency operation, below approximately 5 Hz. This pulsating torque is due to the interactions of low-order harmonics with the fundamental voltage, causing a stepping or cogging motion to the rotor running at low speed. Hence, the pulsating torque limits the low-frequency operation of the square-wave inverter. Appropriate feedback control techniques or flux weakening can attenuate the low-speed pulsating torque problems. The existence of a phase-controlled rectifier to control the voltage of the inverter as illustrated in Figure 2.48 is an inherent weakness of this circuit. The phase-controlled rectifier will present a low power factor to the A.C. supply, at low speeds, and the D.C. link filter capacitor is large and reduces the response time of the system to voltage and hence speed changes. If the drive system is one for which regenerative braking operation is a requirement, the rectifier has to be of inverse-parallel type. The input power factor and response time of the drive can be improved by replacing the phase-controlled rectifier with a diode rectifier feeding a D.C. chopper which regulates the input voltage to the inverter. For recovering regenerative energy of the load, a two-quadrant chopper will be necessary. The alternative supply converter arrangement of a diode bridge plus chopper also provides a fixed voltage link, which is more economically buffered, if mains dip ride through is required. The voltage-fed square-wave drive is usually used in lowpower industrial applications where the speed range is limited to ten to one and dynamic performance is not important. Recently, this type of drive has largely been superseded by PWM-type voltage-fed inverters. Nevertheless, the voltagefed square-wave inverter can be easily adapted to multimotor drives where the speed of a number of induction motors can be closely tracked. It is also used in some highfrequency ( > 1 kHz) and some high-power applications.


phase-to-phase voltages


I\ //




r%. \

I// f
I ~


phase-to-neutral voltage

Figure 2.49 Square-wave voltage-fed inverter, output voltage and current

The advantages of the square-wave inverter are high efficiency (98 per cent), suitability to standard motors, potential good reliability and high-speed capability. However, it suffers from low-speed torque pulsations and possible lowspeed instability. In a square-wave inverter, each harmonic voltage amplitude is inversely proportional to the harmonic order and hence there are no pronounced high-order harmonics. These are filtered out by the motor leakage inductances.

Chapter 2.2



In the PWM inverter drive, the D.C. link voltage is uncontrolled and derived from a simple diode bridge. The output voltage can be controlled electronically within the inverter by using PWM techniques. In this method, the transistors are switched on and off many times within a half cycle to generate a variable-voltage output which is normally low in harmonic content. A PWM waveform is illustrated in Figure 2.50. A large number of PWM techniques exist each having different performance notably in respect to the stability and audible noise of the driven motor. Using the PWM technique, low-speed torque pulsations are virtually eliminated since negligible low-order harmonics are present. Hence, this is an ideal solution where a drive system is to be used across a wide speed range. Since voltage and frequency are both controlled with the PWM, quick response to changes in demand voltage and frequency can be achieved. Furthermore, with a diode rectifier as the input circuit a high power factor, approaching unity, is offered to the incoming A.C. supply over the entire speed and load range. PWM inverter drive efficiency typically approaches 98 per cent but this figure is heavily affected by the choice of switching frequency- the higher the switching frequency the higher the losses in the drive. In practice, the maximum fundamental output frequency is usually restricted to 100 Hz in the case of gate turn-off thyristors (GTO) or about 1 kHz for a transistor-based system. The upper frequency limit may be improved by making a transition to a less sophisticated PWM waveform with a lower switching frequency and ultimately to a square wave if the application demands. However, with the introduction of faster switching power semiconductors, these restrictions to switching frequency and minimum pulse width have been eased. In general, a motor with a large leakage reactance is desirable to limit the flow of harmonic currents and thereby minimise losses.

_ carrierwave A ~ ~ r ] ~ ~ ~ : sinewave ~/~ ~ ~ ~ ~ ~ ~ ~ar,,,~k~referencesignal

i/A' ,,'V","V",,,, ,' ', V '8 V',',V, ,", ,, , ,,,,,


, ,

, t


, ,


+sV"l Fll II II II- -!

0,v.I IIU II U_

per phase output voltage, _~ ~_ lower thyristor on (with respect to mid D.C. link point)

PWM line voltage


nr F U1
PWM phase voltage Figure 2.50 Sinusoidal PWM line and phase voltages CONVERTER-FED SYNCHRONOUS MACHINE
Once rotating, a synchronous machine generates A.C. voltages which can be used for the natural commutation of a converter connected to its terminals. Indeed, the connected synchronous machine behaves as the mains in respect of the A.C. to D.C. converters described earlier. Figure 2.51 shows the basic components of the drive system. A low-impedance or stiff D.C. current source is required and is obtained from a controlled rectifier and a series reactor. With a stiff current source, the output current wave is not greatly affected by the magnitude of the load. The synchronous machine can be approximately represented by a counter e.m.f, in series with an equivalent leakage inductance. The D.C. current is switched through the inverter thyristors so as to establish three-phase, six-stepped symmetrical line current waves. Each thyristor conducts for 120 and at any instant one upper thyristor and one lower thyristor remain in conduction. It is necessary to maintain an approximately constant angular relationship between the rotor and stator e.m.f.s and hence automatically maintain the correct inverter frequency. This is an important point. The inverter does not impose a frequency upon the machine, rather the machine itself determines the frequency. The motor cannot therefore pole slip. The drive is accelerated by increasing the current fed to the motor, which then accelerates and thereby increases the frequency.

At high powers and/or high voltages it is not possible to implement PWM strategies with high switching frequencies. The waveform of such systems can be improved by providing intermediate D.C. voltage levels. Commercial systems of this type exist, but their application is quite rare.
Current Source Inverters

Whereas each voltage-fed inverter can be used with most forms of A.C. machine, a different design of current-source inverter is usually adopted for synchronous and induction motors. Current-source drives are usually, but not always, single-motor systems, and since current is controlled, have simple short-circuit protection. In contrast to voltage-source inverters full four-quadrant operation is inherently possible.


DRIVE CONVERTERCIRCUITS:A.C. to A.C. Power Converters with Intermediate D.C. Link TH5 TH2
,/ I I






,C' TH2' supply rectifier

TIsynchronous machine

Figure 2.51 Converter-fed synchronous machine

As in the D.C. drives, the A.C. supply power factor is poor at low speeds. Full four-quadrant operation is possible without additional components. Special procedures are necessary for starting these drives because at standstill the machine voltage is not available to commutate the current. In essence this is usually achieved by momentarily switching off the D.C. link current every sixth of a cycle. This allows the thyristors in the inverter to turn off so that the next pair can be fired. Above approximately 5 per cent of rated speed the machine generates sufficient voltage for natural commutation and control is undertaken in a similar manner to that of a D.C. drive. Applications for this type of drive fall into two main categories. First, starting converters for large synchronous machines, the converter being rated only for a fraction of the machine rating. Second, as large high-power (and sometimes high-speed) variable-speed drives for a variety of applications. Power ratings typically from 1.5 to 30 MW at speeds up to 8000 r.p.m, are available. Also of import is the fact that high-voltage drives are offered with supply voltages up to typically 5 kV, but systems up to 25 kV are in service where the high-voltage converter technology is similar to that used for HVDC power converters.

upon the converter-fed synchronous machine drive having additional components to provide VAR compensation. Figure 2.52 shows a basic power circuit. The diagram somewhat belies the potential complexity of the VAR compensator. In its simplest form, this could comprise capacitors plus appropriate switches. Control of such a system is somewhat involved and it is often better to use a cycloconverter or even an auxiliary synchronous machine to provide the commutation and motor VARs. This system is only appropriate for high-power drives, generally above 4MW, where an induction motor is preferred.


A forced-commutated induction motor drive is the most widely used current source inverter at power levels in the range 50-3500 kW at voltages up to normally 690 V (highvoltage versions 3.3 kV/6.6 kV have been developed, however they have not proved to be economically attractive). Figure 2.53 shows the inverter and motor of the drive. The D.C. link current Id, taken from a stiff current source, is sequentially switched at the required frequency into the stator windings of the induction motor. The motor voltage waveform is approximately sinusoidal apart from the superposition of voltage spikes caused by the rise and fall of machine current at each commutation. Further distortion is caused by the effects of slot tipple and D.C. current tipple.


Unlike the synchronous machine, the induction motor is unable to provide the VARs or terminal voltage to commutate a converter connected to its terminals. Commercial schemes are available, however, which are closely based

II. . . . . . . . . . . . . . . . . . . . I r
r" . . . . . . . . . . . .

~VAR , compensator Ld .~ Id { _7


supply I


II 2



induction motor

Figure 2.52 Converter-fed induction motor

Chapter 2.2


TH 1

FH 3
,, || I|



Power can flow in only one direction via the diode bridge, which means that motoring torque can be developed only at subsynchronous speeds. For reverse running it is necessary to reverse the phase sequence of the stator supply. This drive can be very economic when designed for operation over a limited speed range below synchronous s p e e d this is the useful operating region for fans, pumps etc. The converter bridges required for such limited speed range operation need only be rated at a fraction of that of the machine which it is controlling. It is necessary in such designs to provide a starter, usually a resistance to run the motor up to the lowest controllable speed. This means that should there be a fault with the converter equipment, the system can be easily designed to run at full speed without the controller. Note that the supply harmonic currents and VARs associated with the converter part of the drive may be substantially reduced by adopting a limited speed range solution. The static Kramer drive finds application mainly at ratings between 1 and 20 MW, with induction motors with four or more poles (stability problems exist with two-pole motors, which can only be resolved with care). Speed ranges of 30 per cent are typical (i.e. 70-100 per cent rated speed). The induction motor stator can be wound for any conventional voltage e.g. 6.6 kV, 11 kV.

,) ~7D5 C'; 7D2

A ,L

\ 703
B J.


~ZD4 I I

7D 6





Figure 2.53 Forced-commutated induction motor

The operating frequency range is typically 5 to 60 Hz, the upper limit being set by the relatively slow commutation process. Special motors with low leakage inductance do offer advantage with this converter and allow reduced capacitance in the inverter and/or higher operating frequency. Below 5 Hz, torque pulsations can be problematic but PWM of the current can be used at low frequencies to ease the problem. This system is most commonly used for single-motor applications such as fans, pumps, extruders, compressors etc. where very good dynamic performance is not necessary and a supply power factor, which decreases with speed, is acceptable.


This final category of power converter converts the fixedfrequency, fixed-voltage A.C. supply to a variable frequency and/or variable voltage without an intermediate D.C. link.

The static Kramer drive is shown in Figure 2.54 and comprises a slip-ring (wound-rotor) induction motor together with an uncontrolled converter, D.C. smoothing reactor and a fully controlled converter in the rotor circuit. The diode bridge gives an output voltage Va that is proportional to the slip of the motor. Vd is opposed by the D.C. voltage of the fully controlled bridge, a small potential difference being sufficient to circulate current corresponding to the required load torque. Ideally, neglecting losses, the fully controlled bridge D.C. voltage sets the speed to which the motor will accelerate. Control is therefore very similar to that for a D.C. drive.

Soft Starter/Voltage Regulator

Figure 2.55 shows a typical soft start comprising inverse parallel connected thyristors in each supply line to an induction motor. Alternative connections are available but the principles are similar. The converter shown in Figure 2.55 is used to control the voltage applied to the motor and in this way soften the effects of switching an induction motor direct-on-line



Figure 2.54 Static Kramer drive


DRIVE CONVERTERCIRCUITS:Direct A.C. to A.C. Power Converters





Figure 2.55 Typical soft start Figure 2.56 Cycloconverter

(DOL). Although the converter will control the current drawn from the supply, its most usual application is in controlling torque to provide smooth jolt-free acceleration. Because the stator frequency is unchanged, a reduced running voltage, and hence flux, equates to a large slip which results in additional rotor losses - care must therefore be taken in its application. In a number of specialised cases, purpose-designed highresistance rotors (or slip-ring motors with external rotor resistors) are used to form a variable-speed drive - the rationale for such a system is based more on history than technology. More recently, such converters have been employed as combined soft starters/power-factor controllers/energysaving devices. The case for significant energy saving by this form of converter is often hard to prove. A crude form of frequency control is possible by modulating (varying cyclically) the thyristor firing angles at the required output frequency. Commercial systems are available but they are of limited value as supply current and motor torque are of poor quality.


J Ap J OAn

J Bp

J Gp



J oAm

J 08m

/0 Grn


A typical scheme for a cycloconverter drive is shown in Figure 2.56. Each motor phase is supplied, in effect, from a dual A.C. to D.C. converter which was described earlier. It is usual to employ circulating current-free converters. To avoid line to line short circuits, isolating transformers are used on the supply side. By modulating the firing angles of the dualbridge converters, a controllable three-phase set of voltages can be produced suitable for feeding polyphase A.C. machines. The drive is inherently four quadrant. The maximum output frequency is limited to approximately half the supply frequency by considerations related to harmonics in the motor currents and torque, stability and dimensions of the drive components. The cycloconverter therefore finds application in low-speed drives. The complexity of the drive also means that only high-power systems ( > 1 MW), or specialised applications (e.g. conveyor drives for use in hazardous environments), are economic. They are used on large ball

Figure 2.57 Matrix converter

mills, minewinders etc. They are also used to feed multimotor loads such as roller tables. Owing to the modulation of the converter firing angles, the harmonic content of the A.C. supply is complex and designs for appropriate harmonic filters somewhat involved. The cycloconverter is suitable for feeding both induction and synchronous machines. In specialised applications such as wind generators, cycloconverters have been placed in the rotor circuit of a slip-ring induction motor. Such a system, known as a static Scherbius drive, is detailed below.

Static Scherbius Drive

The static Scherbius drive is closely related to the static Kramer drive, with the single-quadrant diode bridge in the rotor circuit replaced by a cycloconverter.

Chapter 2.2
The cycloconverter is used as the voltage and frequency changer between the rotor and the supply. The cycloconverter is inherently regenerative, and the output can be controlled up to half the supply frequency in both phase sequences. It is thus possible for the system to operate as a full four-quadrant drive. For a given converter rating the range of speed control is therefore twice that of a static Kramer drive. The relative complexity of the drive limits its application to somewhat specialised high-power applications where a very limited speed range only is required and perhaps stringent harmonic current limits have been imposed by the supply authority.


new, recent advances in power devices offer the potential to overcome many of the drawbacks inherent in the circuit when the switches comprise inverse parallel thyristors. Limitations in the maximum output voltage (86 per cent input) means that its application in the commercial industrial market could be problematic. There are prospects for use in integrated motors and some servo systems where machine voltage is not seen as critical. Commercial systems are available only for very specialised applications at present. It has yet to be proven to be a practical and cost-effective industrial drive.

Matrix Converter
Recently, attention has been refocused on the matrix converter shown in Figure 2.57. Although the basic circuit is not


Speed and Position Feedback Devices

i m

1 2
3 4


90 92 92 97 99 100 101

@ B /I

s 6 7

The precise control of speed, position or acceleration requires appropriate measuring systems to be applied to the controlled variable. Although considerable progress has been made in the development of sensorless drive systems, these schemes tend to result in improvements in dynamic and shaft performance as opposed to precise positional or speed accuracy. This chapter deals with sensors extemal to the drive module itself.

Before considering the various forms of sensor, it is necessary to clarify the, often misunderstood, difference between resolution and accuracy of a feedback device:

The resolution of a feedback device is most simply described as the number of measuring steps in one revolution of the motor shaft. For an incremental encoder it is the number of pulses per revolution.



Resolution is often described in terms of number of bits. This is related to the twos complement; for example, a 12-bit feedback device has the equivalent of 212= 4096 measuring steps per revolution. For analogue feedback devices such as resolvers, where a resolution is quoted it usually refers to an associated resolver-to-digital converter.

shaft, the graduation, the optics and the electronic signal processing. For analogue feedback devices, the accuracy is influenced primarily by the winding distribution, eccentricity of the air gap, uniformity of the air-gap flux and the elasticity of the resolver shaft and its coupling to the motor shaft. When used, the accuracy of electronic signal processing can also heavily influence performance. Electronics can be used to improve accuracy. This is possible as many of the causes of error are fixed and correction techniques such as calibrated look-up tables can be used. To ensure smooth drive performance, a speed feedback device must give high resolution. At low speeds the position deviation of the encoder within one signal period, i.e. accuracy, affects speed stability.

The accuracy of a sensor is best described as the position deviation within one signal period or measuring step. In an encoder the accuracy is influenced primarily by the eccentricity of the graduation to the bearing, the elasticity of the encoder shaft and its coupling to the motor


The D.C. tachogenerator, also known as a tachometer or more commonly as a tacho, is an electromagnetic transducer that converts mechanical rotation into a D.C. output voltage. This voltage is directly proportional to rotational speed. In addition, it has a polarity sense dependent on the direction of rotation. The basic theory of the D.C. motor described in Chapter 1.1 applies to the tacho. There are, however, some aspects which differentiate the tacho from the generic D.C. motor: 1 Linearity of the output- the tacho must provide a D.C. output voltage proportional to the shaft speed with a defined linearity. Smooth output - the output voltage must be relatively smooth/free from ripple, particularly in the frequency range in which the drive is operating. The output voltage for a given speed should be constant with changing temperature.


The ripple voltage is defined as the variation in output voltage caused by the number of armature coils for a given design. Hence, the basic ripple frequency is dependent upon the number of coils and is a measure of the number of cycles in one revolution. The fundamental ripple can be distorted by intemal random and cyclic electrical noise, such as commutation due to varying contact resistance between the brush and commutator interface and brush bounce. In addition, misalignment of the driving medium and the output shaft of the tachometer would cause serious modulation in the output signal.

The tacho output is directly proportional to speed. This relationship can be expressed in the form:

t-~ O t~ " e(D O') 0 cO O



where Vg is the output voltage of the tacho, Kg is the transfer constant of the tacho and n is the speed of the tacho. Kg is often quoted in volts per 1000 min-1. There are two basic forms of D.C. tachogenerator, shuntwound and permanent-magnet types. Most modem tachos are of the permanent-magnet type, which are compact and robust.

shaft speed

Figure 3.1 Tachogenerator output voltage versus shaft


Chapter 3.1
To overcome the internal distortion, high-quality brushes are used, having low contact resistance and sufficient brush force to maintain positive contact stability without degrading the brush life. Enough cannot be said to ensure that the end user correctly aligns the drive and driven shafts because it becomes critical if a good signal is required at the lower speeds. The ripple voltage is usually measured in r.m.s, although peak to peak values are sometimes quoted. High-frequency brush ripple and its harmonics is rarely quoted in performance data since a minimum of electronic smoothing effectively removes these components without adding a significant time delay to transient events.


Linearity of output voltage is also affected by the output current drain. This current in turn depends on the voltage gradient of the tacho, the mechanical speed and the load impedance. A further impact could be armature reaction and saturation effects particularly in very high-speed applications. Linearity should not be confused with the tolerance given to the output voltage which is normally given as +10 per cent.


The stability of the output voltage is mainly determined by the state of the commutator surface. To maintain the commutator in good condition the recommended load RL should be typically between 50 and 500 times RA, where RA is the armature resistance of the tacho. The use of high values of RL, within the above limits, will improve linearity and stability, particularly under conditions of fluctuating load. The use of values below or above the limits can lead to excessive brush and commutator wear, or to increasing instability of the output. The ideal commutator condition is with a carbon track deposited on the commutator surface which has a shiny, dry appearance, evenly distributed around the commutator circumference. A thick deposit of oily or powdery carbon suggests that contamination has occurred and the output will be unstable. A bright, polished copper surface is not recommended and commutator cleaning sticks, abrasive powders or commutator lubricants must not be used. If cleaning becomes essential, the brushes must be removed, the armature may be rotated typically up to 1000min-1 and a hard felt pad moistened with a nonoily cleaning fluid held against the commutator surface to remove carbon and debris. It is important to allow the commutator surface and the intersegment insulation to dry thoroughly before reassembly, and to fit new brushes if possible since the old ones may well be impregnated with the harmful contaminant. In this case it is desirable to run the tacho at speed with a low value of load resistance to bed in the brushes.

Tachogenerators are specified in respect of their temperature error. The quality of this parameter is closely related to the design principle and materials used in the manufacture of the tacho. In order of increasing cost:

Uncompensated generators
This is the lowest cost type, having a temperature error up to 0.2 per cent per K of magnet temperature.

Compensated generators
This type has a temperature error in the area of 0.05 per cent per K of magnet temperature. Low-cost ceramic magnet materials are often used in combination with thermistorbased compensating networks. This temperature compensation system limits the application of this type of tacho to a temperature range up to 75C. The output impedance is generally higher than that of the more expensive types.

Stable generators
This type has a temperature error in the area of 0.02 per cent per K of magnet temperature. They are usually based on Alnico 5 magnets.

Ultrastable generators
This type has a temperature error in the area of 0.01 per cent per K of magnet temperature. To achieve this low level of temperature sensitivity Alnico 8 or Alnico 5 magnetic material in combination with a compensating alloy in the magnetic field circuit is used.

The maximum terminal voltage is a function of the bar to bar voltage seen at the commutator and a value of 12 volts bar to bar is typical to ensure stability in the output signal. To exceed this value will result in poor stability and, in the long term, destruction of the commutator bars and brushes due to the high self-induced currents in the windings under commutation. The maximum voltage is rarely quoted, rather the maximum speed.


Linearity is defined as the deviation from a straight line voltage/speed relationship at any given point within the speed range, as a percentage of the theoretical output at that point. A limiting factor for linearity is the speed at which the machine is rotating. A deviation will occur at speeds beyond the machine's capability, caused by aerodynamic lift of the brushes etc. and the hysteresis losses in the armature core.

Either voltage or commutation will be the limiting factor on the maximum operating speed. These values are normally given in the manufacturer's data sheets. Care should also be taken on very high-speed applications that the mechanical speed limit is not exceeded.


D.C. TACHOMETERGENERATOR:Mechanical Construction

brush assembly bearing

- - t(-.-------.-)

armature assembly

field assembly

Figure 3.2 Typical construction of a D.C. tachogenerator

The mechanical construction of tachogenerators varies considerably. Many are separate devices mounted on the outside of a motor or machine, and others are integral, some

even are wound onto the same armature as the motor they are monitoring. Figure 3.2 shows a typical arrangement used in many drive applications.


A.C. tachogenerators/tachometers generate a three-phase A.C. voltage proportional to speed, which is rectified into a D.C. voltage via an integral, usually three-phase, diode bridge. The polarity of the D.C. output voltage is not dependent upon the direction of rotation so can only be used on drives having only one direction of rotation. The advantage of such generators is that they are almost maintenance free, being of brushless design. The rectifier has a

linearity error of approximately 1.5 volts due to the forward voltage drop of two diodes. This error is essentially constant throughout the speed range. Output voltage ripple is typically in the order of 4 per cent. These are low-cost units with moderate performance used on unidirectional applications.


The resolver is a specific member of the synchro family. A synchro is a general term for a family of angular position

sensing devices which can operate together to provide a rotational torque, for light loads. Alternatively, they can provide

Chapter 3.3 (CX or TX) transmitter ,,

= ,R1 ,_,1 ~s3


(CT or TR) receiver S3 ~ N

supplyfor torquechain



angle @
Figure 3.3 A simple synchro

torquetransmitter (TX) $1 $3

"r Sl I '

i T S2 "', outputsignal ; ~ "'. for controlchain output " , angle

torquedifferential transmitter(TDX)

torquereceiver (TR)


(TX) inp

r TDXA L O / ~ . ~TgX T ~ ~ l . . ( T R , . s h a f f , stator~ R"~..kZ.k~' ~ "~l:pul: angle TXl r rTtX ~S2 S2~ ~_......I~ R2 R1 inputangle to TDXshaft ~S2 [rTtRri R2 R1

Figure 3.4 Torque transmitter~differential transmitter~receiver

a signal with direction and magnitude information enabling a servomotor system to provide rotational torque. A resolver is a modified form of synchro used for resolving angular position into coordinate data for use in control systems.

another similar three-phase stator, an alternating field will be set up in alignment with the first. A rotor within that stator field will have a voltage induced in it the amplitude and phase of which will be an indication of the input rotor position. If the output rotor is energised from the same source as the input rotor, it will develop an alignment torque.

All synchros and resolvers have the same basic construction, a wound rotor, carried on precision bearings, free to rotate inside a fixed, wound stator. Operation is based on the principle of a rotary transformer, so that when windings are in electrical alignment a maximum voltage is induced from the primary to the secondary. When the two windings are situated at right angles to each other, there is no coupling and no voltage is induced in the secondary winding. At any other angle a voltage will be generated which takes the form of a sine wave as the rotor shaft moves through 360 electrical degrees.

Torque synchros are applied when it is required to transmit light torque without additional servo components. Their use is normally limited to driving instrument pointers, compass cards and similar applications. Typical components used in systems are: torque transmitter (TX) torque receiver (TR) torque differential transmitter (TDX)

This simplified description ignores the fact that a single winding on a rotor and stator would not give angular rotation information, so the synchro family is based on a three-phase transmission system. The voltages induced in the three stator windings are displaced with respect to each other by 120 electrical degrees. If the three stator coils are connected to

These components may be used together in a torque chain as follows. From Figure 3.4 it can be seen that when the two rotors are energised from the same source and the stators are connected as shown, there will be a torque developed by interaction of the stator field and the rotor field of the receiver. The



TorqueSynchros ~ . rotor output Ii Vrasigdeu'm's') Ite (arl /7~

torque will be to bring the rotor of the receiver into alignment with that of the transmitter. When aligned, the torque is zero. The torque/misalignment curve is of sinusoidal form between zero and 180 with its maximum value at 90 . Typical alignment accuracy for torque synchro chains is + 1. Where this is not sufficiently accurate gear reduction systems may be used, and then the accuracy can be increased by the gear ratio, less a factor for gear error and backlash. Typical performance figures for torque synchros: transmitter accuracy: receiver accuracy: receiver/transmitter: differential transmitter: + + + 6 . . . 10 arc min 5/60 arc min 6/45 arc min 6 . . . 10 arc min

clockwine" rotatios

minimium couplng

cl "-counterockwise rotation

Figure 3.5 Effect of null output voltage

The control synchro is similar to the torque synchro but with high impedance windings, which reduce the system current loading, magnetic nonlinearity errors and temperature rise. This often enables smaller components to be used in a given application. The magnetic circuit is modified to provide uniformity of output, rather than torque output. The components of the control synchro are the control transmitter (CX), control transformer (CT) and control differential transmitter (CDX), and they may be used in the same way as torque synchros except that the rotor of the control transformer provides an electrical output signal as the input to a servomotor drive amplifier, which provides the error correction torque. When the shafts of CX and CT are aligned, a null voltage will appear at the rotor terminals of the CT. A slight deviation from this alignment will produce a signal in the rotor winding with phase relative to the reference voltage which will depend on the direction of deviation. The control transformer can therefore be considered a null detector and it is most often used in this way. However, the null is never zero, there is always a residual voltage present, as shown in Figure 3.5. This is due to stray coupling within the laminated stator that results in an in-phase voltage, a quadrature voltage, both at fundamental frequency, plus a number of harmonics. The level of this residual voltage is typically in the range 3 0 - 1 0 0 m V depending upon the supply voltage and individual specification, with some 50 to 80 per cent at fundamental frequency. Typical performance figures for control synchros: transformer accuracy: + 6 . . . 7arc min transmitter accuracy: + 6 . . . 8 arc min differential transmitter: + 6 . . . 10 arc min

R1, rotorI R3 stator

$4 I stator =$2 _.$1 -$3

Figure 3.6 The basic configuration of a resolver

Furthermore, the windings are displaced by 90 to each other compared to the 120 spacing of the synchro. The basic configuration is as shown in Figure 3.6. The prime purpose of a resolver is to provide Cartesian coordinate output signals from a polar coordinate input, plus the ability to add mechanical rotation of the resolver shaft. Depending on the control circuit requirements, resolvers may be supplied with one or two windings on the stator. The input may be fed to the rotor or the stator, although rotor-fed units are more usual. It is often convenient to build in a voltage reduction ratio between the primary and the secondary windings known as the transformation ratio (TR), to match into the next stage of the servo control electronics.

Computing Resolvers
Computing resolvers are primarily used for calculating trigonometrical functions and they are available with a feedback winding built in, or without. They are normally stator fed and the nonfeedback versions are for operating on substantially constant voltage sources. When there is a likelihood of the source voltage changing, feedback versions can be supplied. These are units with additional windings in the stator slots, in which a voltage is generated and compared with the input; the difference is amplified by a highgain amplifier and fed to the main windings. This feature compensates for variation in source voltage, and also reduces the effect of variation in frequency, winding temperature and load impedance.

The resolver is a special version of a synchro and has two windings on the stator and one or two on the rotor.

Chapter 3.3


Typical performance figures for computing resolvers: Nonfeedback phase shift: 2 . . . 2 0 null output: 1 0 . . . 60mV sine deviation: 0 . 1 . . . 0.25 per cent Feedback phase shift: 5 . . . 2 0 null output: 1 5 . . . 80mV sine deviation: 0 . 0 5 . . . 0.2 per cent

Multipole Resoivers
Multipole resolvers may be used where higher accuracies are required and emulate the geared systems previously used but without the additional hardware and associated gear errors. Typical performance figures for multipole resolvers: phase shift: 6 . . . +16 null output: 1 ... 15 mV electrical error: 3 ... 7 arc min

Phase Shifting
If a two-phase, or quadrature, supply is applied to a twophase stator winding the electromagnetic field in the air gap will rotate and generate in the rotor voltages of equal magnitude but 90 displaced. The phase relationship of the rotor to the stator voltages will be determined by the rotor angle. This is a useful technique for generating a supply of variable phase for a number of test applications.

A.C. Rotary Pickoffs

A.C. rotary pickoffs are transducers having a singlephase input and output. They normally operate on 400 Hz or higher frequencies, and the output voltage/rotor angle is reasonably linear. There are essentially two types, one having an angular range up to 25 and the other with an angle up to 65 . The narrow-angle version has four stator coils connected in a balanced bridge configuration as shown in Figure 3.8. The output voltage is the result of the impedance change within the magnetic circuit caused by the position/ rotation of the salient pole rotor. The voltage output sensitivity in volts per degree is proportional to the excitation voltage. For the wide-angle rotary pickoff, the primary excitation on the stator is separate from the secondary winding on the rotor. The rotor is eccentric in the stator bore resulting in a variation of flux linkage as the rotor rotates. The sensitivity is proportional to the transformation ratio between primary and secondary turns and may be made larger than that for the narrow-angle type. A centre tap can be provided in the secondary winding to give two balanced outputs, which are in antiphase. A typical output curve is given in Figure 3.10.

Brushless Resolvers
Originally all synchro devices employed slip rings and brushes to feed current to the rotor, but brushless resolvers have been developed in which current is fed to the rotor through a circular transformer mounted at the end of the unit. Generally, this technique is used only for rotor-fed units having a single rotor winding. Two rotor windings can be provided but a second transformer is then necessary and the construction can become uneconomic. Typical performance figures for brushless resolvers: phase shift: - 8 . . . +30 null output: 1 5 . . . 120mV electrical error: 3 . . . 15 arc min

resolver body

resolver stator assembly transformer stator / assembly

rotor assembly resolver rotor stack assembly
Figure 3.7 Sectioned view o f a brushless resolver

transformer rotor assembly


RESOLVER: Resolver

output red 1

Resolver-to-Digital Conversion
Of all angle-measuring transducers the resolver is inherently the most robust and stable in long-term performance and in relatively hostile operating environments. Where a digital output signal is required the analogue resolver can still be used in conjunction with resolver-to-digital (RJD) conversion electronics. R/D converters can be designed to accept input from both synchros and resolvers. An internal transformation circuit is used to convert synchro signals to a resolver/quadrature signal format. A number of techniques have been applied to the R/D conversion. Common is the use of single or double RC phase-shift circuits in which the zero crossing times of the two resolver format signals are compared. The difference is used to gate a clock, the frequency of which can be scaled to indicate the digitally coded angle. Another is the real-time function generator, in which two resolver format signals are applied to trigonometric function generators, providing an analogue output voltage that is integrated and digitised, then fed back to balance a bridge from which the value of the angle could be derived. An improved version is the ratio-bounded harmonic oscillator, although it suffers some disadvantages in that it is not a real-time measurement so, like the RC converters, it experiences staleness errors. With the introduction of monolithic semiconductor devices, ratiometric tracking converters were widely adopted. This technique is based upon a servo loop in which the converter tracks the input continually while there is any change taking place in the ratio of the sine and cosine signals. Since the system is ratiometric, neither changes of input voltage, nor voltage drop in the lines between resolver and converter are of significance. For high-accuracy R/D systems the repeatable nature of a resolver's error profile can be compensated for by means

black Figure 3.8 Narrow-angle A.C rotary pickoff, stator coil configuration

stator excitation R1 $1

"', L ~

~ bidirectional .]]~i tatin

~', " l i d "', $2

rotor output

Figure 3.9 Wide-angle A.C. rotary pickoff, stator coil configuration

typical output characteristics ~10~ 8

7 shaft GW degrees 3 a 80706050403020 1


5 4

I0 2 20304050607080 degrees 3 shaft CCW 4 5 6 output volts 7 in-phase 8 component

9 - 10

phase shift (o lead)

d 3.0

6 2.5
i_ c 0 ..~

Figure 3.10 Typical output curve of wide-angle A.C. rotary pickoff

t~ E ~ " "

2.0 1.5 1.0 0.5 0

transformation ratio ~,..versus frequency ( l e f t - h a n ~

/ ~

15 30 300 ~ 1Ok 30k lOOk Hz phase shift ~"~uency' versus frequency (right-hand scale) ~ 90 100

Typical performance figures for A.C. pickoffs: phase shift: 2 . . . 3 0 null output: 2 5 . . . 4 5 m V voltage sensitivity: 220...750mV/degree output linearity: 0.4... 3 per cent

80 Figure 3.11 Performance characteristics of resolvers with variations of temperature and carrier (supply) frequency

Chapter 3.4


of a look-up table. This technique can be applied to complex system installations or to simple R/D conversions to enhance the performance. Although it is often convenient to utilise a reference supply voltage and frequency in common with other servo components, operation at higher frequencies can be of benefit for R/D conversion. Many designs of converter operate in the

range 50 Hz to 20 kHz with a bandwidth of 1 kHz. It is of value therefore to illustrate the dynamic characteristic of a typical resolver over a range of frequencies and severe operating temperatures. From the typical curves in Figure 3.11 it can be seen that the optimum range is from 400 Hz to 5 kHz.


The encoder principle is simple: a photocell is positioned behind a slotted disc, or more commonly a transparent disc with lines photographically drawn on it. A light source shines through the disc into the cell. The photocell output is monitored as the disc is rotated with the encoder shaft. The number of pulses per revolution (p.p.r.) can range from tens to several thousand. The choice of encoder is important. Physically, the encoder must be compatible with the environment and also with the mechanical system. Encoders have the advantage that their output is in the form of a pulse train the frequency of which is not affected by temperature or attenuation of long cable runs as is the analogue signal of a tachogenerator or resolver. It is therefore potentially capable of contributing to extremely accurate digital speed control. Encoders can be classified into two primary types: (i) incremental encoders that produce digital signals which increase or decrease the measured value in incremental steps absolute encoders that produce a code value, which represents the absolute position directly

etching techniques. Imperfections in the gratings can be averaged out increasing the overall measuring accuracy and reliability. In practical encoders a second signal is produced phase shifted from the first by 90 . These signals are then digitised (squared) and fed into a counter in which counting pulses are derived from the square-wave edges. Simultaneously, the square wave signals are applied to a direction-determining circuit, which transmits the counting pulses to the positive or negative port of the bidirectional counter. This method of determining linear or angular displacement by counting pulses is commonly referred to as the incremental measuring method. Typically, the finest grating used is 8 Finer gratings can be manufactured but cannot be scanned in the same manner as it would be necessary to maintain too high a tolerance on the gap between the two gratings. Scanning signals of a relatively coarse grating are therefore fed into a circuit which interpolates between signals increasing the apparent resolution by a factor of, typically, five. The encoder produces a train of pulses the frequency of which is proportional to speed. The direction of rotation can be determined if necessary by pulses from a second photocell, displaced 90 in phase from the first, as shown in Figure 3.12.


These two types of encoder have somewhat different technologies at their heart and may therefore be considered separately.

The basic technology of incremental encoders is often referred to as the Moir6 principle - a photoelectric scanning method to produce periodic signals using two fine gratings, which are positioned closely to each other and have approximately parallel and equally spaced lines. If the gratings are moved relative to each other periodic fluctuations in brightness can be seen. These fluctuations are converted to electrical signals via photoelectric sensors. Using this method it is possible to scan very fine graduations which can be produced very accurately using photographic



I !





clockwise rotation

I b I I






anticlockwise rotation

Figure 3.12 Pulse trains corresponding to bidirectional

rotation o f an encoder


ENCODER: Incremental Encoder

Incremental encoders are available in the following common forms: Incremental - readily available from a wide variety of suppliers. Almost any line count available up to 5000 per revolution. Special line counts and output options are easily obtained. Incremental with block commutation signals - again, readily available but somewhat constrained by lack of industry standards. Mounting configuration, signal conditioning and power supplies vary widely. Available in line counts up to 8000, plus block commutation signals (120 elect blocks) for two, four, six, and eight-pole motors. They are being developed in both hollow-shaft and modular versions by a variety of encoder suppliers. Incremental with sine wave c o m m u t a t i o n - this type of encoder generally has sinusoidal quadrature outputs, with a 1 volt peak-peak amplitude. Commutation is accomplished using a quadrature one cycle per revolution output. Marker pulses are available on incremental encoders to give a precise mechanical reference position. This is used when the absolute position of the motor shaft is required. In such cases at start up of a machine a procedure often referred to as homing is undertaken which typically requires the motor to
m m ml~m ~m H m,lmm~m mi~m E ~ m mm ml irJI ml;m mm~ ~ ~

be turned at low speed by one revolution or until the marker pulse is located. Where it is necessary to know the absolute position at tum on and without rotating the motor, it is possible to have a battery-backed encoder and control electronics, or to use an absolute encoder.

In absolute encoders the value of the actual position is immediately measured when the system is switched on. Absolute encoders do not therefore need a counter since the measured value is derived directly from the graduation pattern. In most cases the output from the encoder is in the form of a pure binary code or in Gray code. A further coding exists which is a direct derivative of the Gray code, and is called the Gray excess code. This comprises a section from the middle of a Gray code pattern and permits the transmission of other than 2 and yet remains a unit-distance code. In the example in Figure 3.15 where the last two code values have been omitted from the graduation pattern, to give 28-position resolution, the code would be described as a 28-excess-3 Gray code.
~ ~ ~ r~4 w l ~ ~ ~ ~ ~ ~ ~ W

mhl m[a mu ~

m m m m m m m m m m m m m m m mm mm mm mm mm mm mm mm mmmm mmmm mmmm mmmm mmmmmmmm

Figure 3.13 Pure (natural) binary code

mm mm mmmm

mm nm mmmm

mm mm mmmm

mm mm mmmm

Figure 3.14 Gray code

28 positions

Figure 3.15 Gray excess code

Chapter 3.5


The optical arrangement of a sin/cos encoder is similar to that of an incremental encoder. They provide the differential analogue output signals ( i A, 4- B, and 4-R) equivalent to those of an incremental encoder but of sinusoidal wave shape, and of magnitude typically 1 V peak to peak. The sinusodial wave shape allows high-resolution position determination. Direction and speed of rotation is detected as in the incremental encoder. The angular position can be determined by knowing the line count of the encoder and, when between two consecutive increments or lines, deriving the phase from the analogue signals A and B. The reference mark R provides absolute position determination, if the angle at which the encoder is mounted is known. The incremental count and hence the incremental position can be determined by a timer that counts up when A is the leading sequence and counts down when B is the leading sequence. When digitised, both edges of A and B are counted, thus the incremental position t~incr is given by

where incr is the timer count or incremental count, N is the line count of the encoder and ff~0is the zero position. One incremental step is equivalent to a 90 (electrical) phase shift of the signals, A and B. The phase ~ of the sinusoidal signals A and B can be used to interpolate the position between two consecutive line counts or four incremental steps, which are equivalent to each other. It can be calculated as ~5 = 90 + arctan (B/A), q~ = 270 + arctan(B/A), if A > 0 if A < 0

which has the advantage that the absolute amplitudes of A and B, which are a common function of the encoder's rotation speed and supply voltage, do not affect the ratio and hence the angle. Since the arctan-function is ambiguous, one has to check the sign of the sinusoidal signals A and B to identify the correct quadrant. It is this ability to interpolate between the line counts associated with incremental encoders which gives the sin/cos encoder such an advantage in applications where high resolution is required. Care must be taken in the practical implementation of sin/cos encoders due to the low signal levels which are commonly used.

~incr-- [360/(4 x N) x incr] + ~o


The combination of any motor drive package must consider a number of criteria including: Absolute versus incremental p o s i t i o n - the need for absolute or incremental position information is generally application dependent, and many drive systems can develop or handle either type of interface. Tachometers inherently give speed information and do not give position. Position can only be derived by integrating the speed signal. Resolvers inherently provide absolute feedback. In cases where absolute position information is needed, e.g. for robotics applications, a resolver can therefore be a good choice. Rotary encoders can provide either incremental or absolute outputs and, in fact, most absolute encoders now provide both.

High-speed o p e r a t i o n - any feedback system will generate data based upon the resolution of the sensor, multiplied by the interpolation factor of the interface electronics (interpolation is a method for producing measurement steps which are smaller than a quarter of the scanning signal period). For most resolver-based systems, the feedback element provides one cycle of output per revolution and so is an absolute position device. Because the position information is produced by amplitude modulation of a carrier signal, it usually needs to be converted to a digital format before it can be used by a modem motor controller. Resolver-to-digital converters (RDC) interpolate the resolver output signals and provide typically 10, 12, 14, or 16-bit results, depending upon the converter used. When coupled to a 12-bit RDC, a position measurement resolution of 360o/212 or 5.27 arc minutes is obtained.



For an encoder, the maximum speed is limited primarily by the frequency response of the sensor. Encoders generally have a maximum frequency response capability in the range 200-400 kHz, which allows a 4096 cycle per revolution encoder to turn at 6000min-1 without missing counts. Higher performance is available at higher cost, with a maximum frequency response capability of approximately 1 MHz. Low-speed o p e r a t i o n - low-speed operation requires high resolution and accuracy from the sensor. Many encoder-based drive systems have simply selected the encoder resolution/pulse number based upon the maximum speed of the application and the frequency response of the encoder input of the drive. For example, many drives use a 2000 pulse per revolution encoder with a 200 kHz maximum frequency input capability of the drive. This allows the motor to run at 6000 min-1 without missing counts. In a digital system, where encoder pulses are counted over a defined time period, quantisation effects result in what can best be described as a ripple on the speed feedback signal. Whilst this occurs at all speeds, the effect (percentage ripple) is higher at low speeds. Whilst this is rarely a practical problem, in some applications it can be important. In such applications a sine-cosine encoder should be used. d Motor efficiency- in permanent magnet servodrives where the switching of current to the machine windings is determined by information from the position sensor, motor efficiency is a function of the accuracy of the position information. Single cycle sine-wave-type commutation signals generally have accuracies to about 5 mechanical.

Hall effect switches are in the same range, and are typically very difficult to align. Resolver and optical encoder commutation can be good to 10 arc min with some careful planning, and they are very simple to align. e Accuracy- in general, optical encoders have an advantage in this area. Specially fabricated resolvers can be obtained which are accurate to 4-2 arc min, but in general resolvers have accuracy ratings of 4-2 to 4- 20 arc min. The resolver-to-digital converter (RDC) adds an uncertainty of 2 to 4- 8 arc min. Resolver errors also have both static and dynamic contributions, which result from the acceleration error in the RDC tracking loop, offset voltages that are uncompensated, phase shift between the signals and the reference voltage, and capacitive or inductive crosstalk between the resolver signals and the reference cabling. Noise in the interconnect or on the reference will generate speed-dependent errors proportional to the phase shift in reference and inversely proportional to the reference frequency. Modular incremental rotary encoders usually have an accuracy rating of two arc minutes or better over all operating conditions. High-temperature operation - resolvers are capable of operating at high temperatures, and in many cases can be operated at up to 150C. Standard encoders are typically specified for use up to 100C... 125C. Although the encoder is usually mounted within the motor housing, many measures are normally taken to thermally isolate it from the motor core. As a result, it is usually not necessary to require a feedback device to be operated above 100C.


Speed and position feedback devices are obviously subject to the acceleration and speed associated with the motor or load that they are measuring.

angular acceleration 104 rad s -2 max acceleration of 1000 m s -2 for shock and impact valid for 11 ms vibration resistance 100 m s -2 from 50 to 2000 Hz Maximum speed is generally related to bearing life and forms one of the selection criteria for specific applications.

Moment of inertia of the rotor and the coupling of the feedback device form a single vibrating mass system with a natural frequency, which needs to be outside the operation frequency range of the drive. A large number of different coupling types exist, but in general a diaphragm type is usually preferred to give the highest natural frequency. The natural frequency is given by the equation: f = (1/27c). N(C/I)

Chapter 3.7


where f is the natural frequency in Hz, C is the torsional rigidity of the coupling (Nmrad -1) and I is the moment of inertia of the sensor motor (kg m2). It is interesting to note that although many suppliers quote current loop torque bandwidths of servo amplifiers of 2 kHz and above, the natural frequency of the very best coupling is approximately 2 kHz. It is not possible to take advantage of better converter performance.

No one can discount the sturdiness of the resolver. It is a simple device with a similar make up to the motor, consisting of windings, bearings etc. However, for the majority of the environments encountered, an encoder is completely

adequate. Because both encoders and resolvers can be purchased with or without bearings, neither can claim an advantage in this respect; however, frameless resolvers may have slightly less sensitivity to axial play than a modular encoder would. With respect to electronics, it is true that the resolver electronics can be mounted remotely from the motor, in a less extreme environment. However, they are much more complex than those of the encoder. Typical encoder designs use a small number of very basic components. It is possible that for extremely high-impact shock environments a resolver could claim some advantage over an encoder. However, if resolutions of less than 1000 counts per revolution are desired then an encoder with a metal or mylar code wheel can compete favourably with resolver designs.


Absolute measuring system -

the measuring value is determined by reading information from a scale, without counting. The measuring value is immediately available after switch on.

Encoder - apparatus consisting of a measuring standard and

a scanning unit (transducer, sensor).

Gray code - unit-distance code system in which only one code signal changes with the transition from one measuring step to the next. Incremental measuring system - measuring method by which the measuring value is derived by the summation (counting) of increments (measuring steps). Integral coupling - innovative angle encoder design with built-in coupling, located preferably on the stator side. Interferential measuring system - photoelectric measuring

Accuracy grade - grade of quality, determined by the

maximum permissible measuring deviations within a predetermined measuring range (e.g. 1 m).
Amplitude evaluation - method of evaluating signals gen-

erated by dynamic scanning (with cartier frequency): the amplitude variation of two alternating voltages of the same frequency is used to determine the measuring value.
Angle encoder - angle-measuring device, converts the shaft

rotation angle into electrical signals (can be incremental or absolute).

Carrier frequency method - scanning method used mainly

system with a phase grating scale where scanning signals are produced via the interference of diffracted beams.
Interpolation - method for producing measuring incre-

with magnetic and inductive measuring systems (see dynamic scanning).

Direction discriminator - part of a bidirectional counter, which determines the counting direction. Distance-coded reference marks - incremental measuring method, whereby the absolute position can be determined by evaluating the systematically varying distances between consecutive reference marks. Dynamic scanning - scanning method by which two alter-

ments which are smaller than a fourth of the scanning signal period.
Measuring system - consists of an encoder and associated

electronics incorporating interpolation, counter, readout and/or data interface.

Modular angle encoder - angle encoder, consisting essen-

tially of disk and scanning unit assemblies (rotor and stator) which are integrated into a machine or a rotary table.
Moir~ principle - photoelectric scanning method to produce periodic signals using two fine gratings, which are closely positioned to each other and have approximately parallel and equally spaced lines. Multiturn rotary encoder - absolute rotary encoder which

nating signals of constant amplitude and slightly different frequencies are generated and where the phase between the two signals represents the measuring value.
Eccentricity e r r o r - measuring error of an angle encoder

caused by an eccentricity in the mounting of the circular graduation.

determines the angular position of the shaft and the number of shaft rotations.



Phase evaluation - method of determining position by

detecting the phase between alternating voltages having a slight variation in frequency. Phase grating scale - scale with step grating which diffracts the transmitted or reflected light into two or more orders.
Radian - standard unit of angle: the angle at which the arc of

Resolver - inductive angle-measuring device, producing two alternating voltages the amplitudes or phases of which depend on the (shaft) rotation angle. Reversal error - measuring error which results from approaching a position from different directions.
Scanning frequency - response level which limits the velocity of an incremental measuring system. Static scanning - scanning method, which generates peri-

circle has the same length as the radius.

Reference m a r k - random graduation pattern which, when

traversed over, produces a signal peak, which may be used to determine an absolute datum within an incremental measuring system.
Reference pulse - square-wave signal produced when the

odic signals during movement. The signal periods and fractions thereof correspond to a definite linear or angular displacement.
Systematic error - reproducible measuring deviation, which

scale reference mark is traversed over; normally one measuring step wide; may be used to define an absolute datum within an incremental measuring system. measuring step, smallest digital unit of the measuring value.

can be compensated for by e.g. computation.

Torsional stiffness - rotational rigidity of a precision coupling governing the reversal error of a rotary encoder.


Drive Control

iiiiiiiiiiiiii~iiiiiii !iiiiiiiiiiiiiiiiiii!ii!i ~'~'~"~

1 2


103 105 114 116

:,ii!iii iii!iiiiiiiiiii!iiiiiiii!




Many applications exist where something has to be controlled to follow a reference quantity. For example, the speed of a large motor may be set from a low-power control signal. This can be done using a variable-speed drive as shown in Figure 4.1. Ideally, the relationship between the reference and the motor speed should be linear and the control system should respond instantly to changes in the reference. Any control system can be represented, as shown in Figure 4.1b, with an input reference signal, in this case a speed reference, a transfer function G and an output, in this case the speed of rotation of the motor shaft W. For the system to be ideal, the transfer function G would be a simple constant, so that the output is proportional to the reference with no delay.

Unfortunately, the transfer function of many systems is not a constant and so, without any form of feedback from the

mains power supply

speed reference

] v=r,=b,e-I




speed b speed reference ,.]



Figure 4.1 Variable-speed drive and motor


GENERAL: O p e n - L o o p C o n t r o l

output to correct for the nonideal nature of the transfer function, the output does not remain proportional to the control signal. Using an induction motor supplied by a simple open-loop variable-speed drive as an example, the following are some unwanted effects that can occur in control systems.

The simple open-loop drive of the previous section is replaced with a control system as shown in Figure 4.2. The speed of the motor shaft is measured and compared with the speed reference to give a speed error. The error is modified by a transfer function F to give a current reference i* at the input to the current control block. Various methods of current control for motors are discussed in the next section; however, for now it should be assumed that the motor current can be controlled to give a torque which is proportional to the current reference. If the speed of the motor varies from the reference level a speed error is produced and the torque applied to the load is modified to bring the speed back to the reference level. It is necessary to choose a suitable transfer function F to obtain the required performance from the closed-loop control system. The function could be a simple gain, therefore the current reference i*= Kp x Werr. This would give some degree of control over the output speed, but the speed error must have a nonzero value if any torque is required to hold the motor speed. If the speed error is not zero, then the speed would not be at the required reference level, and so the speed would vary with load. By adding an integral term so that the current reference i*= Kp x Werr+ Kifwerrdt, it is no longer necessary to have any speed error even when torque is required to drive the load at the reference speed. The integral term accumulates any speed error over time and builds up a current reference to provide the necessary torque. A closedloop control system with proportional and integral terms is called a PI controller. Although there are many types of closed-loop controller, the PI controller is the most commonly used because it is simple to implement, relatively easy to set up and well understood by most engineers.

The output of a simple open-loop drive can be a fixed frequency, which is proportional to the speed demand signal. Therefore, the frequency applied to the motor remains constant, for a constant speed demand. However, the speed of the motor drops as load is applied because of the characteristic slip of an induction motor, and so the speed does not remain at the demanded level.

It is possible under certain load conditions and at certain frequencies for the motor speed to oscillate around the required speed even though the applied frequency is constant. Another major source of instability in rotating mechanical systems is low loss elastic couplings and shafts.

There are many possible sources of nonlinearity; if, for example, the motor is connected to a gearbox the speed at the output of the gearbox could be affected by backlash between the gears.

Variations with temperature

Some aspects of the system transfer function may vary with temperature, for example the slip of an induction motor increases as the motor heats up, and so for a given load the motor speed may reduce from the starting speed when the motor was cold.

With a simple open-loop inverter and induction motor there can be a delay before the motor speed reaches the demanded level after a change in the demand. In very simple applications, e.g. controlling the speed of a conveyor belt, this type of delay may not be a problem. In more complex systems, such as machine tools, delays have a significant effect on the quality of the system. These are some of the unwanted effects that can be produced if an open-loop control system is used. One method that can be used to improve the quality of the controller is to use a measure of the output quantity to apply some feedback to give closed-loop control.


The step response is one method of assessing the ability of a closed-loop control system to follow a step change in the reference. Some example step responses are shown in Figure 4.3 for a simple second-order system. If the output reaches the reference in the shortest possible time without any overshoot the response is described as being critically damped. If overshoot is not acceptable, then this represents the best possible response giving the minimum delay between the input and the output of the system. If the system damping is increased the response

speed error

speed reference

current reference i*




W~ ~ *

current control



Figure 4.2 Closed-loop control system

Chapter 4.2


~ ,~nder damped

dB -10 -15 -20
~: ed

100 III

,, 1000





-25 -30 010 -10


- "t3"~
1oo "~


rad s-1



-20 -30 deg-40 -50 -60 -70 -80 -90


Figure 4.3 Step responses

becomes slower and is described as being over damped. Alternatively, if the system is under damped the response includes some overshoot and may oscillate about the required reference before settling. These results, which are for a simple second-order system, show that increasing the damping reduces overshoot and slows down the system response. As will be demonstrated in Chapter 4.4, real systems can be more complex and increasing the damping does not always give this result. The step response may be the closed-loop response of the system, where a change to the reference level in the minimum time is required. Alternatively, the step response may show the change of output to some other stimulus, such as a load torque transient. In this case the response should be as small as possible. The step response can be used to assess the controller performance when it is used in isolation. However, if the controller itself is to be included within the closed control loop of another system, the gain and delay of the original system are important as they affect the performance of the outer system. The gain and delay can be measured by producing a bode plot showing the gain and frequency response as shown in Figure 4.4. An ideal controller would have unity gain and zero phase at all frequencies; however, in a real system as shown in Figure 4.4 the gain reduces and the phase delay increases at higher frequencies. A measure of these effects is the bandwidth

rad s-1

Figure 4.4 Bode plot of system gain and phase

of the system, which is often defined as the - 3 dB point for the gain characteristic. In the example this occurs at 670 rad s -1. The corresponding phase delay varies depending on the order of the system. A first-order system has a delay of 45 at the - 3 dB point, whereas a second-order system like the one in the example has a delay of 60 at the - 3 dB point. The transport delays associated with digital systems can further increase the delay at the - 3 dB point. In many cases the bandwidth is quoted as an indication of the dynamic performance of a control system, i.e. the higher the bandwidth the better the system performance. This bandwidth is usually quoted as the frequency at the - 3 dB point of the gain characteristic, and particularly with a digital implementation may give no indication of the quality of the control system at all. If the controller is to be included within another closed-loop control system the phase delay is important. If the delay is too large it may be necessary to detune the outer loop to maintain stability. The amount of overshoot in the step response is also important in many applications. Increasing the frequency of the - 3 dB point of the gain characteristic may result in unacceptable overshoot.


There are many types of variable-speed drive each suited to different applications or for operation with different types of motor. In this section descriptions are

given of typical control systems for a range of different types of variable-speed drive operating with A.C. motors.


A.C. MOTORDRIVECONTROL:General-Purpose Open-Loop A.C. Drive

vectors (U 1 to U6) , and the remaining two states (Uo and U7) give zero voltage vectors where all three output phases are at the same voltage level. By switching rapidly between the various states the average output from the inverter can be a voltage vector at any angle with the required magnitude. In the example given in Figure 4.6 a voltage vector at angle c~ could be synthesised by producing u~ and U2 each for 50 per cent of the time. Changing the ratio of the time during which u~ and I!2 are active allows the vector to be moved from u 1 to u2 with a locus which forms one side of a hexagon connecting the tips oful and u2. The magnitude of the vector can be controlled to be anywhere inside the hexagon connecting the tips of the active vectors by introducing periods of zero voltage with vectors Uo and u7. The active vector periods and inverter output waveforms for the required vector Us are shown in Figure 4.6c. A space-vector modulator produces sinusoidal output voltages provided that the vector locus is a circle. The limit for sinusoidal operation is a circular locus that just fits inside the hexagon in Figure 4.6b. Operation outside this circular locus is over modulation and, although the fundamental output voltage can be increased, odd harmonics of the fundamental output voltage are also produced. At full over modulation, i.e. when the locus of the voltage vector is the hexagon shown in Figure 4.6b, the average output voltages are trapezoidal waveforms. It is possible to further increase the output voltage by making the angular velocity of the voltage vector vary as it moves round until the output voltages are square waves. The fundamental output voltage for these ranges is given in Table 4.1 as a ratio between the drive supply voltage and the output voltage for an ideal drive, which has a simple diode rectifier at its input. The waveforms in Figure 4.6c show that an upper or lower power device in each output phase of the inverter is always on. This may be an IGBT or a diode depending on the direction of the output current from each phase. When the control system switches off an IGBT it must allow a short time, the safety margin, before turning on the other device in the same phase to prevent short pulses of current flowing from the positive to the negative rails of the inverter through both devices in that phase. During this safety margin, the dead time, the output voltage of the inverter is undefined as it is dependent on the direction of the output current. Although the dead time is short it distorts the inverter output voltage and has the most significant effect at low modulation


The open-loop A.C. drive is a high-power variable-frequency voltage source. In its simplest form, the output frequency is defined by the user's reference and a suitable fixed frequency to voltage characteristic used to define the output voltage. Although this type of drive is normally designed to supply one or more induction motors connected in parallel, it can also supply other types of A.C. motor or it can be used as a variable-frequency/variable-voltage power supply. The following description relates to the operation of an openloop A.C. drive with an induction motor. Figure 4.5 shows a block diagram of a typical general-purpose open-loop drive. Each of the blocks within the control system is described in the following sections. For this type of control system feedback is required from the output current of at least two phases of the inverter and the voltage from the D.C. side of the inverter. These feedback signals can be derived from within the drive itself, and so no external feedback, such as the motor speed or position, is required. The control system allows a simple open-loop vector strategy to be implemented. In common with all other A.C. drive control strategies described in this chapter the control system is based on a reference frame as described later.

Space-vector Modulator and Inverter

The power circuit is a diode rectifier and voltage source PWM inverter. The space-vector modulator converts a modulation depth, m, and modulation angle, Om, into control signals to turn the six devices in the inverter on or off. The modulation depth defines the output voltage from the inverter as a proportion of the range available without over modulation, and the modulation angle defines the angle of the space-vector representation of the three-phase voltages at the inverter terminals. The modulator is called a spacevector modulator because it applies voltages to the three phase windings of a motor to give a voltage vector in space with respect to the body of the motor. This voltage vector is conceptual rather than real as voltage does not exist as a space quantity, such as, for example, flux. There are eight possible states for the devices in the inverter, excluding those that result in shoot through, as shown in Figure 4.6. These states give six possible active voltage


I YI i

voltage I characteristic Vs I I I

~ R \


~ ~

' ' vector

Y ,._1 vl


modulator and ]

inverter ]

[~ f IT T

"r Oref(reference ~rframeangle) I xy "~J.~ 'sQI D Q ' ~ , Isv




Figure 4.5 Open-loop drive control system

Chapter 4.2


depths. It is possible to reduce this effect by using dead-time compensation to modify the widths of the IGBT control signals depending on the direction of the inverter output currents. The modulation depth does not define the absolute output voltage of the inverter because the D.C. voltage at the input to the inverter can vary with drive supply voltage or if the drive operates in braking mode. Therefore the required voltage magnitude [v*[ must be converted to modulation using the following equation:

control system used with this type of drive only ensures that the reference frame aligns with the flux under steadystate and not during transient conditions. By using this technique the voltages and currents appear as D.C. quantities under steady-state operation, even though the motor currents and voltages are A.C. quantities. The x and y axis currents relate to the production of flux and torque in the motor, respectively. A reference frame is simply a set of axes as shown in Figure 4.7. The method used to calculate the position of the reference frame and to control the motor voltage is covered in the next section. First, in this section, the translation of the current feedback into the required reference frame and the translation of the voltage references from the reference frame into an angle and voltage for the modulator are described. The current feedback is taken from two of the three inverter output phases (is, and isv). These can be converted into twophase quadrature currents (isD and isQ) and then into current components in the x and y axis of the required reference frame as follows"


v/3 x v*l/V~.c.


where VD.C. is the D.C. link voltage and Iv*l is scaled to a value equivalent to the peak phase voltage.

Reference-frame Translation
It is helpful to apply voltages and measure currents in a reference frame aligned with some quantity in the motor. In an open-loop drive the reference frame is usually aligned with the stator flux. As will be explained later, the type of

isD- isu
isQ -- (i,, + 2isv)/v/3

(4.2) (4.3) (4.4) (4.5)


~_J ~V+ ++J ?+

u3(U LVUW L) u2(UuVuWL)

t'~ -- i~D COSOref + isQ sin O,.~f


isy -- isQ cos Oref - isO sin Oref

where 0ref is the angle of the required reference frame.

inputrectifier D.C.linkand brakingchopper

The voltage reference components (vsx and Vsy) can be converted to a magnitude and angle by simple rectangular to polar conversion.

V s l - v/(v~2 + v~ 2)

(4.6) (4.7)

Ov -- tan -1 (V~y/V~x)

uo(U LV LWLi uT(UuVuWu) u5(U Lv LW U) I i i , i I i i , i I i i , |

ul (UuVLWL)
u6(UuVLW U)

Reference-frame Generation
It is a convention with open-loop drives to use the synchronous frequency related to the required speed of the motor as the reference, i.e. 50 Hz for 1500 min -1 using a four-pole motor. The reference shown in Figure 4.5, f*, is compensated with a frequency from the slip compensation




U7 U7




Figure 4.6 Space-vector modulation

a power circuit b space vectors c inverter output voltages

Vsy ~ R s i s y

Table 4.1 Fundamental output voltages

Range Space-vector modulation without over modulation Space-vector modulation with over modulation Square-wave operation Maximum output/input voltage ratio
1.00 1.05



stator flux (in steady state)


Figure 4.7 Stator flux reference frame


A.C. MOTOR DRIVE CONTROL: General-Purpose Open-Loop A.C. Drive

block. When the reference frame is aligned with the stator flux the y axis current, i,y, is a measure of motor torque, and so the slip compensation block can estimate the change of motor speed with load applied due to motor slip (see Chapter 1.2). By adding the compensation value to the frequency reference, the reference is increased as load is applied to the motor, f~omp is integrated to give the reference frame angle Ore and so the reference frame rotates at frequency fcomp. If fi fixed values of V~x and V~y were used the drive would apply voltages to the motor at the frequency f~omp, which should hold the motor speed constant even when a load is applied. The frequency adjustment applied by the slip compensation block may not be very accurate as the motor slip varies with temperature, but this method gives moderate control of speed without any form of motor position or speed feedback. The motor stator flux is defined as: ~s -- / Vs - Rsisdt (4.8)

the drive when fast acceleration is required or a large transient load is applied to the motor. Therefore, a current-limit system is required to decrease the frequency reference when a large decelerating load is applied, or increase the frequency when a large accelerating load is applied. Figure 4.8 shows a current-limit system commonly used with openloop drives. The frequency reference, f*, is derived from the user frequency reference via a ramp block which limits the rate of change of frequency and hence the rate of change of motor speed. If the torque-producing current in the motor exceeds the current limit then the PI controller is enabled to modify the frequency reference. If the motor is producing too much accelerating torque because a large load has been applied the frequency reference is reduced towards zero and unless the load reduces with speed the motor stalls. If the motor is producing too much accelerating torque because the ramp rate is too fast the PI controller reduces the rate of change of frequency reference so that the current is limited because the acceleration ramp is extended. If the decelerating torque is too large because a large accelerating load has been applied or the decrease in frequency reference is too fast, the PI controller works in the opposite way and either accelerates the motor or reduces the deceleration rate.

where qDs is the stator flux, vs is the voltage vector representing the motor terminal voltages and is is a vector representing the motor currents. In the steady state the voltage vector applied to produce the stator flux leads the flux by 90 and is given by:
V~s - - Vs -



Performance and Applications

The performance characteristics of the open-loop drive control system applied to an induction motor can be summarised as: Moderate transient performance. Full torque production down to approximately 3 per cent of rated motor speed. Although a good estimate of stator resistance (R,) improves torque production at low speeds, the control system will work with an inaccurate estimate, albeit with reduced torque. The stator resistance can be measured by the drive with a simple test. Although a good estimate of motor slip improves the ability of the drive to hold the reference speed, the control system will work with an inaccurate estimate, albeit with poorer speed holding. The motor slip depends on the rotor time constant of the motor (Tr) and this cannot be measured easily. No position or speed feedback is required from the motor shaft.

In Figure 4.7, V~x - Rsisx and V~y = v~s + Rsisy. If V;x and V~y are derived in this way then the x axis of the reference frame is aligned with the stator flux, and the y-axis current can be used as a measure of load for slip compensation etc. For constant stator flux, Iv l must be proportional to frequency up to the rated frequency of the motor. Above rated frequency, Iv ,l is held constant so that the flux falls with increasing frequency to give flux-weakening operation. This method of control gives good and robust motor control down to approximately 3 per cent of rated motor frequency. The motor speed can be controlled below this frequency, but usually at reduced torque. This method gives significantly better performance than the fixed-voltage control methods described in Chapter 1.2.

Current Limit
The motor current cannot be allowed to increase with load without any limit or the inverter protection system will trip

user frequency reference


maximum allowed value of isy


current above limit

Figure 4.8 Current-limit system

Chapter 4.2


This type of drive is used in many applications where moderate performance is required and where providing position feedback would be unacceptable because of the environment or cost, or is simply not necessary. The following are some examples of applications where open-loop drives are used: fans and pumps conveyors centrifuges

the torque is proportional to i~y. The rotor flux position is easy to calculate from the absolute rotor position as shown below. It is clear that the absolute position within each electrical revolution is required, and so the position feedback device must give some absolute position information. Traditionally, resolvers have been used as they can give the absolute position of the rotor (see Chapter 3.3), but these are being replaced by incremental encoders which give more precise feedback with lower cost interface electronics in the drive. However, an incremental encoder does not give the absolute position of the rotor, and so additional Gray code signals are required to find the absolute position at power up. For a three-phase six-pole motor a three bit Gray code is repeated three times per mechanical revolution giving the position to within 60 during each electrical revolution. Provided that the angle between the Gray code signals and the magnets is known the reference frame can be placed within +30 of the correct position. Once the motor has rotated past a change from one Gray code value to another, the position is known exactly and the incremental signals can be used to track the absolute position until the drive is powered down again. The incremental encoder with additional Gray code signals is a relatively cost-effective form of feedback. Other types of encoder that can provide the absolute position of the rotor are described in Chapter 3.4.

The permanent-magnet servodrive is generally used for applications requiring high performance where motor shaft position feedback can be used. Because the rotor is not symmetrical this feedback must give absolute position within each electrical revolution of the motor. Figure 4.9 shows the control system of a permanent magnet servodrive. The inverter control and reference frame transformation is the same as for the open-loop drive. The other blocks within the control system are described below.

Reference-frame Generation
In common with the open-loop drive the reference frame must be aligned with the motor flux so that the current in the x and y axes controls the flux and torque, respectively. In the open-loop drive the stator flux is used because it is easy to calculate its position without rotor position feedback. However, the effects of the current in each axis are not completely decoupled because the x-axis current changes with motor torque even if the stator flux is constant. It is preferable to align the reference frame with the rotor flux as this gives complete decoupling so that the x-axis current controls the rotor flux and the y-axis current controls the torque produced by the motor. In the control system shown the x-axis current reference is set to zero, and so all the rotor flux in the x axis is produced by the magnets. The torque produced by the motor is given by Te = K1 ~rxisy, where K1 is a constant, ~rx is the component of rotor flux in the x axis and isy is the component of current in the y axis. As ~rx is produced by the magnets and is constant,

Current Control
Unlike the open-loop drive the permanent-magnet servodrive uses closed-loop current control giving the drive the ability to change the current and hence the torque produced by the motor very rapidly. A simple PI controller is provided for each axis of the reference frame. The reference for the xaxis (flux-producing current) is zero because the flux is provided by the magnets on the rotor. The y-axis reference defines the torque-producing current and hence the torque produced by the motor. Although the currents and voltages are D.C. quantities in the steady state, high dynamic performance is still required so that the phase delay of the current controllers is as small as possible and the effect on the outer speed controller is minimised.



(3 k/ ~] mo


vector [modulator and inverter

~ frame angle) absolute position feedback

oo ,sv
Figure 4.9 Permanent-magnet servomotor drive control system


A.C. MOTOR DRIVE CONTROL: P e r m a n e n t - M a g n e t Servodrive

rotor flux

Figure 4.10 Rotor flux reference frame for a permanentmagnet motor

The description of the current controllers has been kept simple and does not include anything about compensation for cross coupling or back e.m.f.

The control system described has a zero current reference in the x axis, and so the rotor flux cannot be altered by the drive. This limits the maximum speed of operation as the motor voltage increases with speed. Field weakening is possible but, owing to the large effective air gap of a permanent-magnet motor, a large amount of current is required to reduce the flux, making the drive inefficient in the field-weakening region. It is also necessary to limit the maximum speed with field weakening because, although the motor terminal voltages are reduced by reducing the flux while the drive is enabled, when the drive is disabled the voltages return to the level that would be produced without field weakening and may damage the power electronics in the drive. The drive can be used in simple torque control, without the speed controller shown in Figure 4.9.

The following are some examples of applications where permanent-magnet servodrives are used: * machine tools where precise and dynamic performance is required pick and place applications where the requirements are less precise, but where rapid movements are required more recently, permanent-magnet motors with high numbers of poles (e.g. 32 poles) have been used in low-speed applications such as direct (gearless) drives for lifts.

Speed Control
The outer speed controller is shown as a PI controller using the differential of the position as speed feedback. A more detailed description of this type of controller is given in Chapter 4.3.

Performance and Applications

The performance characteristics of the permanent-magnet servodrive can be summarised as: Good dynamic performance at all speeds. Full torque operation down to standstill. Permanent-magnet servomotors usually have low inertia. Combined with a fast sample rate for the speed controller and fast torque control, this gives a speed controller with a very high bandwidth. Position feedback is required that gives the absolute electrical position of the motor. Permanent-magnet motors exhibit an effect called cogging related to the geometry of the motor, which results in ripple in the motor torque. This effect can be miniraised by good motor design, but can still be a problem.


The closed-loop induction motor drive, often referred to as a closed-loop vector drive, is used in many applications requiring better performance than that of an open-loop drive with an induction motor. To obtain the best performance with this type of drive position feedback is required from the rotor, but unlike the permanent-magnet servodrive only the change of position and not the absolute position is required. The control system is similar to that used with a permanentmagnet servomotor as shown in Figure 4.11. Descriptions are only given for blocks that have not already been covered in previous sections.


flux controller


space vector modulator and inverter Omr(reference frame angle)

incremental 3osition feedback

i~x ,~1 flux


sy ~1

"'1 calculator

Figure 4.11 Closed-loop induction motor drive control system

Chapter 4.2


Flux Calculator and Reference-frame Generation

In common with the permanent-magnet servodrive the reference frame is aligned with the rotor flux. Unlike the rotor flux in a permanent-magnet motor, which remains at a fixed position with respect to the rotor and at a constant level defined by the magnets when operating without field weakening, the rotor flux in an induction motor moves at slip frequency with respect to the rotor and the flux is provided by the x-axis current. Therefore, a flux calculator is required which derives the magnitude and angle of a vector to represent the rotor flux. By convention a current vector is defined called the rotor-magnetising-current space vector, imr, which represents the rotor flux. This vector is aligned with the flux and has a magnitude proportional to the flux, but in units of current. The magnitude and angle of this vector are given by the following equations where Or is the rotor position and Tr is the rotor time constant of the motor:

Wide power range of motors available so that this type of drive can be used for applications requiring less than 1 kW up to more than 1 MW. Suitable for field-weakening applications where motors can be operated up to many times base speed. The drive can be used in torque control, without the speed controller shown in Figure 4.11. The estimate of the flux position to align the reference frame is important as this affects the absolute level of torque produced for a given torque reference. The flux position calculation is dependent on an estimate of the rotor time constant, which varies significantly with rotor temperature. However, it is possible to include a rotor time constant estimator in the drive control system, so that the drive gives consistent torque control. Closed-loop induction motor drives are used in many applications where good dynamic performance is required and especially where an induction motor drive is required to give full torque at standstill. The following are some examples of applications where this type of drive is used: cranes and hoists lifts high-speed spindle applications material winding

]imr[-- i~x/(1 + sty)

Omr -- Or q- f i~y/(Trlimrl)dt


As with the permanent-magnet motor the torque is proportional to the x-axis component of flux and the y-axis component of current, and so Te--Kz[imr[isy , where K2 is a constant. To achieve a linear relationship between the demanded torque at the output of the speed controller, Te, and the torque produced by the motor the torque demand must be modified using ]]mr[ to give the y-axis current demand, isy, as shown in Figure 4.11. The angle Omrgives the angle of the rotor flux and can be used as the reference-frame angle for translation of currents and voltages. As the rotor of an induction motor is simply a symmetrical conductive cage, the absolute position is not important, and so only the incremental position is required to define Or. Therefore, a simple incremental encoder can be used for position feedback.

Operation without Position Feedback

The control system described above requires incremental rotor position feedback, but it is possible to implement the scheme without any physical feedback device. This can be done by estimating the rotor position from information available to the drive through the motor voltages and currents. One class of methods used to determine the rotor position, referred to as model-based methods, uses a model of the motor to calculate the rotor speed and hence the incremental rotor position. When a physical position feedback device is used the drive gives good dynamic performance and operates with full torque even at standstill. When a position estimator is used the dynamic performance is reduced and the minimum speed for full torque operation is similar to that of the open-loop drive. However, a closedloop induction motor drive without position feedback does have the following advantages and disadvantages when compared with an open-loop induction motor drive:

Flux Control
As the flux is produced by the x-axis component of the stator current (equation 4.10) it is easy to control the flux level, and unlike for the permanent-magnet motor the current is reduced when reduced flux is required. Therefore, this type of drive is well suited to high-speed operation with field weakening. The motor is normally operated at rated flux up to the speed at which the terminal voltage reaches the rated level or the maximum voltage that can be produced by the inverter while still maintaining control over the motor currents. Above this speed the motor voltage is limited by reducing the x-axis current reference i~x.

light load instability problems that can occur with an open-loop drive are eliminated torque control operation is improved starting with a spinning motor is faster fast closed-loop current control reduces trips under transient conditions

Performance and Applications

The performance characteristics of the closed-loop induction motor drive can be summarised as: Good dynamic performance at all speeds. Full torque operation down to standstill. A position feedback device that gives the incremental position of the rotor is required.

the motor model is normally dependent on the stator resistance and the rotor time constant; incorrect estimates of these parameters can cause a significant reduction in performance at low speeds; without real position feedback information it is difficult for the drive to compensate for variations in these parameters.


A.C. MOTORDRIVECONTROL:Four-Quadrant Operation

The input inverter can be controlled in a similar way to an inverter supplying a motor, using the reference-frame-based control system described in previous sections. A control system for the input inverter is shown in Figure 4.13.

The drives described in this chapter are normally based on the power circuit shown in Figure 4.12a. This power circuit has the following drawbacks: Although the inverter allows power flow in either direction, the diode rectifier only allows power to flow from the supply into the drive. Therefore the chopper circuit must be used to dissipate unwanted energy in the resistor when braking power is fed from the motor to the drive. The currents taken from the supply generally contain significant harmonics. The voltage at the inverter input is theoretically limited to the peak line voltage of the supply. In practice this is reduced further by voltage drops in the converter. This limits the maximum output voltage without over modulation to less than the supply voltage.

Reference-frame Generation
If the reference frame is aligned so that the vector representing the supply voltage lies along the y axis, the x-axis current controls reactive power flow and the y-axis current controls real power flow. As the input currents should have unity displacement factor the reactive current reference, i~x, is set at zero. The voltage at the D.C. terminals of the inverter is regulated by a PI controller which produces the real current reference, i~y. It is possible to provide voltage feedback from the supply side of the input inductors as the input to a phase-locked loop (PLL) to derive the reference-frame angle. However, this requires voltage isolators connected across the supply which may be at a very high-power level for a large converter. A much more cost-effective alternative is to use the modulation angle, 0m, as the feedback representing the angle of the supply voltage. This is valid provided that the current controllers are operating correctly because this angle is the angle of the voltage vector applied on the inverter side of the input inductors. Normally the phase shift across the input inductors is small even at high loads, and so Omcan be used to represent the angle of the supply voltage vector. If the supply voltage is aligned with the y axis of the reference frame Oref-- Orn -- 90 , therefore Ov= 90 . The reference-frame angle can be defined by a PLL as shown in Figure 4.14. The PLL is required to make the control system stable and to smooth out transients in 0v due to the operation of the current controllers and supply voltage distortion.

If the diode rectifier is replaced with an inverter as shown in Figure 4.12b, the input currents can then be controlled to give almost sinusoidal waveforms with unity displacement factor. It is also possible for power to flow in either direction through the input inverter so that the drive gives full fourquadrant operation.

input inductors inputinverter

* mtt

input rectifier and inverter brakingchopper

+ f o,or

Performance and Applications

The performance characteristics of the four-quadrant drive described here can be summarised as: Full four-quadrant operation. Approximately unity displacement factor and minimal input current harmonics. Fast transient control of the D.C. voltage at the input to the motor inverter even during motor transient operation.

Figure 4.12 Alternative A.C. drive power circuits

a standard A.C. drive power circuit b four-quadrant drive power circuit


PI ~

VD ' ~

~ * ~)_~
PI PI - ~


space vector modulator a ei~ ndl r


ii:.iiiiiiiiiii~i!i]'~ l supply
I I input


I0~ z'-~ ~

Figure 4.13 Input inverter control system

Chapter 4.2



Table 4.2 Effect of different possible voltage vectors on stator flux


Voltage vector
u~ u2 u3

Effect on magnitude of ~s ( ~s ) increase increase decrease decrease decrease increase almost no change

Effect on the angle of ~ (0~) retard advance advance advance retard retard almost no change

Figure 4.14 Reference frame PLL

The D.C. link voltage of a drive with an input diode rectifier is limited to less than the peak line supply voltage. With an input inverter the D.C. link voltage can be boosted to levels above this. No supply voltage feedback required.

u5 u6 u0 or u7

Four-quadrant drives are used in applications where good quality input waveforms are required and/or significant braking energy can be returned to the supply. The following are some examples of applications where this type of drive is used: engine test loading systems cranes and lifts cable laying winders



"""~_~'Z~-~ ~,,

direction of

u4~, N/ ,~u,

~, \ r o t a t i o n





Figure 4.15 Direct torque control The drive control systems described so far all use spacevector modulation to produce the inverter control signals. An alternative method of control that can be used when a motor is connected to the drive is direct torque control. From the earlier description of the space-vector modulator it is clear that the inverter can only produce any one of eight possible voltage vectors. At each sample point direct torque control selects one of the eight voltage vectors to change the motor flux and torque to the reference values as quickly as possible. The principle of direct torque control is demonstrated in Figure 4.15. The stator flux is given by the following equation:

stator flux behind the rotor flux increases decelerating torque. The selection of the voltage vector is further constrained by the change in torque required. If the torque is to be increased at the instance shown in the example then voltage vector u3 or u 4 must be selected or if the torque is to be reduced u5 must be selected. Torque is also controlled by a hysteresis method. A switching table is constructed which contains the voltage vectors to be selected to control the stator flux and the torque to stay within their hysteresis limits. An example of a control system based on direct torque control is shown in Figure 4.16. The torque reference, T~, and the flux magnitude reference, [~Ps[*, are derived in the same way as for a drive using a space-vector modulator. The hysteresis comparators then produce the required change in flux and torque, and the switching table selects the required inverter state to give a voltage vector that will change the flux and torque as required. The angle of the stator flux, a~s, is used to determine which 60 sector the stator flux is in, as different areas of the switching table are used for different sectors. The direct torque control system requires an estimate of the stator flux and the torque for the hysteresis comparators, and an estimate of rotor speed for the speed controller. These are derived from a model-based estimator using the motor currents and an estimate of the motor voltages from the switching table state and the inverter D.C. input voltage, VD.C.. It is important to note that although the principle of direct torque control appears simple and does not depend on estimates of motor parameters, the motor model used to derive the estimates of torque, flux and speed is complex and

~, - / vs - R, isdt


In the example shown the selection of each of the different possible voltage vectors will have a different effect on the stator flux given in Table 4.2. The aim of the control system is to hold the magnitude of the stator flux, CPs, within the hysteresis band indicated by the two dotted lines in Figure 4.15, therefore at the instance shown one of the voltage vectors u3, 114, or us must be selected to reduce the magnitude of ~s. The torque produced by the motor is proportional to the rotor flux multiplied by the stator current at fight angles to the rotor flux. If the angle between the rotor flux and the stator flux is increased in the direction of rotation then the component of stator current leading the rotor flux by 90 must increase because the difference between the rotor and stator flux vectors is Ls'is, where L~' is the so-called transient leakage inductance of the motor. Therefore, advancing the stator flux in front of the rotor flux increases the accelerating torque, and retarding the


A.C. MOTOR DRIVE CONTROL: Direct T o rq u e Control

I VDC Flux controller stator flux ~lq~sl* hysteresis comparator



"~/ torque

switching table


k,I /') "1 "-~


,, ,

h ste ess
comparator switching state

I%1 electromagnetic torque, stator flux and speed estimator


Figure 4.16 Direct torque control system

is heavily dependent on the motor parameters. The following list gives a comparison between direct torque control and a control system based on a space-vector modulator: The calculations for the current controllers, referenceframe translation and space-vector modulator are more complex than for the direct torque control hysteresis comparators and switching table. However, the sample rate required for direct torque control (typically 40 kHz) is much higher than that for a space vector modulator (6-12kHz), because direct torque control uses a hysteresis method. As well as supplying a single induction motor, most induction motor drives can also supply more than one motor in parallel where the motors are different sizes, or the drive can be used as a general-purpose variablefrequency/variable-voltage power supply. Direct torque control cannot be used in these applications, and so a direct torque control drive must also be able to operate with space-vector modulation for these applications.

Because direct torque control is based on hysteresis controllers the inverter has a continuously variable switching frequency.This is considered to be an advantage in spreading the spectrum of the audible noise from the motor, but the range must be controlled so that it does not exceed the maximum allowed by the power electronics of the inverter. Care must also be taken with direct torque control to ensure that changes from one voltage vector to another more than 60 away do not occur repetitively as this can increase the stress on motor insulation. A direct torque control drive inherently delivers a change in torque in the shortest possible time within the limits of the sample rate. Because of the sampling and calculation delays usually associated with space-vector modulatorbased systems a change in torque can take several samples. However, dead-beat type algorithms can be used with a space-vector modulator system giving performance which compares well with that of direct torque control.


Most medium and large-size industrial D.C. motor drives are based on a separately-excited motor. The flux is generated by a field winding and the torque by a higher current armature winding fed via a commutator. These two windings are completely independent, and so the flux and torque can be controlled independently as shown in Figure 4.17a. Although the field winding usually has a long time constant, the armature winding time constant is normally very short allowing fast changes of armature current and hence torque.

The field converter is either a half or fully-controlled thyristor converter. The half-controlled converter can only apply positive voltage to the field winding, and so the current in the winding can be increased quickly, but decays relatively slowly. The fully-controlled converter can apply positive or negative voltages, and so the performance is the same whether the current is increasing or decreasing. When the motor is rotating at a speed below base speed the field

Chapter 4.3


flux controller

i~: v"-~iF
' ~ , ~


v,~ V~ ~

I controller Phase continuous I phase

~ controller

field converter armature converter

J D.(
m c or

cbntinuous supply voltage


speed feedback


firing ~1 angle ~_1 prediction "-I


armature---, supply

=2 ~ "-

positive t ~ ~ ~bridge





converter (fullycontrolled)

~_IL~ + ~L_~egativ e 'i 7- ~ ];ridge armatureconverter

Figure 4.17 Separately excited D.C. motor drive
w w

a control system b power circuit

current reference is constant at the rated level for the motor, and so the motor armature voltage increases with speed. The armature voltage reaches its rated level at base speed, and above this speed the flux controller reduces the field current reference, i~, to keep the voltage at the rated level as the speed increases further. The armature voltage feedback can include armature resistance compensation to avoid the effects of the armature resistance drop on the voltage control loop.

control the firing angle of the converter. When the current is discontinuous the relationship is highly nonlinear and varies with the voltage level applied to the motor. The drive stores the relationship between the firing angle and motor current for different output voltage levels, and during discontinuous current operation the correct firing angle is selected by the firing-angle prediction block for a given current reference. Any errors are trimmed out by an integrator operating on the current error. When a change in direction of torque is required one bridge must stop conducting and the other bridge must become active. It is clear from the power circuit diagram that only one bridge must conduct at a time during this changeover to avoid a short circuit across the armature supply. It is important that this changeover occurs as quickly as possible to give good dynamic torque control. Modem microprocessor-controlled drives enable intelligent methods to be used to keep the bridge changeover delay as short as possible. Commutation from one thyristor to another during continuous current operation is acheived by applying firing pulses to a nonconducting thyristor that has a more positive phase supply voltage than the thyristor presently conducting in the top row (or more negative in the bottom row). This is

For four-quadrant operation two thyristor bridges are used as shown in Figure 4.17b. Both bridges can apply positive or negative voltage to the motor, but the positive bridge can only supply positive current and the negative bridge negative current. Therefore, the positive bridge conducts when positive torque is required and the negative bridge when negative torque is required. The bridges are phase controlled to apply the voltages required by the reference, v], to the motor. Owing to the high voltage ripple in the converter output and the unidirectional nature of thyristors, the current in the armature can be continuous or discontinuous. While the current is continuous the relationship between the voltage reference and the actual applied voltage is a cosine function and the voltage reference can be used directly to


D.C. MOTOR DRIVE CONTROL: Torque Controller

earliest firing point positive voltage


The performance characteristics of a D.C. motor drive can be summarised as: The current controller sample rate is limited by the possible commutation rate of the thyristors in the bridges. In general, the sample rate and hence the bandwidth of a D.C. drive is ten times lower than that of an A.C. drive. The D.C. drive with separately-excited D.C. motor is used in similar applications to closed-loop induction motor drives. Because the flux and torque are controlled by separate winding, the decoupling of flux and torque control is not dependent on knowledge of the motor parameters. Therefore accurate control of torque is more easily achieved. The thyristor power circuits used in a D.C. drive cost less than an IGBT inverter for an A.C. drive of equivalent power rating. However, a D.C. motor is generally more expensive than an A.C. motor of equivalent power rating, at power ratings below approximately 200 kW. In general, D.C. motors require more maintenance than do A.C. motors. A full four-quadrant D.C. drive can be constructed with two thyristor bridges as shown in Figure 4.17b; however, the input power factor is poor. A full four-quadrant A.C. drive using an input converter as described in Chapter 4.2 takes currents with significantly less harmonics and with almost unity displacement factor. The drive can be used in torque control, without the speed controller shown in Figure 4.17.

negative voltage

inversion limit

Figure 4.18 Voltage produced by top row of thyristors

demonstrated by the example shown in Figure 4.18. The supply has some inductance, and so the current does not transfer instantaneously from one thyristor to the other, but takes a finite time known as the overlap time. During this period the voltage applied to the motor is derived from the average of the two conducting phases. This presents no problem during motoring, when the bridge operates as a rectifier, except for distortion of the supply known as notching. However, during braking the bridge operates as an inverter and care must be taken to ensure that commutation is complete before the inversion limit shown in Figure 4.18. If it is not complete by this point the current builds up again in the thyristor which should turn off and a large pulse of current flows in the motor. This is known as inversion failure. The traditional way of preventing inversion failure is to limit the firing angle to allow commutation to be completed before the inversion limit; however, this limits the motor voltage during braking so that it is less than the limit when motoring. Modern D.C. drives include algorithms which monitor the situations that lead to inversion failure, allowing the firing angle to move right back to the inversion limit under some conditions.

Although it has been predicted for many years that A.C. drives will replace D.C. drives, this has only happened slowly and many D.C. drives are still manufactured. These are used in many industrial applications especially in larger sizes. The following are some examples of where D.C. drives are used: cranes and hoists lifts material winding


The speed controller performance is important in many drive applications. In this section a speed controller is used first as an example to demonstrate the theoretical analysis that can be used to predict the performance of a closed-loop controller and then to show the unwanted effects produced by delays often found in digital control systems.


For the purposes of this analysis it is assumed that an ideal speed controller is used with a load that is a pure inertia. This can be used for most simple systems, especially in servo applications where the load usually is an inertia with very

C h a p t e r 4.4


speed controller

current controllers


inertia load


Kv+ KJs







Figure 4.19 Ideal speed controller with inertia load

little friction. An s-domain diagram of such a speed controller is given in Figure 4.19 where:

w*(s) = speed reference (rad s -1) w(s) = speed (rad s -1) isy = torque-producing current (A) Te = electrical torque from the motor Ta = torque disturbance (N m) Tm = mechanical torque applied to the load (N m)
J = load inertia (kg m 2) Kp = proportional gain of the speed controller (1/(rad s-l)) K; = integral gain of the speed controller (1/(rad s-l)) Kc = closed-loop gain of the current controllers (A) Kt = motor torque constant (N m A - 1) The closed-loop gain of the current controllers depends on the current rating of the drive. The gain is defined so that a current reference from the speed controller of unity gives a torque-producing current of K~ amps. If an output of unity from the speed controller results in a torque-producing current that is equal to the rated current of the drive, then Kc can be defined as the r.m.s, rated current of the drive. The motor torque constant defines the torque produced per unit current. Normally, Kt is defined in N m per amp of r.m.s. phase current for a servomotor. Therefore, an output of unity from the speed controller produces K~ x Kt N m. The closedloop response of the speed controller is given by:

In addition to the closed-loop step response the stiffness of the system can be assessed by its response to a torque transient. From the s-domain block diagram w(s)= (Te + Ta)/sJ and Te = - Ko(Kp + KJs)w(s). Combining these equations gives:

w(s)lw*(s) = [llXoKi][s/(sZJ/KoKi + sXplXi + 1]

(4.17) Dividing both sides by s is equivalent to integrating in the time domain. If Or is the rotor position:

Or(s)/Ta = [1/XoXi][1/(sZJ/XoKi + sXp/Xi + 1] (4.18) Or(s)/Ta- [1/KoKi][1/(sZ/w 2 + s2~/w, + 1]


This equation defines the response of the speed controller to a torque transient. The compliance angle is defined as the steady-state change of angle to a steady-state torque disturbance of rated torque which is the torque produced when the drive is delivering rated current. The responses shown in Figure 4.20b are the position changes as a result of a step torque disturbance of rated torque for a typical servodrive with a compliance angle of 0.07 rad (4). The response to a torque transient is similar to the secondorder response shown in Chapter 4.1. As the damping factor is increased the overshoot is reduced and the response becomes slower. Once the steady state is reached the position has changed by the compliance angle of 0.07 rad. If a steady-state change of position after the application of a load is not acceptable an additional outer position controller must be used. The closed-loop step response is different from the response to a torque transient because of the s term in the numerator. As the damping factor is increased the overshoot is reduced, but with unity damping factor there is still 10 per cent overshoot. Also, the rate of response improves with increased damping factor. It would appear from these results that the higher the damping factor the better the response, i.e. slower response to a torque transient and faster response to a step change of reference. Although this is the case with an ideal speed controller some of the nonideal characteristics of a real system limit the value of the gains and hence the best possible quality of the responses (see Chapter 4.4).

w(s)/w*(s) = G(s)/[1 + G(s)]


where G(s) is the gain in the forward part of the loop. If Ko -- KcKt, then:

+ (X,, + Xi/ )Xo/ J]

Rearranging gives:


w(s)/w*(s) - [sKp/Ki + 1]/[s2j/(KoKi) + sKp/Ki + 1]

(4.15) The natural frequency of the system is defined as w,,= x/(KoKi/J) and the damping factor ~=w,,Kp/(2Ki), then:

w(s)/w*(s) = [s2~/w,, + 1]/[sZ/w 2 + s2~/w,, + 1] (4.16)

This equation gives the closed-loop response of the speed controller in terms of its natural frequency and damping factor. Step responses are shown in Figure 4.20a for a typical servodrive speed controller to a 1.0 rad s -1 reference step for different values of damping factor.


The gains for an ideal speed controller should be set up in two stages. First, the integral gain is set up to select


ANALYSIS OF AND SET UP OF A SPEED CONTROLLER: Calculating t h e Required G a i n s


rad s-1 1.0

/ ~= 2/3


= 3/2






time, ms


4= 1/3





Figure 4.20 Step response of an ideal speed controller
a closed-loop speed response b response to a torque transient

time, ms



the compliance angle or the bandwidth of the system and, second, the proportional gain is selected to give the required amount of damping. The gains are calculated in the units used in the previous section. It is unlikely that these could be entered directly into the gain parameters of an actual drive as the units used vary from one drive manufacturer to another, and so some scaling is normally required to convert these values. The values calculated for an ideal speed controller can be used in many applications, but care must be taken not to exceed the bandwidth limit of a particular drive. Equation 4.18 represents the response to a torque transient. After a transient has been applied the resultant steady-state change of position can be derived from this equation as Or-Ta/KoKi. If Tdrated is the rated torque and Or complianc e is the compliance angle, the equation can be rearranged to give the required integral gain for a given compliance angle:

If rated torque is defined as the torque from a torque-producing current that is equal to the rated current of the drive, then Tdrated-- K c K t - - K o. Therefore, K i - l/Orcomplianc e. The compliance angle may be defined for a different current level, such as drive full-scale current, which may be two or three times the rated current. Therefore equation 4.20 may be more appropriate. Alternatively, the integral gain can be selected to give the required bandwidth of the speed controller. The integral gain defines the natural frequency of the controller, wn. The relationship between the natural frequency, wn, and the bandwidth is defined by the proportional gain. The natural frequency has already been defined as W n - v/(Kogi/J). This can be rearranged to give the required integral gain:
Ki - wZnJ/Ko


g i = Tdrated / (KoOrcompliance)


The damping factor has already been defined as ~ - w,,Kp/Kil 2. Although the user must select the required damping factor for a particular application, a damping factor of unity is often

C h a p t e r 4.4


used for servo applications. By rearranging this equation and substituting for w, from equation 4.21, the proportional gain can be calculated as:

sample period for speed measurement

speed controller delay

~ v

torque controller delay

Kp = 2~v/((KiJ)/Ko)
and with unity damping factor:


Kp -- 2v/((KiJ)/Ko)


difference in position feedback over this p e r i o d gives speed feedback

torque reference calculation

time lag due to the torque controller

The relationship between the natural frequency and bandwidth can be determined from equation 4.16. If the bandwidth, Wbw,is arbitrarily defined as the point where the gain is - 3 dB, i.e. w(s)/w(s)*= 1/x/~, then:

Figure 4.22 Unwanted speed controller delays

characteristics of the speed controller response. The most significant delays are shown in Figure 4.22. Figure 4.23 shows how these delays cause additional overshoot in the speed controller step response. In both cases the calculations for an ideal speed controller have been used to derive the proportional and integral gains to give unity damping factor. In Figure 4.23a the required bandwidth is 672 rad s -1 (107 Hz), and is probably at or beyond the limit of acceptable performance for this particular drive. The ideal speed controller overshoot of 10 per cent has been increased to 25 per cent by the additional delays. In Figure 4.23b the required bandwidth has been increased further to 1000 rad s- ] (159 Hz), and the performance has deteriorated further. The effect of these additional delays can also be seen in the bode plots of the closed-loop system response. Figure 4.24 shows the bode plot for the speed controller set up to give the results shown in Figure 4.23a. The frequency at the - 3 dB point of the gain plot has increased significantly from the characteristic of an ideal speed controller, whereas the frequency at the 60 point of the phase plot is similar to that of an ideal speed controller. This reinforces the point made in Chapter 4.1 that the phase delay of a controller should be used as a measure of the response time, and not the frequency at the - 3 dB point of the gain characteristic. If the frequency at the - 3 dB point had been used in the example given, the bandwidth would have appeared to be 2000 rad s -1 (318 Hz) and yet the phase delay limits the real performance to that of a controller with bandwidth close to that of the ideal system, i.e. 672 rads -1 (107 Hz).

Wbw/Wn- - x/'[(2( 2 + 1 ) + v/((2( 2 + 1) 2 + 1)]


The value of W b w / W n for different values of damping factor is shown in Figure 4.21. It can be seen that with a damping factor of unity Wbw/W.= 2.5.


Most modem drives have digital current and speed controllers which include various delays that affect the







damping factor, ~
Figure 4.21 Effect of damping factor on bandwidth


real speed controller

4e~-ideal speed controller


a limit of drive bandwidth b above limit of drive bandwidth

I =
i i i i i i

Figure 4.23 Effect of system delays on speed controller step response


ANALYSISOF AND SET UP OF A SPEEDCONTROLLER:Nonideal Effects in a Real Speed Controller

1~~ _ i,,]-"~"-[-~."


real speed controller

dB 101 -15 -20 -25 -30

-10 -20 -30

Other effects can limit the maximum possible bandwidth of a speed controller. For example, the quantised nature of the encoder feedback can cause high-frequency torque ripple which manifests itself as acoustic noise from the motor. A high proportional gain in the speed controller increases this effect, and so the proportional gain and hence the integral gain must be limited to limit the motor noise.






de,.,-40 u_50. -60. -70. -80 -90 rads-1

Figure 4.24 Bode plot of speed controller closed-loop response

Switched-Reluctance and S t e p p e r - M o t o r Drives


121 122





All the motors described earlier have the characteristic that they can be operated, albeit at essentially fixed speed, when fed by an appropriate D.C. or A.C. supply. In this section we consider a group of drive systems where the motor can only be used in conjunction with its specific power converter and control, and consequently only overall characteristics are relevant. Switched-reluctance drive systems are of growing importance notably in applications where high low-speed torque is required, and less importance is placed on smoothness of rotation. Although considerable advances have been made in improving the noise characteristic of this drive, it can still

be a limiting factor where a broad operating speed range is required. Stepper-motor systems are somewhat in decline. Their operating characteristic of being controlled by a computer pulse train is now a common feature of many modem servodrives. Also, where rapid settling times are required, stepper drives, which are inherently open loop, are not ideal in both fundamental performance and the fact that varying mechanical friction has a significant impact. This variability can make stepper drives unacceptable in applications where the transient performance is important.


SWITCHED-RELUCTANCEMOTORSAND CONTROLLERS:Basic Principle of the Switched-Reluctance Motor



Like the stepper motor, the switched-reluctance motor (SRM) produces torque through the magnetic attraction which occurs between stator electromagnets (formed by winding coils on salient poles) and a corresponding set of salient poles formed on a simple rotor made only of electrical steel (or other ferromagnetic material). The stepper and switched-reluctance motor share the same basic principle of energy conversion, and both are members of the family of variable-reluctance motors. The intuitively straightforward principle of torque production is easily visualised in the very simple reluctance motor illustrated, in cross section, in Figure 5.1. The motor illustrated is commonly referred to as having a 'two-two' pole structure, the two numbers referring to the number of stator and rotor poles, respectively. Intuitively, it can be seen that if current is passed through the stator windings, with the rotor position as shown above, then the rotor will experience a clockwise torque - as indicated by the arrows. Note that, since no permanent magnets are involved, the polarity of the phase current is immaterial. This permits the use of a wide range of power electronic drive circuits, as will be shown later. If the rotor (and any associated mechanical load) is free to move, this torque will cause the rotor to accelerate clockwise. Torque will continue to be produced in a clockwise sense until the rotor reaches the fully aligned position shown in Figure 5.2. The fully aligned rotor position is sometimes referred to as 'top dead centre' (TDC) by analogy with the internal combustion engine. At top dead centre (TDC), the magnetic circuit - completed by the rotor - offers minimum opposition to magnetic flux. This opposition is known as the magnetic circuit's reluctance, and is analogous to resistance in an electrical circuit. Hence, at TDC, the phase reluctance is at a minimum. This

means that, for a given value of phase current, the magnetic flux linked by the windings is maximised, and therefore the phase's electrical inductance - defined as flux linkage per unit current - is at its maximum value. Assuming that mechanical self inertia carries the rotor past TDC, it can be seen that the polarity of torque produced by the motor will reverse if we continue to energise the phase windings beyond the fully aligned position; see Figure 5.3. Although the rotor is still turning clockwise under its angular momentum, the torque is now applied in an anticlockwise sense, and will first reduce the rotor's clockwise angular velocity, and eventually may - depending on the initial rotor speed and the total moment of inertia - cause it to reverse. Thus, the polarity of torque can be reversed, and braking (i.e. generating) can be accomplished without reversing the phase

i current phase

Figure 5.2 Rotor at top dead centre (TDC); zero torque produced, magnetic circuit reluctance at a minimum and electrical inductance at a maximum

phase current laminated steel stator

i current phase


laminated steel rotor

Figure 5.1 "Two-two" reluctance motor Figure 5.3 Torque reversal beyond TDC

Chapter 5.2 phase


TDC (Figure 5.2) when we must switch the phase current off. The rotor will coast beyond TDC under its own momentum (Figure 5.3), until it reaches BDC (Figure 5.4). The phase can then be switched on again and the whole cycle repeated. If we fail to switch the phase current off at TDC then, while the current, or to be strictly correct the magnetic flux associated with it, persists, anticlockwise torque will be produced as the rotor turns further clockwise. This anticlockwise torque will reduce the average motoring torque produced by the machine over a cycle of operation. Note that the machine will operate equally well as a motor in either direction of rotation, the stator coils just need to be energised over the appropriate range of angular position.

Figure 5.4 Rotor at "bottom dead centre" (BDC); zero torque produced, magnetic circuit reluctance at a maximum and electrical inductance at a minimum

Operation as a Brake or Generator

So far as basic control of the machine is concerned, there is no fundamental distinction between the terms generating and braking. As an aside, it could be argued that generating is concerned with the efficient conversion of mechanical work to electrical energy, whereas braking is simply the removal of mechanical energy from the load regardless of where that energy ends up. For the purposes of our present simple explanation, the two terms will be regarded as being equivalent. It is fairly easy to see that generating with the SRM is a mirror image of the motoring operation. Generating requires that the torque be of a polarity such that it opposes the present direction of rotation. Following on from the previous example, again assume that the present direction of rotation is clockwise. We therefore require anticlockwise torque to brake or generate. This time, considering Figures 5.1 to 5.4 reveals that we should energise the phase winding when the rotor poles are moving away from the stator poles, as shown in Figure 5.3. In other words, current should be applied when the phase inductance is decreasing (or reluctance increasing). The current should ideally be switched on at TDC, and off again at BDC. Again, if magnetic flux, and current, persist when the rotor has turned beyond BDC, then motoring torque will be produced over part of the machine's electrical cycle, and the average braking (or generating) torque will be reduced.

current, in spite of the fact that the machine has no magnets or windings on its rotor. When braking or generating, mechanical work performed on the rotor is converted into energy in the magnetic circuit, which can then be recovered as electrical energy to the power supply by means of the phase winding. If the rotor turns still further clockwise, it will eventually reach a second position of zero torque, this time when its poles are fully unaligned with respect to the stator poles. Qualitatively, we may say that the clockwise and counterclockwise forces now balance each other out, and the net torque is zero. The fully unaligned position is commonly referred to as 'bottom dead centre' (or BDC), as illustrated in Figure 5.4. In contrast with Figure 5.2, the magnetic circuit's reluctance is now clearly at its maximum possible value, and the electrical inductance of the phase is correspondingly at a minimum. Now consider how to use the torque-productive intervals to operate the switched-reluctance machine as a motor or generator.

To Summarise so Far
Motoring torque is produced when a phase is energised while its inductance is increasing with respect to time. Conversely, braking (or generating) torque is produced if the phase is energised while its inductance is falling with respect to time. We can therefore choose motoring or generating operation of the SRM simply by timing the excitation of the phases with respect to the rotor position. Hence it is a reluctance machine in which the phases are controlled, or switched, in accordance with rotor position, giving rise to the name switched-reluctance motor. Note that it is a selfsynchronous machine, i.e. the stator excitation frequency is locked to the rotor speed. This is in contrast to, say, the induction motor, where rotor speed is a result of the applied frequency. Figure 5.5 shows the variation of phase inductance as a function of angle for a typical SR machine. If we apply current over the intervals discussed above, then the polarity

Operation as a M o t o r
Motoring operation of the SRM requires that the torque generated by the machine acts in the same sense as the actual direction of rotation. In other words, the torque should be of a polarity- i.e. direction - such that it reinforces the present direction of rotation. Through study of Figures 5.1 to 5.4 we can see that, to operate the machine as a motor, we should energise the winding of the machine only while the rotor and stator poles are approaching each other. In other words, ideally, apply phase current only while the magnetic circuit reluctance is decreasing - or, equivalently, while the phase's electrical inductance is increasing - with respect to time. Assume that we want rotation in a clockwise direction, starting with the rotor positioned as shown in Figure 5.1. We should energise the phase winding until the rotor reaches


SWITCHED-RELUCTANCE MOTORSAND CONTROLLERS: Basic Principle of t h e Switched-Reluctance M o t o r

/ X .... Lmax


The simple control methodology discussed so far assumes that (for motoring operation) a controlled phase current is switched on when the rotor reaches BDC, and off again at TDC. By varying the magnitude of the phase current, the average torque produced by the motor can be regulated. This approach is indeed used in practice to control the motor at relatively low rotational speeds. The power electronics applies the full D.C. supply voltage to the phase winding(s), thereby causing the magnetic flux - and hence phase current - to rise at the maximum possible rate. When the phase current has reached its working value, which at low speeds will occur within a relatively small rotor angle, the converter must limit the current, which it does by reducing and controlling the average voltage at the winding. This is usually accomplished by switched-mode action, or so-called chopping of the phase current. When the rotor reaches TDC, the power converter applies the full D.C. bus voltage in reverse across the winding. This forces the flux linkage, and hence the phase current, to fall at the maximum possible rate, until both the flux and current are zero. Again, at low rotor speeds, this happens in a small mechanical angle.








mechanical angle (degrees clockwise)

I I o, I


I I o,



I I off

Figure 5.5 a typical inductance with angle b phase current for clockwise torque c phase current for anticlockwise torque

of torque can be seen to depend on whether the inductance is rising or falling with angle. If the machine is only lightly magnetically loaded (when it will be producing only a modest amount of torque for its size), then the steel from which the rotor and stator are made will behave magnetically in an approximately linear fashion. That is, for a given number of turns on the windings, the phase's magnetic flux will vary approximately in proportion to the phase current. If linearity is assumed, then it can be shown that the torque produced as a function of angle is:

What Happens as Speed is Increased?

An important point, not considered earlier, is the recognition that magnetic flux and phase current cannot, in practice, rise or fall instantaneously. Faraday's Law states that the rate of change of flux linkage (equal to magnetic flux, qS, multiplied by the number of winding turns Nph) is proportional to the applied voltage, Vph, hence:
dch/dt = (Vph/Nph)


This relationship shows independence of polarity of phase current (due to the i2 term) and that the torque polarity depends on the slope of the inductance curve with angle. Given that the angular spacing of Lma x and Lmi n a r e fixed by the machine's rotor pole pitch, the magnitude of the inductance's gradient with angle will depend on the difference between Lma x and Lmi n. Thus the output of the machine, for a given current depends on this difference, which should ideally be as large as possible. In a practical machine, the ratio of Lma x t o Lmi n will typically lie in the range 4:1 to 10:1.


Inductance is defined as flux linkage per unit current, i.e.:

L = (Nph" ~ ) / i


We can rewrite equation 5.3 to yield an expression for current:

i = (Nph . dp)/L


Relationship between Torque Polarity and Motoring/Generating

Reversing the polarity of torque, for a given direction of rotation, switches the machine between motoring and braking operation. To do this, it is necessary to simply alter the excitation pattem so that each phase is energised either when the inductance is rising or when it is falling. Note also that braking torque becomes motoring torque - and vice versa if the direction of rotation is reversed, provided that the excitation pattem with respect to rotor angle remains unchanged. The choice of rotational direction defined as positive (increasing) angle, as shown in Figure 5.5, is an arbitrary one. Once this is decided, we can define positive or forward torque and speed in the same sense. Motoring operation then occurs when the torque sign and speed sign are the same; braking occurs when they are dissimilar.

The relationship of equation 5.4 is key to understanding the current waveforms in the switched-reluctance motor phases. The current at any instant is determined by the ratio of magnetic flux linkage to inductance. Remember that magnetic flux is a function of the applied winding voltage and time (equation 5.2), and the phase inductance is a function of rotor angle and hence has both speed and time dependency.

Medium-Speed Operation
Equation 5.2 shows that the rate of change of magnetic flux is constrained by the available D.C. supply voltage. Consequently, from equation 5.4, the current will take a finite time to reach its working value when switched on at Lmi n and a second (longer) time to fall back to zero when we switch off at Lma x. At low rotational speeds, these time intervals will occupy negligible rotor angle, and it is possible for the phase current waveforms to closely approximate the ideal squarewave functions shown in Figures 5.5b and 5.5c.

Chapter 5.2


control the motor, and this approach is not used in practical drive systems. A preferred approach is normally adopted to mitigate this effect - angle advance. The switch-on point of the phase current is advanced with respect to the rotor angle, so that the magnetic flux and current have already reached their working values by the time the phase inductance starts to rise. Similarly, the turn-off angle is advanced, so that by the time the inductance starts to fall, the magnetic flux and current have already been forced to relatively low values, and little braking torque results. If the angle advance is too great, however, torque will again be lost. Too early a switch-on angle will place significant flux in the machine when the inductance is still falling from the previous cycle. The resultant braking torque will reduce the average output of the machine while adding to the phase current, hence also worsening efficiency. Switching off too early will avoid the braking torque otherwise incurred as the inductance falls beyond top dead centre, but it will unnecessarily reduce the torque. Increasing the peak phase current can restore the output, but this may worsen the rootmean-square phase current, hence increasing the copper losses in the winding and again worsening efficiency. Therefore, an optimum degree of angle advance must be chosen for both switch on and switch off, so as to maximise efficiency and output for the torque and speed level in question.


60 120 180 240 300 mechanical angle (degrees clockwise)


Figure 5.6 Phase inductance and realistic current waveform at medium speeds

However, as the rotational speed increases, the time occupied by an electrical cycle of the machine falls with the reciprocal of speed, and the rise and fall times of the current become significant. This yields the sort of phase current waveform shown in Figure 5.6. The finite rise time of the phase current means that a little torque will be lost because the current has not reached its working value when the inductance starts to rise. However, the persistence of current beyond Lma x (TDC) is more significant, because it results in a short period of braking torque. Not only does this reduce the average motoring torque, but the extended phase current also brings a small increase in winding losses. This means that the output has fallen and the losses have increased, and therefore this tail current can noticeably reduce the efficiency of the motor. The nonlinear shape of the current at turn off is due to the fact that not only is the magnetic flux falling linearly with time, but the inductance also falls. This means that, by equation 5.4, the current decreases less rapidly than might otherwise be expected. The effect is less marked at Lmin because here the inductance changes less rapidly with angle (hence, for a given speed, with time). It is the rate of change of magnetic flux relative to the rate of change of inductance with time that determines whether the rise and fall times of the current are significant or not. Note that the rate of change of magnetic flux is determined by the number of turns on the winding as well as the supply voltage (equation 5.2). This means that the same effect will be noticed at a lower rotor speed if the number of turns on the phase winding is increased or if the D.C. supply voltage is reduced.

High-Speed Operation
As the rotor speed rises still further, the rate of change of inductance with respect to time increases along with it and the effects illustrated in Figure 5.6 are exacerbated. Eventually, a speed is reached where the phase current can be naturally limited by the phase inductance, while still maintaining a sensibly broad pulse of current and flux. Under these conditions, there is no need to limit the phase current by reducing the phase voltage, and the torque can be efficiently controlled simply by adjusting the switching angles with respect to rotor angle. This is sometimes referred to as the single-pulse control mode, so called because the phase voltage is applied as one continuous pulse, rather than being chopped by repeated switching of the power electronic control circuit. If the speed is high enough, it is possible for the current to inflect and actually decrease, despite the full positive supply voltage still being applied to the winding. This is shown in Figure 5.7. Note the inflection of the waveform at peak current. This occurs in spite of the fact that the full supply voltage is applied to the winding, and that the phase magnetic flux is therefore increasing. The current rolls over
full D.C. supply voltage applied to winding; magnetic flux increasing ~ / full D.C. supply applied in reverse; fl x.~decreasing

How is Performance Maintained as Speed Increases?

The effect of tail current could be minimised by using relatively few turns on the motor windings, thereby ensuring that the rate of change of flux was always large enough to force the flux and current up and down within a negligibly small rotor angle, regardless of speed. However, this would mean that, for a given working value of magnetic flux in the motor, very large phase currents would be needed. This would increase the cost of the power electronics required to

Figure 5.7 Effect of large rate of change of inductance on phase current at high speed



The use of angle advance, with the switch-on and switch-off angle both independently adjustable, means that it is possible to maintain a high level of energy conversion efficiency as operating conditions are varied. The SR machine, when controlled in this way, is capable of producing high efficiency over a very wide range of torque and speed. By choosing appropriate switching angles and current levels, together with an appropriate electromagnetic design, the torque-speed characteristic of the switched-reluctance drive can be tailored to suit the end application. Furthermore, simply by changing the control parameter selection with torque and speed, a given machine design can be made to offer a choice of different characteristics. It is usual to store the control parameter variations within the motor control system, e.g. as mathematical functions (of torque demand and motor speed), or as look-up tables, either of which can readily be embedded into the controlling microprocessor's software code.

because time rate of rise of inductance exceeds the time rate of rise of magnetic flux, and therefore, by equation 5.4, the current must fall. Eventually, if the speed increases still further, a point is reached where it is only just possible to reach the chosen working flux level, and then drive the flux back down to zero, within one electrical cycle. The peak flux, and hence current, is then limited by the available D.C. bus voltage, and any further increase in speed will necessitate a reduction in peak flux and hence in output torque. This defines the base speed of the switched-reluctance motor. It is possible to operate the machine with a component of continuous (standing) current and flux. That is, the current and magnetic flux associated with a given phase do not fall to zero before the next cycle of operation begins. This socalled continuous-current mode has some further implications for the control of the machine, but is useful in increasing the available power output above base speed. It is especially useful in systems where a wide constant power range is needed or where a large transient overload capability is needed at high speeds. This technique has been patented by Switched Reluctance Drives Ltd and Emerson.

Control of Speed and Position

When controlled as described, the switched-reluctance machine and its basic control system form essentially a torque-controlled drive. This can be compared in performance to a separately-excited D.C. machine with controlled armature current. This is sometimes what we want, e.g. for tension control in winding machinery or for web handling. However, more commonly, controlled speed (or even position) is what the end user requires. If torque is still produced as the motor speed increases, the SRM's speed will increase without limit if additional measures are not taken to control it. To do this, we simply include the basic torque-controlled SR machine within an outer speed regulating loop, as is done with the traditional D.C. drive. Feedback of speed can be readily derived from the rotor position sensor, if one is used, otherwise from sensorless control position data. Figure 5.8 shows a typical control structure for a basic speed-controlled switched-reluctance system. Position control of the SRM can be achieved (in the conventional manner) by adding a further position control loop around that shown in Figure 5.8. Note that, under closedloop position control, the SR machine is not limited to shaft positions corresponding to its natural detents. The shaft position can be controlled to any desired resolution using the

Summary of Typical/Practical Control

The phase currents are always switched synchronously with the rotor's mechanical position. At low speeds, the phases are energised over the entire region of rising inductance, and active current limiting is required from the controller. Torque is controlled by adjusting the magnitude of the phase current. As speed increases, the rise and especially fall times of the phase current occupy significant rotor angle, and it is usual to advance the turn-on and turn-off angles with respect to rotor position. The torque is now controlled by both the current limit level and by the switching angles, although current is usually used as the primary control variable. At high speeds, the rise and fall times occupy still greater rotor angles. The current naturally self limits and it is possible and indeed usual to control the torque using only the switching angles. The shape of the current waveform is greatly influenced by the high rate of change of inductance with respect to time.


torque demand

switching angles and current control level

~ position f e e d b a c k for c o m m u t a t i o n

error amplifier




H parameter.~ "control cntr'

Ikup I I core I l

.~ .J

,, ,motor

power switching

motor speed for parameter determining and for speed control Figure 5.8 Basic speed-controlled switched-reluctance system

Chapter 5.2
position control loop. This will in turn demand speed, and thence torque, which will ultimately be translated into the appropriate currents in each phase required to maintain the set position.


Each phase circuit then comprises four stator coils connected and energised together. Increasing the phase number brings the advantages of smoother torque and self starting in either direction. (Self starting is possible with two phases, but only in a predetermined direction, generally using a profiled or stepped rotor design. Self starting in either direction requires at least three phases.) A higher phase count does increase the complexity of the associated power electronics and, to some extent, the signal level controls. However, increased power electronics costs are mitigated by the power throughput being equally distributed between the phases of the controller, and its total volt-ampere rating may therefore be no g r e a t e r - and in some cases may actually be less - than that required for a machine of low phase count.


The simple single-phase machine we have discussed so far is capable of producing torque over only half of its electrical cycle (which, for the two-two pole structure, repeats twice per revolution). Motoring or braking torque from such a machine will necessarily be discontinuous, and hence starting in the desired direction is not possible from all rotor positions. Single-phase motor starting can be ensured by including a small parking magnet within the stator, positioned such that the rotor always comes to rest in a torqueproductive position. These limitations are acceptable for some applications, and the single-phase SR motor is especially useful for low-cost high-speed applications such as vacuum cleaner fans. More demanding applications use higher pole numbers on the machine's rotor and stator, with the stator poles wound and connected into multiple identical phases. Figure 5.9 illustrates the cross section of a three-phase six-four machine. Here, diametrically opposed coils are connected together to form three-phase circuits, denoted as phases A, B and C. The excitation of the phases (in this case, three in number) is interleaved equally throughout the electrical period of the machine. This means that torque of the desired polarity can be produced continuously, thus greatly reducing the variation in output torque with respect to angle, the 'torque ripple'. Furthermore, machines with more than two phases are able to start in either direction without requiring special measures. The number of phases can in theory be increased without limit, but phase counts from one to four inclusive are the most common for commercial and industrial applications. Many different combinations of pole count are possible. It is sometimes beneficial to use more than one stator pole pair per phase, so that, for example, the twelve-eight pole structure is commonly used for three-phase applications.


Like any other electric motor, the losses in an electric machine may be categorised as follows:
a Copper losses (in the p h a s e winding) - resistive heating from the main phase current, plus any losses due to circulating or eddy-current heating in the copper (sometimes called proximity effect and an important consideration at high frequencies). Skin effect, which effectively raises the winding resistance by concentrating the current around the outside of the conductor, must be taken into account when calculating the effective phase resistance and the copper losses in a highfrequency machine. Iron losses (in the rotor and stator steel) - caused by the changing magnetic flux within the steel. Eddy-current losses, caused by induced circulating currents in the steel, generally increase with the square of frequency, and hysteresis losses follow the frequency of excitation directly. The overall rate of increase of iron loss with frequency therefore lies somewhere between the two cases. The iron losses are heavily dependent on the electromagnetic properties of the steel used. Friction and windage losses - mechanical losses resulting from the rotational speed, such as bearing losses, air friction, turbulence and shearing of air layers at rotor edges.

At relatively low rotational speeds - and especially in smaller machines - the iron losses are usually small, and the copper losses dominate. At higher speeds, the losses in the iron become important, and they are a key factor in the design of large high-speed SR machines. Relatively high windage losses at high speeds result from the rotor structure. For a given grade of steel and lamination thickness, the iron losses depend not only on the frequency of excitation seen by the steel, but also on the magnitude of the magnetic flux and the flux excursions. The magnetic loading of different sections of the rotor and stator steel varies throughout the cycle of operation as the phases are switched on and off, and as the rotor position changes. Furthermore, in a polyphase machine, many sections of the rotor and stator steel carry flux from more than one phase. The problem therefore has

Figure 5.9 Cross section of three-phase six-four SRM



Unipolar excitation is normal for the switched-reluctance machine, and as such the associated power circuits apply winding current in one direction only. There are many different power converter topologies which can be used with the switched-reluctance machine. Discussion of all these variants is beyond the scope here, but a brief summary may be helpful to establish a choice of preferred power circuit. Most power-circuit topologies used for the SR motor have either one or two solid-state switches per phase, each of which requires an actuating circuit to open and close it. In spite of the number of possible power-circuit variations, it will be found that the total VA rating (volt-ampere product) of the switching devices used never falls below a certain minimum value. For example, power-switching topologies using only one power-switching device per phase may be superficially more attractive in cost terms than a two-switchper-phase circuit. But, the single power switch will have to handle at least twice the voltage seen by each device in the two-switch circuit, and so the total VA rating is the same as, or possibly greater than, that for the two-switch case. This rule will be seen in the following examples.

both time and spatial dependency. Modelling and calculating iron losses in the SR machine is a highly complex matter, and obtaining accurate results is notoriously difficult. Good results depend on sophisticated finite-element analysis combined with a detailed knowledge of the machine's operation and of the steel type in use.

Each stator pole must complete one full cycle of energisation for each and every rotor pole that passes it, and hence the frequency of excitation per phase is determined by the product of the number of rotor poles and the speed of the rotor in revolutions per second. That is: fph = (nr S)/60 where nr is the rotor pole count, and S is the rotor speed in revolutions per minute. Since there may be multiple interleaved phases, the total excitation frequency will be given by:

for = (nphfph S)/60

These figures are important in the calculation of motor and power electronics losses, and also in determining the time periods with which the control electronics will have to work.

Single-Switch-per-Phase Circuits
Figure 5.10 shows one phase of a possible single-switch power circuit. The power switch S is here shown as a MOSFET, but could equally be an IGBT, GTO, bipolar transistor or other switching device. Two parallel rails are required, the mains D.C. supply (+ VD.c.) being connected across C2. When S is closed, VD.c. is applied to the phase and flux increases accordingly. When S is opened, the phase current commutates to diode D, and the winding sees a negative voltage equal to the potential across C1. Flux and current are then driven down by this negative voltage, which is typically equal to VD.c. in magnitude, as shown, but need not necessarily be so. Only a single-gate drive circuit is needed per phase, referenced to the zero volt rail (which may be connected directly to the signal level controls and, in a low-voltage system, be at ground potential). This can offer savings in low-voltage or nonisolated systems. Simple single-switch-per-phase circuits are occasionally used in low-cost appliance and automotive applications where the saving in the number and/or

The switched-reluctance drive, like a conventional A.C. inverter, controls the motor using power switching devices which connect the various phases of the motor to a D.C. supply of relatively constant voltage. Typically, this D.C. supply will be derived from standard single or three-phase A.C. power lines by means of a bridge rectifier, filter capacitors and, optionally, a filter choke. If dynamic braking is required, the D.C. supply will receive an average current flow back from the SR machine as mechanical energy is converted back to electrical energy. The D.C. supply must be receptive to the generated power, which will initially cause the voltage across the filter capacitors to rise. The capacitors may be sufficient to accommodate transient braking energy, but thereafter it must either be dissipated in a braking resistor or returned to the A.C. power supply through an active bridge. In this respect, the switched-reluctance system is again no different to a standard A.C. drive.



Power Switching Stage

It was shown earlier that the rate of change of flux linkage is equal to the applied winding voltage, and that it is necessary to get the flux, and hence current, to the correct value in the shortest possible time (hence angle). It then follows that, at the switch-on and switch-offangles, we want to force the flux up and down, respectively, by applying the maximum possible winding voltage. When the phase is excited, we further need to be able to control the current to any desired level.

phase winding

+ VD.c


lgate driveLI T circuit T I



Figure 5.10 A single-switch-per-phase topology

Chapter 5.2


complexity of gate drive circuits is attractive. Their use in high-power industrial systems has become less common, however, partly because of the availability of MOS-gated power switches such as MOSFETs and IGBTs which can be operated from simple low-cost gate drive circuits. Additionally, the circuit above has only two switching states and as a result offers less control flexibility than the two-switch circuit, which is discussed later. Note that the switch sees the full winding current and (when returning energy to the supply via D) twice the supply voltage VD.c.. The circuit of Figure 5.10 requires a centre-tapped power supply, which can conveniently be obtained by splitting the supply reservoir/filter capacitance into C 1 and C2, as shown. The motor will draw energy from C2, some of which will be returned to C 1 at the end of the electrical cycle via the diode D. Some means must therefore be provided either to dissipate energy from C 1, or to return energy from C 1 to C2. If this is not done, the respective capacitor voltages will drift towards an unbalanced condition, and ultimately most of the total supply voltage will appear across C 1. This circuit is sometimes referred to as the C dump, presumably because a capacitor (here C1) is used as a dumping ground for unconverted energy returned from the motor. An additional switching device is often used to return this energy from C1 to C2, so that an n-phase motor will require (n + 1) switches.

and returning it to C2, the average current flow in the two capacitors can be equalised and the average (D.C.) voltage balance will be maintained if the phase currents are identical. The circuit is in fact self stabilising if the phase energisation is not current limited, since an increase in capacitor voltage will lead to an increase in phase current. If current control is used, h o w e v e r - as it usually will be at low and medium speeds - then additional measures are needed to ensure that voltage balance is maintained. Note that the gate drive circuit for the second phase (B) is not referenced to the 0 V rail, and in fact its common-mode potential swings rapidly between VD.c. and 0 V as the switch $2 is opened and closed. The gate driver must be carefully designed in order to deal with these conditions. (The common-mode voltage may swing at a rate of many thousands of volts per microsecond using modem IGBT or MOSFET switches.) Each switch in Figure 5.11 experiences the full supply voltage (VD.c.), but the windings only see half this voltage during flux up and deflux. This means that, for the same rate of change of flux as before, the number of turns on the windings must be halved, necessitating twice the phase current for the same working flux level. The total switch VA per phase is therefore (once again) 2 x VD.c. x IpH.

Multiple-Phase Operation
If the motor has two - or any even number o f - phases, then the drift of capacitor voltage may be eliminated by connecting the second phase as shown in Figure 5.11. The D.C. supply can then be conveniently connected across C1 and C2 so that the capacitors form a centre tap or midpoint on the power supply. The idea can be extended to four phases, six phases etc. in ti~e obvious manner, by connecting equal numbers of the phases as high-side and low-side switches. By drawing the second phase's energy from C1

Single-Switch Circuit using Bifilar Winding

Another well known single-switch topology uses a bifilar (two strand) winding in the motor, in which the phase energy is supplied via one winding and returned by the other, Figure 5.12. This circuit has the advantage that all the power switches can be referenced to the zero volt supply rail, which is attractive for low-voltage systems, e.g. automotive applications. Because the two halves of the winding are connected in the opposite sense to each other, transformer action means that



+ VD.c./2
phase A

from control m

gate drive circuit


D2 2

Figure 5.11 Single-switch-per-phase topology with two phases (H circuit)



SWITCHED-RELUCTANCE MOTORS CONTROLLERS: AND Power Electronics for the Switched-Reluctance Motor
phase B gate driver of Figure 5.10, withstand rapid swings in common-mode potential as $2 opens and closes. Each power switch is exposed to the full D.C. supply voltage VD.c., and switches the full winding current Iei-i. The total VA rating per phase is therefore again 2 VD.c x IH-I. The power electronics rating is therefore no worse than that for the single-switch cases (and better than that for bifilar). Although additional power switches are required, the total switch heat losses are distributed among two packages per phase, which eases thermal management, especially in highpower drive systems. More important in many applications is the fact that the asymmetrical bridge circuit offers additional control flexibility, which is useful in managing the phase current and acoustic noise of the machine. Note that, because the motor phase is connected in series with the two switches across the power supply rails, this power circuit is not susceptible to the shoot-through fault condition which can arise in conventional inverters. Indeed, closing both switches at the same time is a prerequisite for normal operation of the machine. This can simplify protection of the power electronics, although it should be noted that a shoot-through fault would occur if the terminals of the motor phase were inadvertently short circuited, e.g. by incorrect user installation. Figure 5.14 shows an outline schematic for a complete threephase SR power converter, with no braking.

each switch sees theoretically twice the D.C. supply voltage during energy return (deflux). However, the electromagnetic coupling between the two strands of the winding is inevitably imperfect, and this results in sometimes large additional voltage overshoots across the switches when they open. This means that the switches must be of significantly higher voltage capability than the theoretical 2 VD.c., but still carry the full winding current. The VA rating therefore in practice always exceeds 2 VD.c IH-I per phase. Furthermore, the machine winding is more complex and requires additional connections to the power electronics, both of which increase cost. The efficiency of a bifilar machine is marred because, at any instant, the phase current is flowing through only a part of the available conductor area in the winding. The copper fill factor (i.e. the proportion of winding space actually containing copper) is also a little worse, because of the additional, separately insulated, winding required.
Two-Switch Asymmetrical Bridge

A two-switch asymmetrical bridge topology is the most common one used for switched-reluctance motor control. Two switches and two diodes are employed, as shown in Figure 5.13. Here, the gate drive of S 1 is referenced to the zero volt rail (the low-side switch), and $2 has a floating gate drive (the high-side switch). The drive circuit for $2 must, like the

+ vo.c

+ Vo.c.


S2 ~ L.[gatedrive L

/circuit /

phasewinding ,gatedrive U ~'circuit I Sl D2

gate [ d"ve _

l T



Figure 5.12 Bifilar SR power electronics topology RLI

Figure 5.13 Asymmetric bridge power circuit

' 1o 2o




phase C

, ~

-=O1 Q;




~ _~~~6



A.C. line input







Figure 5.14 Basic three-phase power converter circuit




simple thermal



simple rotor: no windings no conductors no magnets no commutator

mechanically robust


~ii!i~ii~i!iJi~i!iii!i~iii~s~i~i~ii!!ii!iiiliiliiiii~iiiiiiiiiiiiiii!~i~iiiii:~~~```~````~`~`~``~~~iiii! ili~ii ii ii \ no phase overlaps

high dV/dtwithstand lowercapacitanceto frame reduces RFI

short winding overhangs lower hot spots less movement

Figure 5.15 Key advantages of the switched-reluctance machine


The switched-reluctance machine combines the advantages of relatively high specific output with great simplicity and robustness.

means that the stall endurance of the machine is limited by the thermal time constants associated with the stator. These are generally long, owing to the large stator mass, and the SRM performs well under conditions of prolonged stall. By way of a summary, some of these key machine-related advantages are illustrated in Figure 5.15.

Rotor Construction
Because the rotor carries no windings, magnets or conductors of any kind, the machine is well suited to high-speed operation, and is a good choice for use in harsh environments and/or at high temperatures. Heating of the rotor is confined to eddy-current and hysteresis losses in the steel and, as a result, the rotor runs relatively cool in the majority of applications, enhancing bearing life.

Electronics and System-level Benefits

The SRM requires the use of power electronic controls for commutation and control, in many ways similar to the inverter used to vary the speed of an induction motor. However, in contrast to the A.C. drive, the motor does not require sinusoidal supplies in order to operate efficiently. As a result, the power converter used with the SRM does not need to switch at high frequencies. This reduces switching losses in the power semiconductors at low motor speeds, and is especially useful in medium and high-power drives (e.g. 10kW+) where switching losses can otherwise be significant. To avoid tonal components in the drive's acoustic noise, it is common to use current-control schemes with randomised or spread-spectrum switching frequencies. At higher speeds, the power semiconductors switch on and off only once per electrical cycle of the machine (Figure 5.7). The switch turn on occurs at zero current (which implies zero loss), and turn off may occur at a current less than the peak phase current due to natural roll over of the current. The switching losses at high speeds are therefore negligible. The relatively low electronics switching losses, combined with high torque per amp of phase current, mean that the power semiconductor ratings in the switched-reluctance drive can be somewhat lower than those in conventional systems. The phases of the switched-reluctance drive system operate independently of each other, and in the event of a fault developing in one phase, the others are able to continue to produce torque as normal. This gives the machine an almost unique inherent fault tolerance with the ability to limp home in the event of a partial failure. The SR machine is capable of yielding very high overload torque; its ability to do so is really limited only by the

Stator Construction
The stator is also very simple and robust, requiring only short-pitched coils which are placed over the salient stator poles, and which can easily be prewound on a former or bobbin. The stator windings - unlike those of an induction motor - are not distributed over many slots, and the phases do not cross each other in the end-winding region. This largely eliminates the risk of a phase-to-phase insulation failure. The simplicity of the coils allows the end windings to be much shorter than those typically found in induction motors, and the losses associated with the end windings (which do not contribute to the output of the motor) are reduced. This improves efficiency and allows the construction of relatively short stack (pancake) motors with minimal penalty on specific output. The winding construction tends to yield a lower capacitance to the frame than does a conventional A.C. motor, typically 20 to 30 per cent less. This improves electromagnetic compatibility and reduces radiofrequency interference, because coupling of high-frequency currents to the stator is somewhat reduced. As mentioned above, with the possible exception of machines operating at very high speed and at high power, the machine's losses are concentrated in the stator, which is relatively easily cooled. Thermal management of the machine is therefore relatively simple. Furthermore, the fact that the rotor heating is minimal, especially during stall,


SWITCHED-RELUCTANCE MOTORS AND CONTROLLERS" A d v a n t a g e s o f t h e S w i t c h e d - R e l u c t a n c e System

Figure 5.16 High-performance SR Drive T M for textile applications (courtesy of Picanol, n.v., leper, Belgium) (SR Drive is a registered trademark of Switched Reluctance Drives Ltd)


o~ 800 .m


._o 70(D

E >, 6 0




2000 2500 speed, min-1




Figure 5.17 System efficiency (mains to mechanical output) of 7.5 kW SR Drive T M system at rated output (SR Drive is a registered trademark of Switched Reluctance Drives Ltd) thermal time constants associated with the stator windings. This high peak overload torque capability, combined with relatively low rotor mechanical inertia (due to material being removed at the outer diameter to form the salient poles), means that the very high rates of angular acceleration are possible. Torque control stability and dynamics do not depend on precise motor parameters, and good dynamic response is inherently obtained. The electrical time constants associated with the windings are generally short, and there is no need (unlike with the vector-controlled induction motor) to keep the machine fully or partially fluxed when it is lightly loaded, in order to secure rapid response times. This can bring efficiency and thermal benefits, especially in applications which may spend much of their time lightly loaded, and operate only at full torque to skip quickly from one speed or position to another, e.g. as shown in Figure 5.16. Careful design and optimisation of control parameters can yield excellent system efficiency over a wide range of torque and speed, e.g. Figure 5.17. This shows the true system efficiency (measured as mechanical power output divided by the raw A.C. electrical power input to the power electronics) of an IEC 132-frame 7.5 kW SR Drive system. The drive is delivering a constant 50 N m torque from stall to 1500 min-1 and a constant mechanical output power of 7.5 kW from 1500 to 4000 min- 1. The switched-reluctance system is capable of operating over a wide speed range at constant power without the efficiency or power electronics cost penalties associated with permanent-magnet and induction-motor technologies under these conditions. This is proving to be an attractive benefit for automotive and traction applications, where significant interest is being shown in the machine, especially when these merits are considered along with its robust and simple construction.

Although the machine is very simple in both construction and concept, in practice saturation of the steel means that it is highly nonlinear. As a result, it does not lend itself to the relatively straightforward and well understood design methodologies used for more traditional machines. The

Chapter 5.2


analysis, design and control of high-performance switchedreluctance motors is a complex matter, and without the necessary tools and expertise it is easy to design a system of poor or indifferent performance. The motor and controller must be designed as a system and operated together. The motor cannot be connected directly to a mains power supply if its associated electronic controller fails. Because each phase operates independently of the others, two cables per phase are usually required, although this is a function of the power circuit used. With the most frequently employed asymmetrical bridge power converter, a threephase system requires six motor leads. It has been said of the SRM that it requires a short air-gap length to perform well. It is true that the SRM can benefit from a short air gap, which increases Lmax and hence the difference between Lma x and Lmi n. It is also true that excellent results can be achieved using air-gap lengths equal to, or even greater than, those normally employed in the production of standard induction motors.

such as current profiling. This involves modulating the phase current (or flux) with respect to the rotor angle, so as to cancel out the torque variations which otherwise occur. Torque ripple can be reduced to a few per cent using such techniques. Further smoothing is possible, but is probably best accomplished using adaptive or self-tuning techniques which are able to take into account the inevitable minor variations from one motor to the next which will occur in mass production. Adaptive techniques can also minimise the effect of rotor position measurement error due to encoder imperfections or misalignment.

Acoustic Noise
Acoustic noise was a well known problem with early SR systems. Some of this may have been due to torsionallyinduced vibration (from torque ripple, as discussed above), but it is important to distinguish this source of noise from that due to the large normal forces experienced by the stator poles in any electrical machine. These tend to distort the stator from its desired round shape, Figure 5.18. Because the normal forces exerted on each stator tooth pulsate (as the rotor turns and as each phase is switched on and off), the stator will tend to vibrate in various modes at the phase frequency and harmonics, hence radiating acoustic noise. Noise due to stator distortion and vibration can be minimised by careful attention to mechanical design (including the choice of pole structure and considerations of back-iron thickness), as well as by electronic means. The latter can make a significant impact by controlling the spectral content of the normal forces seen by the stator poles, and hence that of the resulting surface acceleration which generates the noise. Care must also be taken with general mechanical design and stiffness. Good concentricity must be maintained if low noise performance is to be attained. Sufficient progress has, however, been made in the control of acoustic noise to permit the use of SRMs in high-volume domestic appliances, e.g. washing machines, where noise is a major concern, Figure 5.19.

Torque Ripple
The torque produced by the SRM, when excited by constant phase currents, is not inherently smooth even if three or more phases are used. The SRM produces some residual torque ripple; for a three-phase system this may typically amount to 30 to 40 per cent (expressed as the peak-to-peak excursion divided by the mean torque). Contrary to popular wisdom, however, there are surprisingly few applications where the presence of such torque ripple causes noticeable process or control problems. Indeed, SRMs have been successfully used in traditionally demanding applications, such as positioning drives and for web tensioning, without recourse to torque smoothing by exceptional electronic means. Care must, however, be taken that higher-frequency components (harmonics) of torque ripple do not cause undue acoustic noise through exciting resonances in the load or other associated mechanical arrangements. If the natural torque ripple really is unacceptable, the torque produced by the SRM can be smoothed using techniques

Figure 5.18 Exaggerated normal-force bending of stators


SWITCHED-RELUCTANCE MOTORSAND CONTROLLERS:Disadvantages of the Switched-Reluctance System

! i
Figure 5.19 Low-noise SR Drive T M system for domestic washing machines (courtesy of Maytag Inc.) (SR Drive is a registered trademark of Switched Reluctance Drives Ltd)


Stepping motors are a group of motors characterised by the fact that the shaft rotates in angular steps corresponding to discrete signals fed into a controller. The signals are converted into current pulses switched to the motor coils in a specific sequence, and the motor acts as an incremental actuator, which converts digital pulses into analogue output shaft rotation. The speed of rotation is dependent on the pulse rate and the incremental step angle, whereas the angle of rotation is dependent on the number of pulses fed to the motor and the incremental step angle.

This dependence means that the motor is eminently suited to open-loop position and speed control, within limitations.

There are three basic types of stepping motor:

permanent magnet - PM variable reluctance - VR hybrid



Figure 5.20 Step sequence for a PM rotor stepping motor

Chapter 5.3




Figure 5.21 Step sequence for a VR rotor stepping motor

The Permanent-Magnet Motor

The permanent-magnet motor has a laminated, slot wound, stator, usually with two, three or four phases. The rotor mounted on a bearing in each end frame is usually a solid cylinder magnetised in a two, four, six or eight-pole configuration. Rotation of the motor shaft is achieved by switching currents between coils to produce a change of the electromagnetic field alignment. By controlling the sequence of switching, the field can be made to rotate within the stator bore and the rotor will rotate in synchronism. When the electromagnetic field is in a fixed position, the rotor will be aligned with it and the torque exerted will be zero. By switching current to the next set of coils, the electromagnetic field will move to align with those coils and thus out of alignment with the rotor. This will exert a restoring force on the rotor to bring it into alignment again, which results in a rotational torque being developed at the shaft. Compared to the VR motor (described below) the PM version develops a higher torque due to the magnet flux strength, and it has a preferred axis of alignment because of the polarised rotor. It should be mentioned that when the rotor is 180 out of alignment the torque is zero, but this is a very unstable position and any small movement to either side of 180 results in the rotor returning to its correct alignment position. When the stator is not energised the PM rotor still tends to align itself with a pair of poles, normally remaining in the same position as when last energised; this is known as the detent torque. Care must be taken that the current rating of PM steppers is not exceeded, otherwise demagnetisation of the rotor can occur. When stepping in synchronism with the rotating stator field a back e.m.f, is generated in the field coils due to the PM rotor. The amplitude is proportional to the stepping rate, which reduces current input as speed increases. Further, as a consequence of the inductive nature of the windings, the current input reduces as the switching frequency increases. There are a number of techniques, generally known as current forcing, to overcome this problem, which are described in the following sections.

soft iron material with a number of equally-spaced poles, which form paths of minimum reluctance in the overall magnetic circuit. Since the rotor is not magnetised, polarisation is determined only by the stator excitation and the step angle will be a function of the number of rotor poles compared to stator poles. These are often not the same number. Although the VR motor has a lower static torque rating compared to that of the PM motor, the absence of the permanent magnet in the rotor allows a higher speed range to be achieved for similar input. Furthermore, the detent torque is almost zero (a small amount may be present depending on the type of rotor and its remnant magnetic flux), so the motor can be moved freely when not energised. There is not a problem of demagnetising the rotor, so torque output can be uprated for short duty, although magnetic saturation of the stator and winding temperature will determine the limits.

The Hybrid Motor

The hybrid motor is a combination of the PM and the VR motors in that the rotor has a permanent magnet core

i ~' I

--- ~

' :

| ~"F""

stator winding I rotor

| | i | i i i |

i . . . . . . . . . . . . . . . . .

. . . . . 1-I. . . . . . .
i |

| ! t . . . . . . ~ . . . . .

_ i

, i


flux due to permanent magnet in rotor

i i i i i

The VR Motor
The variable-reluctance motor is similar to the PM motor, except that the rotor is not magnetised; it is formed of


Figure 5.22 Hybrid stepper, sectioned to show magnetic circuits


STEPPER-MOTORDRIVES: S t e p p i n g - M o t o r


The most common configuration of the hybrid motor is a four-phase wound stator, each pole having teeth spaced equally at 1/48th revolution, and the rotor with 50 teeth equally spaced. By switching currents to each pole pair in sequence, the rotor moves round to align itself with the nearest set of teeth and the angle of rotation is given by: step angle - 360/(rotor teeth x stator teeth) = 360/(50 x 4) -- 1.8 The above assumes unipolar switching, which uses the simplest circuits, but by half-step switching the angle can be reduced to 0.9 and by ministepping the angle can be reduced by a factor of ten or more.

Figure 5.23 Hybrid stepper sectioned on line AA' of Figure 5.22

with soft iron end pieces. The magnet is polarised axially and the magnetic path flows out radially through the end pieces and back via the stator yoke. So the principle of operation is quite different from that of either the PM or the VR motor. Nevertheless, the generation of torque is due to the forces involved in the alignment of rotor teeth with stator pole teeth and rotation controlled by switching of current to the coils in a prescribed sequence.


The performance characteristics of stepping motors are significantly affected by the type of drive employed and the way in which the phases of the motor are energised. Typical logic modes are shown below for three and four-phase VR and PM steppers. In Figure 5.24 the number preceding the mode letter indicates the number of motor phases to which the mode applies. Mode A (unipolar, single coil) - only one phase at a time is energised, i.e. one phase is switched off at the same time as the next phase is switched on. The motor will execute one basic step for each input pulse.

stepping modes mode 2RA


mode 2RAB

mode 4A

2.1 2 2 12

1.2 2 2.1 4

mode 4B 4.1 1.2

mode 4AB

mode 3A 1.2





mode 3B 3.1


mode 3AB 1



mode 3RBC 1.2 1.2.3 1.3 1.2 1.2.3 ",,~,,,," 1.2.3 2 . 3 ~ - 2 . 3



3 2.3

I 2.3



i ',


1.3 ~.2.3 1.2

Figure 5.24 Logic modes for three-phase and four-phase VR and PM stepping motors




E, C

The resulting pattern of currents is given in Figure 5.26.



Reversing the sequence will cause a reversal in the direction of rotation. With the switching pattern of Figure 5.26, the motor steps through its basic or full step angle of 1.8 for the hybrid stepper. The pattern can be modified as shown in Figure 5.27, introducing half stepping, i.e. a step angle of 0.9 with the following advantages: higher resolution smoother drive rotation step resonance minimised reduced settling time


Figure 5.25 Unipolar switching

Mode B (unipolar, two coils) - two phases are energised at any one time. This mode will produce higher holding and dynamic torque and will reduce rotor oscillations, but power input to the motor is double compared to mode A. The motor will also execute one basic step for each input pulse. Mode AB (unipolar, half step) - this is a combination of modes A and B in which the motor phases are energised sequentially in modes A,B,A,B,A,B,A etc. The motor will execute half the basic step for each input pulse. This mode has the advantage of a smaller step angle which enables the motor to operate at higher pulse rates. Mode RBC (partial bipolar) - this is a more complex mode requiring reversal of coil polarity as indicated in the diagram by a bar placed above the phase number. This mode is applicable only to a three-phase PM motor and enables it to operate at a quarter of the basic step angle.

The sequence of the switching pattern is also reversible resulting in reversed motor shaft rotation. Unipolar switching results in current flowing in one direction only through the winding. Although the direction of flow is not important, at best only two of the four motor coils are energised at any one time. Improved performance can be obtained by energising all four coils for each full step. This can be achieved with bipolar switching.

By connecting pairs of coils in series or in parallel and reversing the flow of current in coils, use can be made of all four coils to give a higher torque. It should be noted that with coils connected in series the inductance is four times that of coils connected in parallel. So although the low-speed performance is similar, the current

unipolar full-step switching step number phase


2 3

0 0

1 0

1 1

0 1


Figure 5.26 Current pattern for unipolar switching

unipolar half-step switching step number phase

2 3 clockwise 4 5

0 0 0 0

1 1 0 0

0 1

0 0 anticlockwise



0 0 0

1 1 0




Figure 5.27 Current pattern for half-step unipolar switching


STEPPER-MOTOR DRIVES: B i p o l a r S w i t c h i n g

Figure 5.28 Parallel connection

A v

Figure 5.29 Series connection

full step step number 1 2 3 4 phase A +1 -1 -1 +1 phase B +1 +1 -1 1 step number 1 2 3 4 5

half step phase A +1 0 1 1 1 phase B +1 +1 +1 0 1




Figure 5.30 Current pattern for bipolar switching


R2 R2




Figure 5.31 Four-step L/R drive configuration

Figure 5.32 Eight-step L/R drive configuration

Chapter 5.3


and hence the torque, will fall off earlier at higher stepping rates for the series-connected configuration. The choice of coil connection should take this into account in addition to the current and voltage capacity of the controller.

control without the need for position or velocity feedback devices, i.e. open-loop operation. But, there are certain operating characteristics which have to be respected to obtain satisfactory performance. The stepping motor behaves like a slave motor, in fact in the earliest designs such motors were known as repeater motors, rotational velocity and position being determined by an external pulse source. The dynamic characteristics are shown diagrammatically in Figure 5.33; the top curve shows the pull-out torque and the bottom curve the pull-in torque, the area between the two curves representing the slew range in which the motor cannot be started or stopped without losing synchronism. Operation in this region requires controlled acceleration and deceleration, i.e. ramping. The optimum ramping rates and stepping rates are determined from motor performance data, after allowing for load friction and inertia, in order to maintain synchronism with the input pulses, over the total traverse. The ramping down time can be less than for ramping up, because load friction will assist in the deceleration.


As stepping increases, the coil inductance and back e.m.f. limit the current and the torque falls off with increasing speed. Various techniques are used to overcome this problem. In the simpler drives, such as the L/R type, voltage forcing is employed by using a supply higher than the coil voltage and limiting the standstill current by a resistor in series. This reduces the time constant of the coil/resistor circuit and allows a faster rise time at each switch sequence. Owing to the IR loss in the resistor, the use of L/R drives is limited to relatively low-power motors. When operating a unipolar drive in the four-step mode (full stepping), the use of two resistors rather than a single resistor will result in a better damped drive capable of operating at higher stepping rates, yet the overall standstill efficiency is the same. With the eight-step mode (half stepping) it is necessary to achieve a balance of torque between adjacent steps, that is when one coil is on and two coils are on, so a three-resistor network is used. Voltage forcing by means of series resistors may also be used with the bipolar drives.

Effect of Inertia
The inertia of the load connected to the motor shaft can have a significant effect on dynamic performance. Best performance is obtained with a load inertia equal to that of the motor rotor and higher values of load inertia will result in an increase in the mechanical time constant. As a consequence of this the motor will exhibit a lower pull-in rate and a general decrease in dynamic response.

To overcome the losses in voltage-forcing resistors and produce a more efficient drive, essential for the larger motors, a chopper-regulated technique is adopted. A highvoltage D.C. source is used to obtain a fast build up of current in the motor coils. When the nominal level is reached the switching device is turned off, but the current tends to be maintained by the motor coil inductance. When the current has decayed below the nominal value, the switching device is turned on again. This process of chopping is repeated with increasing frequency as the motor step rate increases, until the time constant of the motor coil circuit does not allow the current to reach nominal value, when the supply reverts to a constant voltage source. The chopping principle can be applied to both unipolar and bipolar drives.

An important point to mention is the phenomenon of resonance suffered by all stepping motors, to some degree or other. This can occur under two particular circumstances.

torque holding to rq__.~ pull-out I torque .[

pull-in torque

star range I


pulse \ range ~ rate

A further variation of design is the bi-level drive in which the supply voltage is maintained at a high level for acceleration and deceleration, but is reduced to a lower level once the motor is running at a constant load/speed. This permits higher peak ratings to be applied to motor and drive combinations.

Figure 5.33 Dynamic characteristics

o o >

ramou? constant velocity


The stepping motor is a very versatile device with a wide range of speed and torque, plus the ability to remain under

time Figure 5.34 Effect of inertia



step position


and velocity-control systems. But as control systems became more sophisticated the advantage of closing the loop became evident. Since the stepping motor is a digital actuator it is usual to employ a digital feedback device, such as an encoder or resolver with R/D conversion. It is also possible to design units in which the feedback component is close coupled to the rear of the stepper.

Space-Rated Steppers
time Figure 5.35 Resonance effect
For operation in outer space long-term reliability is paramount and the simplicity of the stepper-motor construction, coupled with the integrity of digital control, makes stepper systems a favoured choice. What is not so obvious is the attention to detail in the manufacture of motors for space, the care required in the finishing of individual parts to close tolerance, in the handling and transit between operations and the care of assembly to ensure that there is no bruising of mating surfaces or stress applied to threads, bearings or wire connections. Following thorough cleaning of all parts the units are assembled and tested in a clean room maintained at a positive pressure to exclude foreign particles down to a size of 1 micron. To further ensure the long-term operational reliability of the system it is quite normal for the system to be dual channel so that, in the event of failure at any point, operation can continue by switching to the second channel. Various degrees of redundancy are used, in some cases a complete redundant system, but in other cases the motor may be provided with a redundant rotor/stator assembly in the same housing. This is known as a tandem unit and then additional care has to be taken in correct alignment of the two rotor/ stators to ensure a smooth transition on switchover.

The first is when, at the moment the drive pulse is switched on, the rotor lags so far behind the new step position that the torque developed is not sufficient to accelerate the rotor to the new position. This corresponds to point D in Figure 5.35. The second is when the rotor swings so far forward that the new step position is actually behind the rotor, represented by point B. This will produce a reverse torque, which will slow down the rotor so that it will not be able to respond to the subsequent pulse, and may even result in a reversed direction of rotation. There are ways to overcome this phenomenon. Increasing load friction can damp it out, but at considerable expense of motor performance. Viscous friction dampers can be used, which affect the transient response but not the steady-state performance. Various electronic circuit modifications have been devised to reduce oscillation and even changes to the motor magnetic field geometry can be economic for largequantity applications. In practice, by using half-stepping drive circuits and arranging to operate motors away from the resonant speed regions, combined with the inherent damping of load friction, resonance does not usually present too much of a problem. If the application permits, altering the acceleration rate will often avoid resonance effects.

Fuel Control Actuators

A rather similar application to the above is for tandem steppers to control the fuel flow in gas turbine engines. In this case there are other considerations, such as rotation over a precisely specified angle with stops to prevent complete engine shut down, the possibility of complete fuel contamination, the need for the motor to survive in the event of an engine fire and reliable operation in ambient temperatures from - 5 0 C to +125C.


Originally the great advantage of the stepper motor was its ability to operate in the open loop mode in position-control

Practical Drives

i i i i i i i i i i i i i i i i i !i

iiii!i!ii!ii!i!ii!iiiii 1 iiiiii!iiiiii!iii 2
..!i............ 3 ..
i i !i i i i!!ii ~!~i~!i!~i~i~i i i~ i i i i i!i!i!i i i i i i i i i i i



141 145 151 169

i i i i i i!i i i i i i i i i i i l



Thus far this book has dealt largely with the theory of variable-speed drives. The capability of commercially available drives deviates from the theory, both in regard to practical limitations of performance at the motor shaft and also in regard to issues like control input performance. It is important to also recognise the importance and value of additional features that are incorporated in modem commercial products. This chapter deals with some of these practical aspects by describing commercial products, taking as examples variable-speed drives manufactured by Control Techniques.

Before considering the detail of specific drive products, it is worth considering a few aspects of commercial drives which may be considered common. When considering the theory of drives it is convenient to limit the control aspects to torque, speed and possibly position loops. In reality a large number of other features are built in to modem digital drives as standard. As highlighted in the Preface, the purpose of this book is not to endorse or promote the products from this manufacturer. Equally the publisher does not endorse products from Control Techniques or any other manufacturer.



Typically, the following functionality is incorporated in the software structure: selection between different types of speed and torque reference programming of analogue and digital inputs/outputs selection of different ramp functions setting up of the drive speed and current loops monitoring of drive status parameters, trip log programming of special application features such as different types of torque control mode, digital lock function etc. programming of additional logic blocks and PID functions which are available for user application set up

It can be seen from the Figure that an offset to the demanded speed can be input using parameter 1.04. This can be set to a fixed value or can be programmed within the drive on an option card to be a complex variable based upon any parameters input by the customer or a derived value within the drive itself. Consideration of the control diagram will start to provide an insight into some of the control capabilities of a commercial drive. These capabilities can often be enhanced by the addition of option boards. These may provide additional connectivity in the form of a fieldbus interface, or additional control capability in the form of a second processor. These options are discussed in sections 6.3 and 6.5. Although modem drives can be operated directly from the product-mounted keypad, it is more common to integrate drives into larger automation systems. In such systems the interface to the drive may be through a digital serial

Figure 6.1 provides an overview of the typical software structure of a modem, digital drive. Although this specific diagram is for the Control Techniques' Mentor D.C. drive, the basic structure remains broadly the same for D.C. and A.C. drives.


speed offset 1.04 I post offset /


1"-"3 i biplar

I "10 I


I,.,31 ref inch


reverse select


ramp 2.02 enabl~ post ramp ref 2.04 2.05 2.06 2.07 PID

I 1.05 ]
inch ref preramp ref speed output 4.04 4.05 4.06 current limits final current demand

Lix., i J
maximum speed forward refon I 1"06 I

i !



1.11 I - /

I~ ( ' ~ ~ - <

speed error

maximum 11-09 I
firing angle

3 oo > - - 3.~0



speed reverse

i " ~ ' ' ~ / " "".,. feedback - V """ ~losed loop only

t current actual current

Figure 6.1 Typical control structure of a modern digital drive

< user

drive >




[ 8.,61 an read rite

'bit' parameter invert 8.26 I

Figure 6.2 Programmable digital input


0 0

C h a p t e r 6.1


link, or hard-wired digital and analogue inputs and outputs. Such inputs and outputs are typically of the following form.

DIGITAL OUTPUT Programmability

An example of a digital output function would be an at-speed signal. The output state changes when the actual value reaches the demand signal. A typical programming flow is shown in Figure 6.3.

DIGITAL INPUT Programmability

With reference to Figure 6.2 it can be seen that by setting the contents of parameter 8.16 it is possible to direct the input command at terminal 26 to any read/write bit parameter in the drive. The logic of the input can be reversed via parameter 8.26.

Typical Specification
Digital outputs are usually referenced to the drive control zero volts line (0V). There is often integral overcurrent protection to ensure that damage is avoided if the output is inadvertently short circuited. The output is invariably taken from a supply rail ~24 V: output voltage + ve control rail less voltage drop across transistor switch and current sense resistor (IEC 1131-2) current capacity typically of the order of 100 mA

Typical Specification
Digital inputs are usually referenced to the drive control zero volts line (0 V). There is protection afforded by a series resistor, clamp diodes and a filter capacitor: input threshold 10 V + 5 V hysteresis 0.3-0.8 V(IEC 1131-2) mode positive logic (IEC 1131-2) or negative logic The response time of digital inputs is very dependent upon the way in which the software is implemented. The inputs are scanned by the control software and it is this scan time that is important to the response time. As it is a scan, the time quoted is invariably the time between scans and is therefore the worst case. It is common to have a number a fast inputs ( < 345 gs) for use in time-critical applications such as limit switch monitoring; others may be relatively slow (~7 ms).
< user 33 drive >

Again the response time of digital inputs is very dependent upon the way in which the software is implemented.

ANALOGUE INPUT Programmability

With reference to Figure 6.4 it can be seen that the analogue speed reference can be input via terminal 3, via a voltage reference selector (parameter 7.26) and scaling (parameter 7.20) to the speed demand parameter in the speed loop (parameter 1.17) by means of the speed reference destination parameter (7.15).

+24 V sou rce

19 91
read only 17 ST3


i i

volt 0-10 V current speed 4-20 mA 20-4 mA 0-20 20-0

Figure 6.3 Programmable digital output

< user TB1 1 +10V 0V -10V currentloop mode 17.281 1 1 ~ / drive >

read only scale

destination: any read/write parameter default = 1.17

analogue speed




Figure 6.4 Programmable analogue input


GENERAL" Analogue Input

Typical Specification
The specification for analogue inputs does vary considerably with the type of drive being considered and reference to the specific drive specification is important. As a guide only we can consider precision analogue inputs which would be associated with high-performance drives, and general analogue inputs which may be associated with products for less demanding applications or as auxiliary analogue inputs on a high performance drive: (i) Precision analogue inputs - usually these inputs accept differential input voltages: input voltage gain accuracy linearity resolution full-scale asymmetry zero crossing error input impedance +IOV 1% O. 1% of actual input delta of 150 gV will give a change 0.1% < 300 gV > 10 kf~ (differential) > 1 Mf~ (common mode)

It is usual to incorporate some filtering, equivalent to approximately 30 ps to ensure a reasonably smooth output signal. It can be seen that the features of a modem electronic variable-speed drive embrace far more than the control of torque, speed or even position. Although it is not possible to discuss all features of a modem drive, in this section we hope to reveal some of the less obvious features of commercial drives. The utilisation of some of these features is highlighted in Chapter 12 where applications are discussed. Typical environmental conditions for standard commercial drives are"

Ambient temperature: 0C to 40C (32-104F). At ambient

temperatures above 40C (104F) derate 1.5 per cent per C up to 55C (131F).

Storage temperature: - 4 0 C to 55C ( - 4 0 F to 131F). Maximum storage time" 12 months. Rated altitude: sea level to 1000m (3300ft). At heights
above 1000 m reduce the full load current by 1.0 per cent for each additional 100 m (320 ft).

(ii) General analogue inputs - usually these inputs are referenced to the control zero volts: input gain accuracy resolution input impedance -+- 10 V, 0-20 mA and 50 f~-5 kf~ 3% 10 bit (A of 10 mV will give a change) 100 kf~ (voltage mode) < 200 f~ at 20 mA (current mode)

Relative humidity: noncondensing to 85 per cent at 40C. Degree of ingress protection" IP 00 (unit for building into an
electrical enclosure).

Starts~hour (A.C. drives): unlimited by electronic control;

_<20 by interrupting the A.C. supply.

ANALOGUE OUTPUT Programmability

In the example shown in Figure 6.5 the analogue output is programmed to provide a meter signal. It could equally well be used to provide a speed reference, or other process variable, to another drive or process equipment.

Speed reference: - 10 to + 10 V or 0 to 10 V; 4 to 20 mA; 20 to 4 mA; 0 to 20 mA; 20 to 0 mA; digital speed input. Serial communications interface: RS-485 serial port, optically isolated.

Materials: flammability rating of main enclosure - UL945VB; glands - UL94-V0.

Typical Specification
Analogue outputs are usually referenced to the drive control zero volts rail: modes range gain accuracy resolution voltage and current + 10 V and 0-20 mA 3% 10 bit

Vibration (random)" packaged and unpackaged- tested to 0.01 gZ/Hz (equivalent to 1.2 g r.m.s.) from 5 to 150 Hz for one hour in each of three axes as in IEC68-2-34 and IEC682-36. Bump~shock: packaged - tested to 40 g, 6 ms, 1000 times/
direction for all six directions as in IEC68-2-29. Unpackaged - tested to 25 g, 6 ms, 1000 times/direction for all six directions as in IEC68-2-29.

The update rate of the output may be better than 100 gs on high-performance drives, and is determined by the implementation of the control software. The analogue output is not usually considered critical in this respect.

EMC immunity: EN50082-2. Following specifications from

IEC61000-4: Part 2 - electrostatic discharge, level 3 Part 3 - radiofrequency field, level 3

< user
+lOV 12

drive >

1 .o81


o o

Figure 6.5 Programmable analogue ouput

C h a p t e r 6.2


Part 4 - transient burst, level 4 at the control terminals level 3 at the power terminals

Part 6 - conducted radiofrequency, level 3

EMC emissions -


Part 5 - surge at the A.C. supply terminals (as specified by EN50082-2 informative annex) level 4 line to ground level 3 line to line

Aspects specific to the alternative types of commonly-used drive system are detailed in the rest of Chapter 6.


D.C. drives are widely used in applications which require regeneration, precise speed control, dynamic performance and constant torque over wide speed ranges. The technology is mature and extremely reliable. Applications that traditionally favour D.C. drives include web handling, winders, extruders, paper production, cranes, hoists, plastic production and wire drawing. Control Techniques Mentor II is a family of D.C. variablespeed industrial drives. All sizes (7.5 to 750kW) share common control, monitoring, protection and serial communications features. Units are available in either singlequadrant or four-quadrant configurations. Both types offer comprehensive control of motor speed and/or torque.

Operating parameters are selected and changed either at the drive keypad, through the serial communications interface, or through MentorSoft, a WindowsrM-based configuration software tool. Mentor has many embedded configurable functions which are easily adapted for virtually any application. These configurable functions include items such as assignable I/O, autotune, feedback selection, ratio control etc.

The rating of a Mentor drive is based on the maximum continuous current that it is designed to deliver. An overload capacity of 150 per cent of the rated current for 30 seconds is

Figure 6.6 Control Techniques Mentor D.C. drive family


D.C. DRIVES"Drive Selection

Figure 6.7 EMC data and wiring recommendations

Chapter 6.2


alternative l~ation ~:

i ~i~i~i!~i~i~i~!ii!~~i~i!i ~~ii i i ~i!i i i i i ~i i i i i i i i !i~i i i i i

ii lilii
RFI filter =:.:::=:: ......................................................... .=: :: :=::...........

i::ii i i i ::.i:.i:.i:i:i i:.i::i::ii i i i::ii i i:~i:iLi i:.ii !iiliiil i=/:iiiii::iii:.iii:!i::::i~:iiii:::~i:.::i::~i ::i :. i

:~~ii:':~:~R~!iii~fii!~:~ i i~~'~'i~~~:~~~"~:~:~i ~:~~i~:~:~~i~:~~'~~~'i:i;~~:~'~':i!'~;i~'i:i '~i ::i ~;~:iii!i : ::;~:~'~ ':~ ':~: '?: ~:;i~:~.i i~ :i;;~i ~: : : ;i: :~ i~

~ii~j~i~!iiiii!~~iiiiiii~iiiii~iii!iiiiiiiiiiii!~i!i~ii~iiii~ii~iiiiiii~iii~i~!~ii! i!~ i


~ i i ~ i i ~ i ~ ~ ~ i ~iii~ i ~ ELI LU "~~ ~) :~~ ~~i'~ ~: ~~'. ~ ~:~d i~(~: ~ ';~,~h~:~:;ii~~ :~ / :~ ~;:~':;i~ ;' /



i .....

~i~i!!~ii~i~i"~i~!~i~i~!!~i~i~i~i~i~i~ ~,



~:i:.:::~i::: i i: : ~i;i:iiiiii.ii!ili::i::iiii! i~:i::ii!!:.:::. i: ::i .i:iiii !i!!i : !i!ilili::;iiii:.:: i!iil


/ /
.............: ...............................................................................................

-< 100 mm (4 in)

ii~iiii~iiii~!~ii:~ii i i !i i?i iii!iii:ii:iiii iii iiii! iiii~dliii~||iiiiii~6~iiiii~6i~iiiii::iii i:~iiii~iiiiil

ii~i~i~X~l~ii~iiiii~i!~ii~i~iiii~~iii~iiiiiiii~~~ ~iii~i~~ii~ii~iiiii~iiii~~ii~i~~~~~iii ~
E1 E2 E3FI+ F2-

J <-100 mm (4 in)

. . . . . . . . . . . . . . . . . . . . . . . . . . .!.i. . . . . . . . . . . .



,~ -,,m

,- B i=m

B m



back plate

alternative safety ground connections for the motor

Figure 6. 7


D.C. DRIVES: Drive Selection

permitted, which may be utilised, for example, during acceleration. If higher maximum currents are required, a drive with a higher continuous current rating should be used, so as not to exceed the maximum current limit setting of the selected unit. Additionally, in applications involving cyclic duty, the r.m.s, current over the whole cycle must not exceed the continuous current rating of the drive.

If optimum response is required from the drive, the current loop, which is the inner control loop of the drive, must be set up to enable the outer speed control loop to function optimally. The dynamics of the current loop are principally a function of the electrical characteristics of the motor as discussed in Chapters 1 and 4. The Mentor has a built in self-tuning procedure for its current loop, providing an easy route to full performance.


A. C. supply requirements mains supply 208-480 V A.C. + 10% three phase voltage: (optional 208-525/660 V A.C. + 10%) input frequency 48-62 Hz autosensing phase rotation nonsensitive D. C. armature output voltage Six-pulse full wave. Table 6.1 shows maximum armature voltage for common mains supply voltages.

PC-Based Commissioning Tools- MentorSoft

MentorSofl is a WindowsTM-based software tool. The functions mentioned above are very similar throughout Control Techniques' complete range of A.C. and D.C. drives. Although the parameters that are used to control and adjust the various features can be manipulated via the drive's onboard keypad, there is a set of software packages available which make this manipulation easier, quicker and far more visual/intuitive.

T a b l e 6.1 M a x i m u m armature voltage

recommended (1Q or 4Q)


The current-loop performance of a D.C. drive is limited by the switching times of the drive as well as the control-loop performance. The following traces were obtained on a Leroy Somer D.C. motor type LSK 1122 S04; 9.5kW, 400V, 28.5 A with tachogenerator feedback.

Mains supply voltage (V) 380 400 415 440 460 480

Maximum D.C. armature voltage (V) 440 450 460 500 510 530

Speed and Current-loop Response

Figure 6.8 shows the current-loop performance during the request for a speed reversal using default speed-loop gains. The armature current rises from zero to maximum in approximately 30 ms. Figure 6.9 shows the optimised performance. Figure 6.10 shows the speed change due to the application of a 100% step load change using product default speed loop

Supply phase sequence Loss of one or more phases of input automatically detected; the drive will run irrespective of input phase rotation Speed feedback motor armature voltage, or tachogenerator, or encoder (pulse generator) full PID speed loop algorithm Current feedback resolution 0.1% current loop linearity 2% Control All analogue and most digital inputs configurable by the user for specific applications; provision for encoder inputs for position control applications; on-board provision for tachogenerator calibration; in-built field-weakening controller, with digital programmable control; drive software includes current-loop self-tuning algorithm; user-defined menu for quick access to most used parameters.

.... i .... i .... i .... i .... i .... i .... i .... i .... i ' "






; ....

: .....

; .....

; ....

: ....



All adjustment and application set up is via a menudriven parameter set controlled from a five-button keypad. This provides a simple intuitive commissioning interface.
Figure 6.8 Current (default response to a speed demand change

loop gains), timebase

- 50 ms/div




.... .....

,. .... :, ....

~ ....

: .........

,. ......... :. ..........

: .... :. ....

,. ........ i ..........

~ ................


Very simple low-power analogue D.C. drives, for use on a single-phase supply, are widely available. They tend to be simply constructed with a very much reduced feature set compared with the larger digital system drives.

2> speed

The 4Q2 D.C. Drive

The 4Q2 is a 7.5 kW four-quadrant regenerative thyristor D.C. controller for operation on a single-phase mains supply. The D.C. motor speed controller is designed to provide full four-quadrant control of conventional shunt-wound and permanent-magnet motors, using either armature voltage or tachogenerator feedback.


U 1 :ref~2{/50ms



Figure 6.9 Current response to a speed demand change (optimised loop gains), timebase - 50 ms/div


gains. Speed is seen to transiently fall by 10% before the speed loop returns it to the original set value.

Mentor D.C. drives can be configured for use in a wide variety of applications, some examples of which are"

mechanical handling pick and place handling systems, complex transfer stations and automatic warehousing plastic extrusion drives for single-screw extruders, involving the forcing of molten plastic through a die to form a continuous product of film, sheet or pipe paper-making machines steel processing - many applications including rolling mills, metal forming and wire drawing
_I I I I I I I I I I I I I I I I

switch-selectable features (easy to configure) speed/torque control tacho/armature voltage speed feedback modes status/fault or low-speed/zero reference full-wave bridge, full control 20:1 constant torque speed range 2per cent regulation for 100 per cent load change with armature voltage feedback (0.5 per cent with tacho feedback) overload protection 150 per cent FLC for 15 s with trip indication ambient - 10C to + 4 0 C

Cheetah- P u m a - Lynx
The Cheetah, Puma, Lynx range of D.C. motor controllers is designed for the efficient speed control of both wound-field and permanent-magnet D.C. motors from 0.18 to 7.5kW. All are single-quadrant controllers.



. . . . . . .x. .~.~ ..~-...v,r~-. . . .



.~ .~.__I i . . :':= .= .v:~'.'.~r~.q

armature current


-1 refA2V50ms

I I I I I I Ii I I I I I I I I I I I I I I I I I I I


-2 ref B lV 50 ms J

Figure 6.10 Speed change due to the application of 100 per cent step load change (using default speed loop gains), timebase - 200 msldiv


D.C. DRIVES:Low-Power Analogue D.C. Drives

Figure 6.11 4Q2 D.C. drive controller

Figure 6.12 Lynx D.C. drive controller

speed/torque control operational modes tacho/armature voltage speed feedback modes

selectable status/fault or low-speed/zero reference relay operation full-wave bridge, fully controlled 20:1 constant speed range

Chapter 6.3



FEATURES C O M M O N TO ALL A.C. DRIVES Power Terminal Layout

Typical power connections to an A.C. drive are shown in Figure 6.13. Typically, the power terminal layout allows the following flexibility: ability to power up the drive from an A.C. or a D.C. source paralleling of the D.C. links of multiple drives, where required in specific applications most Control Techniques drives have a built-in braking transistor circuit, to which an external braking resistor can be connected for dynamic braking

Control Terminal Layout

Typical control connections to an inverter are shown in Figure 6.14. Typically, the control terminals to an inverter provide the following functionality: a number of programmable digital inputs and outputs a number of programmable analogue inputs and outputs connections for a motor thermistor, to protect the motor against overheating one or more programmable relays connections for serial communications with the drive

I + IDB"I braking resistor thermal protection device


" I '2 I '3 I P-I u I v I w I~1

optional RFI I filter I I I I optional line reactor fl~tses

motor earth


start/ reset optional iil

[] [] []
' L1


L2 L3 - - mains supply supply earth

Figure 6.13 Typical power connections to an A.C. drive

10 kD (2 k.Q min)

0 V common local voltage speed reference input (A1) +10 V reference output 0 V common

0-10 V 4 - 20 mA +24 V 0 v_i-~--

remote current speed reference input (A2) analogue output (motor speed) +24 V output digital output (zero speed) drive enable reset run forward run reverse local (A1) remote (A2) speed reference jog select +24 V output 15 --ok 16

module socket

I:'" "1
RJ45 ! ' connector for t

serial c o m m s m
| i

. i




Figure 6.14 Typical control connections to an inverter


A.C. DRIVES: F e a t u r e s C o m m o n

t o All A.C. Drives

key to symbols

single powercable three-core powercable or three single power cables ground cable connection to cable armour or shield

;P:~ii:al r:::;nra' i ...........
" drive



i N

-- ------

I"..... ~ ..... I
- - , I ) : _-.'.......... .


| ................... ' ~ - i

maximumlength:50 mm(2in)
m m m m m m

alternative safety ground connection armoured or shielded cable (3-phase + ground) no sensitive circuits permitted in this zone ~ ///// ::i:LI: ~ L3~I .J--:i: VI: +

~ t l - '- , ~
I /

,~,L _ L L ~ t

~-~ a" u = w "" " -

w --'_LLL

control cables to the drives output 3 output 2 output 1 0V ground

host controller

LI" L2" L3" E" LOAD
RFI filter ........ : LINE L1 L2 L3 E








power-ground bus-bar

:: i: :: i I1: : : i: : :i ii~

i m-,,m IN




l i b ....... ~ , v ,


(,~ v

l l m l l U l l l l l l I l m l l i n

General features

1 single-power ground busbar or low-impedance ground terminal 2 incoming supply ground connected to power ground busbar 3 connect grounds of any other circuits to power ground busbar 4 site ground if required 5 metal back plate, safety bonded to power ground busbar 6 system isolator, circuit contactors and fuses/MCB 7 alternative position for drive fuses/MCB 8 optional braking resistor mounted externally, protected and shielded by a metal grille 9 thermal overload device to protect braking resistor 10 alternative safety ground for motor 11 motor frame ground connection, if required
Special features for E M C

12 the A.C. supply cable must be shielded (screened) or steel wire armoured-the shield or armour should then be bonded to the enclosure wall using standard cable gland fixings 13 back plate should be electrically bonded to the enclosure wall using a short low-inductive connection. Two flat braided cables of nominal size 12 2.3 mm are suitable, or a single braided cable of equivalent dimensions 14 drive heatsink should be directly grounded to the back plate using the metal mounting brackets. It should be ensured that the screws make direct electrical connection to the back plate by using screw threads tapped in the back plate 15 RFI filter should be mounted 150 mm (6 in) from the drive. The RFI filter casing is directly grounded to the back plate by the fixing screws. The length of cables between the drive and RFI filter should be minimised 16 a shielded (screened) or steel wire armoured cable must be used to connect the drive to motor. The shield must be bonded to the back plate using a noninsulated metal cable clamp. The clamp must be positioned no further than 150 mm (6 in) from the drive 17 the shield of the motor cable should be connected to the ground terminal of the motor frame using a link that is as short as possible and not exceeding 50 mm (2 in) in length. A full 360 termination of the shield to the motor terminal housing (usually metal) is beneficial 18 the a.c. supply and ground cables should be at least 100 mm (4 in) from the drive and motor cable 19 sensitive signal circuits in a zone extending 0.3 m (12 in) all around the drive should be avoided 20 unshielded wiring to optional braking resistor(s) may be used, provided the resistor is either in the same enclosure as the drive or the wiring does not run external to the enclosure. A minimum spacing of 0.3 m (12 in) from signal wiring and the supply-side wiring of the RFI filters should be ensured 21 if the control circuit 0 V is to be grounded, this should be done at the host controller (e.g. PLC) and not at the drive to avoid injecting noise currents into the 0 V circuit

Figure 6.15 Precautions f o r p r e v e n t i n g EMC p r o b l e m s

Chapter 6.3


Wiring Precautions to Prevent Electromagnetic Compatibility (EMC) Issues

Because modem inverter drives run at significant switching frequencies, precautions need to be taken to prevent interference between the drives and other equipment connected to the same power supply. Figure 6.15 shows the precautions that should be taken to prevent electromagnetic compatibility problems.


EN50082-2 and EN61800-3 for immunity EN50081-2 and EN61800-3 for emissions using the optional EMC filter where required

Unidrive VTC is an open-loop inverter designed specifically for the fan/pump applications market, with a tailored quadratic (square-law) volts to frequency curve. Its general appearance and mode of operation are very similar to those of the Commander GP. Unidrive VTC covers the range 0.75 up to l l 0 k W (1 to 125 HP). At higher powers up to 1 MW, standard Unidrive can be applied.

An open-loop inverter is characterised by the lack of any form of actual/measured velocity feedback. Without feedback, precise speed control on an induction motor is difficult due to the inherent slip of the motor. Most commercially available open-loop inverters use a common design strategy in respect of the power conversion circuit, but there are many variants of control strategy. Control Techniques' open-loop drives use an open-loop vector control strategy, which maintains almost constant flux in the motor by dynamically adjusting the motor voltage according to the load on the motor. Control Techniques' open-loop inverter range includes the Commander SE, and the Unidrive VTC. These products provide a cohesive range from 0.25 kW, 220 V up to 1 MW, 480 V.

Features and Options

The Commander SE and the Unidrive VTC incorporate the following features:

Specifications and Ratings

Commander SE has been designed as an easy to use (indeed, SE stands for simple and easy), ultra-reliable, rugged openloop A.C. inverter. It covers the range 0.25 to 15 kW (0.35 to 20 HP). The design philosophy incorporates a three-level software parameter set. The first level contains only ten parameters which allow quick drive-to-motor set up and access to the next level. In simple applications this may be all that is required. Level two contains a further 37 parameters which allow the user to configure the drive for the majority of open-loop applications. Level three can only be accessed by using the drive's serial communications port, and contains a multitude of parameters which allow the drive to be configured for almost any conceivable open loop application.

fully compliant with EU EMC regulations optional external EMC filter plug-in control signal connectors full digital control programmable preset speeds automatic resynchronisation to a spinning motor coast to stop, deceleration ramp and D.C. injection braking stopping modes dynamic braking control terminal, keypad or serial communications control voltage or current speed reference signals selectable PWM switching frequencies option module for easy upload/download of parameters WindowsTM-based commissioning software package

The drives also contain a more advanced set of features which can be used by the system designer or the more experienced user. These include: speed control adjustable precision speed reference adjustable skip frequencies with three adjustable skip bands adjustable preset speeds ramps preset acceleration ramps preset deceleration ramps separate acceleration and deceleration ramps for each preset speed separate acceleration and deceleration ramps for jogging adjustable S ramp torque control stopping adjustable D.C. injection braking current level and time programming I/O fully programmable analogue and digital I/O for alternative functions motor protection current limiting (short-term overload) motor thermistor protection (long-term overload) protection trips with trip log monitoring

frequency accuracy: resolution: starts per hour: 0.01% 0.1 Hz by using the electronic control terminals, unlimited; by switching of the supply, 20 starts per hour maximum (three minute intervals between starts) one second maximum (allow at least one second before monitoring the state of the status relay contacts etc.) 3, 6, and 12 kHz are available

power-up delay:

switching frequencies:


A.C. DRIVES:Open-Loop Inverters

Figure 6.16 Commander SE family

Figure 6.17 Unidrive VTC family

programmable drive status logic status and diagnostic information k w h meter run time log adjustable speed sensing levels running costs auxiliary functions

auto reset PID controller undedicated programmable logic undedicated programmable threshold comparitor motorised potentiometer second motor selection with second motor parameters


Chapter 6.3
Access to the advanced parameters is via easy access serial communications: The Commander SE has an easy access serial communications port which enables one or more drives to be used in systems controlled by a host unit such as a PLC (programmable logic controller) or computer. The communications link for the drive uses the EIA RS-485 standard for the hardware interface. The drive has a standard two-wire RS-485 half-duplex interface that enables all drive set up, operation and monitoring to be accomplished if required. Therefore, it is possible to control the drive entirely by the RS-485 interface without the need for other control cabling. The host controller can operate up to thirty-two EIA RS485 devices with the use of one line buffer. Further buffers will increase this number if necessary. Each transmitter/ receiver within a drive (with the internal termination and external pull-up and pull-down resistors disconnected) loads the RS-485 lines by one unit load. This means that up to 15 drives can be connected in a single group to one line buffer. However, with the serial addresses available, it is convenient to only have up to nine drives in a single group. When additional line buffers are used, up to 81 drives can be operated by the host controller. In this case the drives are organised in a maximum of nine groups of nine drives each. A particular drive or group of drives can be given commands without affecting other drives or groups of drives, respectively. The serial communications port of the drive is situated at the RJ45 connector. The EIA RS-485 two-wire port is isolated from the power stage but not isolated from the other control terminals. The EIA RS-422 hardware interface is also supported. As can be seen, the flexibility contained within a modern commercial drive is considerable. In addition a comprehensive range of options is available to provide more flexibility to application or ease of use.



Serial communications converter- the EIA RS-232

hardware interface cannot be used with a two-wire EIA RS-485 interface. Therefore, a suitable adapter for connection to a computer RS-232 interface port is essential. A purpose-designed 485 to 232 converter is available as an option that simply has a D-type connector on one end for connection to the host PC and an RJ45 plug on the other for quick connection to the drive. The converter is built into the lead for simplicity. SE S o f t - SE Soft is a software package designed to aid set up and commissioning of Commander SE drives. It connects to the drive via the drive's twowire RS-485 link. Bipolar reference option - this is a bipolar analogue input card, which offers the user the possibility of inputting a + 10 to - 10 V speed reference signal into the drive. The option is mechanically mounted under the drive's terminal cover and is then directly connected to the drive's control terminals. Profibus DP option c a r d - allows communication with the Profibus DP high-speed fieldbus system. DeviceNet option card- allows communication with the DeviceNet high-speed fieldbus system. CANopen option c a r d - allows communication with the CANopen high-speed fieldbus system. Interbus S option card- allows communication with the Interbus S high-speed fieldbus system. Cable screening- a set of cable screening brackets and screening clamps is available for the drive to provide a convenient way of connecting supply, motor and control cable screens to ground. A C input reactors - a specifically designed set of input line reactors is available allowing the drive to operate on problematic industrial supplies containing large amounts of disturbance, such as voltage spikes, notching etc. Braking resistors - a purpose designed set of externally mounted braking resistors is available with a



(vi) (vii) (viii) (ix) (x)



(i) QUICKEY- this is a small, plug-in, key-like device that allows the user to extract a parameter set from one SE drive and implant the same parameters very quickly and efficiently into one or more SE drives. The key plugs into a small header located under the drives terminal cover. (ii) Electromagnetic compatibility (EMC) f i l t e r - there are three EMC filter options available for Commander SE, these are: a footprint filter that mounts under the drive and will only increase the depth dimension of the drive and not require additional panel space in an enclosure; this filter can be mounted by the side of the drive if required a filter as above but with low earth leakage a low-cost panel-mounted filter suitable for limited inverter-motor cable lengths

Figure 6.18 Commander SE together with a variety of options


A.C. DRIVES: O p e n - L o o p


tailor-made mounting kit, making optimum use of the drive's overall space envelope. Unidrive VTC has a similar set of options available, where required, tailored to its specific application requirements.

motor rating plate: rated motor current in amps rated motor speed in min- 1 rated motor voltage in volts motor power factor The drive has the ability to perform a current-loop autotune, which is parameter selectable. There are two levels of autotune: 1 The first level carries out tests without spinning the motor. Stator resistance and voltage offset measurements of the motor are taken. The second level carries out the static tests and a test where the motor is rotated. In addition to the above test, the total leakage inductance and the rated magnetising current are measured.

Methods of Control
The drive can be controlled by any of the following methods: terminal mode, applying signals from electrical contacts, a controller, or PLC, to digital inputs on the drive keypad mode; manual operation of the keypad on the front panel of the drive; apart from an overriding safety switch to be connected to the control terminals, no external signal connections are made to the drive serial communications using a system controller or PLC connected to the drive by a two-wire RS-485 serial communications link (can be used in conjunction with terminal or keypad mode)

Once these parameters have been set up the performance of the Commander SE, as can be seen in Figure 6.19, is extremely good. (i)
Low-speed t o r q u e - Figures 6.19 and 6.20 show a typical torque/speed curve for a Commander SE drive when used with a standard off-the-shelf induction machine. The drive motor used was rated 2.2kW (3 HP). As can be seen, 150 per cent motor-rated torque was available down at just less than 2 Hz operation. S t a r t i n g t o r q u e - starting torque is defined as the amount of torque that can be applied to the motor shaft under test, with the drive still able to perform a motor start. The starting torque of Commander SE with a standard induction motor equates to approximately 170 per cent of the rated motor torque.

Performance of the Commander SE Open-loop Drive

The Commander SE drive uses an open-loop vector control strategy, which maintains almost constant flux in the motor by dynamically adjusting the motor voltage according to the load on the motor. This strategy does require the user to match the drive to the particular motor or type of motor being used if high performance is required. The parameters that need setting are all available from the (ii)

25 150% torque 20

~10 5 0
0 1 2 3 4 5 6 speed, Hz 7 8 9 10

Figure 6.19 Commander SE low-speed torque performance (150 % full load, 2.2 kW, 3 p.h.)

16 14 12

100% load

~ 6
~ 4

2 0

speed, Hz


Figure 6.20 Commander SE low-speed torque performance (100 % full load, 2.2 kW, 3 p.h.)

C h a p t e r 6.3





the speed range is defined as the controllable speed range over which 100 per cent motor-rated torque can be achieved. The speed range of Commander SE is approximately 1 to 55. In tests, a 50 Hz machine achieved speed control with 100 per cent rated torque from 0.9 to 50 Hz. It should be noted that the above definition of controllable speed range is not universally accepted within the industry so care needs to be taken when making comparisons. S p e e d a c c u r a c y - speed accuracy is defined as the percentage speed error between the displayed speed and actual measured speed when controlling a machine at base speed and 100 per cent load. Speed accuracy of the Commander SE is + 0.6 per cent. S p e e d r e s p o n s e - the speed response is defined as the ability of the drive to follow a small signal sine wave applied as a speed reference. The speed response of Commander SE is 10 Hz, as shown in Figure 6.21. The tests were performed on a 3000min -1 4 k W machine coupled to a two to one inertia. Above 10Hz the speed feedback waveform reduced in amplitude below the - 3 dB point. (Speed feedback was provided on the test rig by a 60V/1000min --1 tachogenerator.)
Full torque speed range -

THE UNIVERSAL A.C. DRIVE The Concept of a Universal Drive

As A.C. drives have developed over the past decade, there have been parallel advances in power-stage design and in the control methodologies used. Essentially, the power stage (including all the power devices, current and voltage feedback systems etc.) remains very similar in all A.C. drives - whether used for the open-loop vector control, closed-loop vector control or for the control of brushless A.C. servomotors. The advances in A.C. drive control strategies as described in Chapter 4 have also led to a uniform approach which makes the implementation of a universal drive intuitively logical. Operation of a standard drive as a regenerative converter is not such an obvious extension to this development. However, reference to the control strategy described in Chapter 4 shows that this can be readily incorporated. The Control Techniques Unidrive is therefore able to operate in the following modes: 1 2 3 4 5 Open-loop V/F for the control of parallel induction motors (or other loads). Open-loop vector for the control of single induction motors. Closed-loop vector for the control of single induction motors. Closed-loop servo for the control of brushless PM motors. Regenerative operation to provide a sinusoidal A.C. supply front end to an inverter system and also to allow power flow, both to and from the A.C. supply.

Typical Applications

fans and pumps conveyors cranes and hoists (hoisting and traverse control) simple winders mixers and agitators grinders spinning machinery (textile industry) circular saws cutting and slicing (meat industry) flow control valves spindle control (lathes) simple engine test rigs

tek run" 5ks/s

. . . . ! . . .

. . . . ! . . . . . . . . ,




. . . .

[ . . T .. ... .. . . . !. .. ... .. . . . ~ .. ... .. . . .!. .. ... .. . . .!. ,. . .1., ? . ! . . . . , - ,

RB freq 10.02 Hz low signal amplitude

. .

. .

. . . .

. . .

. . . .

. . . .

J . . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

1 V 50 ms

50 ms chl


2.24 V

Figure 6.21 Low-speed torque capability~frequency characteristic is 1.9 Hz at 150 per cent rated motor torque and 0.9 Hz at 100 per cent rated motor torque

Figure 6.22 The Unidrive family (0.75 kW to 1MW)


A.C. DRIVES:The Universal A.C. Drive

Having a single product to meet the requirements of all A.C. motor control applications offers key advantages to users in respect of ease of use, common interfaces and programming, single product training etc. for operating personnel, simplified/rationalised spares requirements. Considerable flexibility is inherent in a product which meets such a diverse number of requirements. Such flexibility and universality does imply and involve some compromises. These can be summarised as follows: A common product footprint where servo products tend to a slim book form and open-loop drives tend to a shoe box form. Simple general-purpose drives do not require the accuracy of, for example, current measurement needed for high-performance drives. Further, the necessary flexibility of a systems product does result in additional components. These issues have some cost impact which is significant at lower powers. Great efforts have been made to ensure that flexibility and performance does not mean complexity and difficulty in use. However, compared with the ten parameters found in menu 1 of Commander SE, Unidrive is feature rich!

Unidrive Option Modules

The operation of the drive can be easily upgraded and adapted to different applications with the use of a number of option modules. When an option module is plugged into the drive, it is automatically recognised and the relevant parameter sets become visible. The option modules are of two different types.

Figure 6.23 Unidrive small option module


A number of option modules are available in a small format as follows: additional inputs and outputs a second encoder input for use for example in a digital lock application alternative speed/position feedback inputs: resolver and sine/cosine encoder cloning module for selecting/storing/transferring/copying up to eight full drive parameter sets

A number of option modules are available in large format as follows: standard serial communications (RS-232/RS-485) applications module - this provides a very powerful parallel processing capacity, which can be programmed in an IEC 61131-3 language; this can be considered as a PLC within the drive and when supplemented with directly connecting remote I/O modules provides a very powerful tool for system designers fieldbus interfaces: Profibus DP, Interbus-S, DeviceNet, CAN open, Modbus P l u s . . . CT Net module - CT Net is a communications module for Control Techniques' own fieldbus system and allows the drive to function as a master or a slave on the bus system.

Figure 6.24 Unidrive large option module

The Unidrive has its own, Windows T M PC-based, commissioning and set-up software called Unisoft. The use and

C h a p t e r 6.3


application of this type of software package has already been described in a previous section.
MACROS status relay drive healthy analoque frequency/speed reference 1 (remote) 0-10 V

signal connector

The Unidrive has a set of eight built-in user set ups which are targeted at specific applications. These allow a very fast and simple method of setting up the drive for these applications. The available macros are: macro 1 easy mode macro 2 motorised potentiometer (frequency/speed control by up and down contacts) macro 3 four preset speeds (selected by digital control signals) macro 4 torque control macro 5 PID (set point) control macro 6 axis limit control macro 7 brake control macro 8 digital lock Setting up the drive in any particular macro mode not only sets the terminals to the appropriate configuration but also customises the easy access menu 0 to contain the parameters specific to this mode of operation in one simple step. For example, on the Unidrive if the drive is required to be set up in the motorised potentiometer mode (where two, userfitted, push buttons can be used to increase or decrease the speed), enabling macro 2 will automatically reconfigure the terminal set up to be suitable for such a use. This is shown in Figure 6.25. As can be seen, the digital inputs are automatically configured to be appropriate for taking inputs from the user-fitted push buttons.

motor thermistor

0 V common

0 V common
speed torque up reset and MOT. POT. reset down run forward run reverse MOT. POT. enable

analogue RP MOT. POT.

OL> external trip CL> drive enable 0 v common

Figure 6.25 Terminal configuration of drive when macro 2 (motorised potentiometer function) is enabled

Open-loop Operation
The operation and performance of the Unidrive in open-loop mode remains essentially similar to that of the Commander SE described earlier in section 6.3. The parameter set of the Commander SE has also been harmonised with that of the Unidrive to a large extent minimising confusion/conflict within the overall family.

C~sed-loop Operation
Closed-loop control of induction motors is now sufficiently advanced so as to become increasingly used in applications where in the past D.C. drive + motor combinations would have been the natural choice. It is often said that closed-loop induction motor drives match the performance of a D.C. drive. This is not true, the A.C. drive is far superior. With the addition of a feedback device to the motor, the performance of the drive + motor combination significantly increases as compared with the open-loop drive: the drive can generate full torque from standstill up to the base speed of the motor the speed can be very precisely regulated due to the feedback now available the dynamic performance is greatly improved the magnetising and torque-producing components of the current are now more precisely controlled - accurate torque control operation is now possible applications which require accurate positioning, now become possible

The Unidrive in open loop would be preferred over the Commander SE in applications where: there is a requirement to have a common product for different parts (open loop, closed loop and/or servo) of the same application the application is complex enough to require the parallel processing capability available via the UD70 second processor option module additional I/O capability is required high-power ratings are required

The performance of the Unidrive in open-loop mode is similar to that of the Commander SE.


The closed-loop mode of operation has various advanced features.


A.C. DRIVES: The Universal A.C. Drive

(i) Autotuning Once the motor name-plate data has been entered into the drive, the user can enable an autotune function which measures some of the other key parameters of the motor, such as: power factor (the real value rather than the name-plate figure) machine inductance the magnetic saturation characteristic of the machine


The following examples give an indication of the level of performance that can be obtained using Unidrive in closedloop vector mode. If the drive is set up correctly it should be possible to obtain at least 175 per cent rated torque at standstill (assuming that the drive and motor are matched). Figure 6.26 shows the drive operating under speed control, with a step of torque to 100 per cent applied and then removed. The rotor position changes linearly (constant speed) until the torque is applied when the position changes to wind up the integral term to obtain the necessary torque. Higher settings of the integral term reduce the step change of position. Although there is a step in the position (which occurs with any PI-type speed controller) the speed on and off load is the same. In Figure 6.27, the drive is operating in torque control. The speed of rotation is defined by another motor connected to the test motor shaft. The Unidrive is enabled, applies 100 per cent torque and then is disabled. The transient change of position is due to the limited speed holding ability of the drive controlling the other motor. Typical applications include:

In addition, the drive can tune itself accurately to the slip of the motor during normal operation. It can thus adapt itself to changes in slip owing to changes in stator resistance due to heating effects in the motor. This function is also of particular advantage owing to the fact that in most cases the slip stated on the motor name plate is a batch or a design value and not one derived for that particular machine. (ii) Alternative types of speed/position feedback device The Unidrive can operate with a number of different types of feedback device in closed-loop vector mode: standard encoders with A, A\, B, B\ quadrature channels and Z, Z\ marker pulse (optional) outputs resolvers sincos encoders of up to 1024 sine and cosine waves per revolution - the drive can then interpolate to 2048 parts of each sine wave and hence have a total resolution of up to 1024 2048 = 2 097 152 counts per revolution

(iii) Application features The drive has built-in capabilities for application set ups such as:

digital lock to another motor/a master encoder orientation on stop e.g. for tool changes on a machinetool application the motor shaft must stop in a specific orientation torque - control modes specifically suited to winder, unwinder applications torque control with dancer feedback, using the built-in, standalone PID control loop

cranes and hoists lifts winders and unwinders wire drawing extruders plastic production paper-making machines rolling mills metal forming mechanical cam replacement applications applications requiring digital slaving CNC machine spindle drives cable laying from ships

Servo Operation
The basic control algorithm used for the closed-loop vector operation is very similar to that used for the servo mode of operation. The main differences are that the magnetising

More advanced applications can be easily performed using the large applications module.
i 1 1 1 1 1 1 1 , i l l 1 1 1 1

1 1 1 1

i 1 1 1

i , i i

1 1 1 1

1 1 1 1

1 1 1 1 ~

rotor position



r " r

CH440Nn CHPMTB500s
, , , , i , , , , | , , | , i , i | , I , , , , , i | , , , , , , , , , , , , ,"

C H 2 14 A

Figure 6.26 Application and removal of 100 per cent load at 5 m i n - 1 (speed-controlled mode)

Chapter 6.3
i n , | t n i n u , , , , , , | , , | , i u n n , , | | n ] | u n n , , , ,


rotor p o s i t i o n Itorq



torque \

/ :

, /.,



: ~ , w , , ~ - ~

,~, ~

' ~ . , ~ . ~



..., J

........................ -. .
* .


5 . ~


- C H 2 14 A - C H 4 ~.0 N rl ~C H P M T B 5 30 s
!'. | | | , | , . | , , , , , , | , | n n | n | , | , , , n , | , | , , , n , , , ~

Figure 6.27 Application and removal of 100 per cent load at 5 min- 1 (torque-controlled mode)

current of a servomotor is controlled at a value of 0 and it, of course, has no slip. However, this should not affect the bandwidth of operation. If the drive is set up correctly for closed-loop operation i.e. it has the correct value of slip, power factor and all the other motor parameters, then there should be no essential difference in performance from the drive point of view. The key differences between closed-loop vector and servo modes of operation are:
Motor shaft performance - the most significant difference occurs in this area. Servomotors are often specifically designed with low rotor inertia to facilitate very good dynamic performance. P o w e r d e n s i t y - for a given power rating, the physical size of an induction motor would be larger than that of a servomotor. H i g h e r p o w e r r a t i n g s - for applications that need more than relatively low power ratings (a few kWs), servomotors are not commonly available and tend to be expensive. C o s t - for a given power rating, a servomotor would tend to be more expensive than an induction machine. Note, however, for closed-loop operation an induction motor would need to have a feedback device fitted and this must be considered in any cost comparison. M o t o r d e g r e e o f p r o t e c t i o n ( I P r a t i n g ) - servomotors are generally available as standard to IP65. Induction motors tend to be available as standard to IP54. O v e r l o a d r a t i n g s - servomotors are generally designed for very short overloads. Induction motors can, in general, handle overloads above their nominal currents, for longer time periods. Operation above base speedan induction motor can be operated with ease in the field-weakening region i.e. up to several times above base speed. In theory, PM servomotors can also be operated in the fieldweakening range, however it is difficult to protect against high overvoltages should control be lost above base speed and consequently such operation is rarely considered.

rotor when the drive powers up. Suitable devices are: encoders with A, A\, B, B\ quadrature channels and Z, Z\ marker pulse (optional) outputs. In addition, U, U\, V, V\ and W, W\ commutation signals are also required to indicate the segment position of the rotor magnets at power up resolvers sin-cos encoders of up to 1024 sine and cosine waves per revolution - the drive can then interpolate to 2048 parts of each sine wave and hence have a total resolution of up to 1024 2048 = 2 097 152 counts per revolution


Application set ups similar to those for the closed-loop vector mode are also possible in the servo mode of operation. For example: digital lock to another motor/a master encoder orientate on stop e.g. for tool changes on a machine tool application torque control modes specifically suited to winder, unwinder applications torque control with dancer feedback, using the built-in, standalone PID loop


The performance of a servo controller is commonly described in terms of bandwidth. The bandwidth of a servo speed controller is important because it defines the dynamic performance of the controller. If it is assumed that a 10:1 bandwidth ratio is required in a position controller with inner speed controller, then the speed-loop bandwidth limits the maximum bandwidth of the position controller. A very good position controller has a gain of 50 and a position loop bandwidth of 50rads-1 (8 Hz). Therefore the speed controller needs to be able to achieve a bandwidth of 500 rad s- 1 (80 Hz). The speed-loop bandwidth can be measured by using a sine wave excitation signal as the input to the servo controller

The servo mode of operation requires feedback devices which can provide indication of the absolute position of the


A.C. DRIVES: The Universal A.C. Drive

and monitoring the actual response to this demand. If this test is done using a benchmarked set of gains in the servo controller, standardised Bode gain and phase plots for the servo controller can then be plotted. As described in Chapter 4.1 and 4.4, the phase plot is more critical and represents the impact of the delay created by the speed controller which has the effect of: limiting the possible bandwidth of an outer controller, or producing a delay between the required speed at any point and the actual speed at any point; a system with a position controller with speed feed forward will follow the profile transiently due to the speed feedforward term and any lag in the speed controller results in transient position errors (i.e. following error) The speed-loop Bode plots for Unidrive, Figures 6.28 and 6.29, show that the bandwidth, i.e. the point at which the phase lag is 60 , is approximately 100 Hz.


robotics dynamic pick and place applications axes drives in all types of CNC machine woodworking machines embroidery machines cut to length lines

Regeneration Mode

The input stage of a nonregenerative A.C. drive is usually an uncontrolled diode rectifier, therefore power cannot be fed back into the A.C. mains supply. In the case of a Unidrive operating in regenerative mode, the IGBT bridge can be used as a sinusoidal rectifier, which converts the A.C. supply to a

5o o -5o -lOO



-150 -200



~- -250 Q. -300 -350 -400 frequency, Hz

Figure 6.28 Bode phase plot



rn -o



ooo oo I--I-T
- 15.00000

frequency, Hz

Figure 6.29 Bode gain plot

- 0 - 20% overshoot




controlled D.C. voltage. This D.C. voltage can then supply one or more Unidrives, which control the motor/s. An explanation of the theory behind this mode of operation is given in Chapter 4.2. Some additional, external components are required to build up a Unidrive regeneration system. These are: main regeneration inductors start-up circuit (for a controlled charge up of the D.C. bus of the regenerative system) switching frequency filter

base speed in one direction to base speed in the other direction without speed ramps. Results show that the control system is fast enough to limit the change in D.C. link voltage to approximately 25 V (0.03 p.u.) with very rapid changes (less than 5 ms) in power flow of over 60 kW. Should load changes exceed the tracking capability of the controller, or the D.C. link power exceed the maximum A.C. power, then the PWM rectifier will be forced into current limit. Figure 6.33 shows transient overload operation where the PWM rectifier goes into current limit but the system remains stable and line synchronisation is maintained. During the overload the D.C. link voltage deviates by approximately 80 V from the set point and does not recover until the D.C. terminal power decreases. If the D.C. link power exceeds the PWM rectifier capability and these conditions are sustained then depending on the direction of power flow the D.C. link voltage will either rise until an overvoltage trip occurs or the voltage collapses to the point where the antiparallel diodes in the PWM converter act as a simple uncontrolled diode rectifier. Loss of one supply phase or all three is detected by monitoring the D.C. link voltage, input terminal voltage and x-axis current. In the event of phase failure the converter is inhibited before allowing the D.C. bus voltage to stabilise and attempting resynchronisation. Synchronisation and running has been proved to be satisfactory even with the high levels of supply distortion and notching found in some industrial applications as demonstrated in Figure 6.34. The supply notches were produced by a D.C. motor drive connected directly to the PWM rectifier supply without line chokes or filter capacitors. Note that although the supply voltage notching is almost 100 per cent, and the rectifier line current slightly distorted, line synchronisation is maintained.

A regenerative system using the Unidrive, is typically connected as shown in Figure 6.30. The main advantages for an A.C. regenerative system are: energy saving the input current waveform is a sinusoid the input current has a near unity power factor the output voltage for the motor can be higher than the available A.C. mains voltage the regenerative unit will synchronise to any frequency between 30 and 100Hz, provided that the supply voltage is between 380 V - lOper cent and 480 V + 10 per cent under conditions of A.C. mains instability, a Unidrive regenerative system can continue to function down to approximately 270 V A.C. supply voltage without any effect on the D.C. bus voltage and hence on the operation of the motor drives the regenerative and motor drives are identical

Many sine wave regenerative drive systems have been supplied with ratings from a few kilowatts to several hundred kilowatts. System complexity varies from single motor drives to systems with many motor drives connected to a common D.C. link and PWM rectifier. Waveforms in Figure 6.32 show the system response to a transient when using a 37 kW drive as a PWM rectifier to supply another 37kW drive supplying an induction motor operated under vector control. The motor was reversed from

The performance of high-end servodrives is restricted by the physical limits imposed by high-resolution position acquisition. These limits are currently set by the use of analogue (sin-cos) position feedback signals and the resulting feedback signal degradation resulting from typical system applications where varying cable lengths between

power flow during regenerative operation

R filter Y -t - ~ | ana I supply


switching ~ s t a r t - u frequency t----~'-t---" . . . . I filter I I circuitl ....

unidrive operating unidrive in operating D.C. bus open-loop in vector regen or servo


contactor control signals drive enable control signal

power flow during motoring operation Figure 6.30 A regenerative system configuration


A.C. DRIVES: High-Performance Servodrives

the controller and the position feedback device are a practical reality. Significant advances for overcoming this limitation have been achieved by the integration of position acquisition

and position/velocity control, local to the position feedback device i.e. within the motor housing. Control Techniques has utilised this approach and combined the latest technology in high-resolution position acquisition (sin-cos encoders) and DSP technology to achieve a very substantial,

Figure 6.31 450 kW sine wave regenerative drive system for engine test rig



- ..........................

V. . . . . . . . . . . . . . . . . . . . . . . . . .



i i i |

Figure 6.32 Input converter currents during high-speed motor reversal; D.C. link voltage VD.C. 180 V/div, input currents Ix 85 Aldiv, ly 85 A/div, time 200 ms/div

Chapter 6.3


.................., ii
q ~ vT','v Ivy

i 1 ..........


700 V


' V

lV V

r .|

Ivl '

Figure 6.33 Input converter currents and D.C. link voltage during transient overload; phase current l OOA/div, D.C. link voltage VD.C. 100 V/div, input current Ix 85 A/div, time 20 ms/div



Figure 6.34 Effect of supply voltage disturbances on input converter line current

application-invariant increase in position feedback resolution and the capability for active torque compensation. This new concept in servo system design eliminates the need to transport noise-sensitive analogue feedback signals by providing a dedicated high-speed two-wire data link, thereby drastically reducing the number of signals. It should be noted that digital encoder signals are subject to edge degradation, and noise pick up when transmitted via cable in an industrial environment. Indeed, encoder cable hygiene, made the more problematic by the existence of typically 10 to 14 wires per encoder, is the primary cause of site problems with conventional servodrives. The reduction of the wires per encoder to four (data link plus power supply) is in itself a very significant practical benefit of the system. The simplified structure reduces the maintenance and commissioning cost drastically. In other words, it is a higher

performance servo which is easier to use and carries no cost penalty. The combination of high-resolution feedback techniques and new drive topology provides a quantum improvement in servo performance. Typical position/speed feedback devices used in today's servodrives tend to use either resolver or digital incremental encoder technology. The former of these has limited accuracy and linearity due to distortion introduced by the magnetic elements used in resolvers. The signal-processing mechanisms which typically comprise a resolver-to-digital converter also set a limit on resolution and dynamic accuracy leading to an overall system performance which, although satisfactory for many applications, falls short of that required in many others. On the other hand, there is a practical limit on the number of lines per turn for incrementa encoders, a limit which does not allow the


A.C. DRIVES: High-Performance Servodrives

Figure 6.35 M'Ax Servodrive family

a M'Ax, standalone servodrive and Unimotor SL b MultiAx, three drives in one package for direct SLM interface to motion controllers

resolutions required by today's high-end servos to be achieved. Interpolation techniques can be used to improve the effective resolution, but under dynamic operating conditions interpolation is another source of distortion. The sin-cos encoder provides high-resolution analogue position information which has a greatly reduced level of linearity distortion. The availability of small outline signalprocessing components provides the opportunity to process the information at source, i.e. directly behind the motormounted encoder. This eliminates the opportunity for noise injection. This philosophy of the speed loop motor or SLM TechnologyT M of servo control has been incorporated into Control Techniques' M'Ax drives which include: M'Ax - a standalone single-axis drive which offers the high performance and ease of use and commissioning offered by the SLM Technology in a very compact package. MultiAx - a drive which has three power stages built in, which in conjunction with an SLM-equipped controller forms a very compact and cost competitive three axes servo system.

internal, basic, programmable logic controller 24 V auxiliary back-up supply

Servodrives fall into two main categories of application, each with somewhat different requirements. First, point-topoint positioning where what is important is the speed of getting from point A to point B. In such cases it is the accuracy of positioning and the dynamic performance that is key. In the second category where the precision of the trajectory of the motion is important, there is an additional requirement for smooth and controlled motion between points. For example, the quality of typical machining, robotics and high-performance process applications depends largely on the smooth running of the motor and on a stable, dynamic response during system disturbances. In the following subsections we put some typical figures on these subjective statements.

The M'Ax drive is described in detail in the following sections. The following features are key in considering the M'Ax product:

200 per cent overload rating internal braking resistor high-precision synchronisation of axis position within 50ns high-precision synchronisation of speed loop within 50ns eight million counts per revolution position resolution high-speed drive status line responding within 1 las automatic recognition of the motor (motor data is held in the encoder EEPROM and read by the M'Ax at power up)

The positioning accuracy, and as important the repeatability of positioning, is affected by a number of issues, but primarily the mechanics of the machine and the position feedback device. Figure 6.36 shows a M'Ax servodrive performing speed reversals between - 6 0 0 0 and + 6000 min-1. During the period of steady speed the error between demanded position and actual was measured. To give a practical feel to the data, it is convenient to consider such measures when reflected into a linear movement with typically one revolution being equivalent to 10 mm of travel. On that basis, the measured position error at 6000 min- 1 was < 1 ~tm. During the deceleration/acceleration the maximum position error was <2 mm which is impressive when the time base of the oscillogram reveals that the 12 000min-1 speed change is achieved in 120ms. It is even more impressive when it is revealed that the load inertia was seventy-eight times the motor inertia.

Chapter 6.3



MSPEE0 .... " .............


.,ONE j
.......................................... ~i~i~:~:ii~:~;~i;i~s~:iis~iii)~i~:ii;i~i~;F#;ii~:~i~F;::iis::: ii~s: i;;;~ili::sii:;ii::si@:;i::::iii::s !~ ~s;;: @:~i::il!::iii::ii

- +.~r~..---f: .......................................... " Nsi





......... i



:~. . . . . .


i i

Figure 6.36 Rapid speed reversal (+6000 min- 1 in 120 ms) with M'Ax; second trace is following error

The following error in a control system is in effect the lag in the system, the ability of a drive system to follow a demand signal without delay/lag. The key to minimising following error is to have the ability to operate with the highest possible gains in the control loops while retaining stability. The SLM system provides the opportunity to operate with gains up to five times higher than conventional technology, primarily as a result of the high-quality position information upon which its control is based. The above data is for position accuracy during the rapid deceleration/acceleration (~ 175 rad s-2).

figure shows that at a feed rate of 1 m min- 1 the maximum deviation from the ideal is within 8.1 pm. Tests at feed rates of 3 m min- 1 and 5 m min- 1 gave a maximum deviation from the ideal within 9.3 lam and 10.4 pm, respectively. It is also interesting to note that for the out-of-box performance, i.e. when the user has not made any adjustments to the drive parameters, the maximum deviation from the ideal was within 11.6 pm at 3 m m i n - 1.

What is important to the user is the ability of a drive to withstand and recover from an external disturbance, or change in demand - how big the deviation from the ideal trajectory is, how quickly it recovers back onto the trajectory and the area described by these actions. This is often referred to as the stiffness of the system. Unfortunately drives engineers, or more specifically control engineers, complicate life by talking about the bandwidth of the drive. The bandwidth is the ability of the drive to respond in a controlled/stable manner to a small signal reference or demand signal. Often the torque bandwidth is discussed and often engineers can end up 'comparing apples with oranges'. As discussed in Chapter 4, what is critical in a digital system is the closed-loop phase delay of the torque/current controller rather than the closed-loop gain ( - 3 dB point) which may be used in classical stability theory. There can be a factor of more than 2:1 between these measures because of the digital nature of the control loops (note that even an analogue drive with a digital PWM generator should be measured in this way). A servodrive with a torque bandwidth of 1000 a z is more than adequate for most applications. It is interesting to note that the resonant frequency for

Here we are talking about the speed variation during a revolution of the motor. In a positioning servodrive system this is not critical but in an application requiring contouring or smooth motion a variation of <<0.1% is required. In machine tool applications, the smoothness of a drive system is measured by performing a ballbar test. This type of test is undertaken by feeding two axis drives with a sine wave and cosine wave position reference, respectively. When the two drives are controlling the linear quadrature axes of an X - Y table then the resulting motion should be a perfect circle. The ballbar test measures the deviation from the ideal. The deviation is not purely attributable to drive performance as mechanical and motion controller imperfections impact also, however it is a good basis for comparison. Figure 6.37 shows the results of a ballbar test carried out on a machining centre equipped with M'Ax servodrives. The


A.C. DRIVES: H i g h - P e r f o r m a n c e


high-quality encoder mounting systems is of the order of 1.5... 2 kHz and this needs to be well away from the torque loop bandwidth of the drive if interesting performance is to be avoided. In terms of the speed loop bandwidth this is a factor of ten below the torque loop and the position loop bandwidth is typically a factor of four below that of the speed loop.

b c d e f g h

Summary of Practical Advantages of SLM Technology

Ease of set-up - M'Ax offers easy start technology where the motor parameters and encoder accuracy error

compensation data are stored in the motor. Once the motor is connected to the drive the drive default set up is automatically updated, and for up to 80 per cent of applications no further tuning is necessary. Minimum connections/cabling. SLM performance - positioning accuracy, smooth rotation, dynamic performance. Matched motor range. SLM flexibility - ability to down load parameter changes or complete control-loop structure changes. Lower installation costs. Higher system reliability. Direct connection to industrial 380-480 V supply.

Figure 6.37 Ballbar test results for machining centre equipped with M'Ax servodrives

21 -Jun-00 15:52:55 ~?M1 . . . . . . 20 ms 2.00 V

HARDCOPY i-output to m

|GPIB |RS232 I Centronics Page feed--


220 ms 2.00 V

2r protocol "|HP 7470 |HP 7550 ITIFF

-"1 / | /

....... p ...... .,. ...... ,, . . . . . . . , ....... , ....... , ....... , . ~ .

without torque compensation 20 ms BWL 1 .1 V D.C. X0 2 .2 V D.C. ,oX _.j-T 100 kS/s 1 HFREJ 1.34 V n STOPPED

Figure 6.38 Position deviation of a M'Ax drive system subject to 100 per cent load impact; effect with and without torque compensation is shown, vertical scale is 3.2 per division position error

Chapter 6.4


an. ref + enable + gnd.~ 4 wires




power 1

H I mtr 1 I



4 wires Drivelink
4 wires

Drivelink 1 power 2 multiAx



lencoder I SLM I

power supply 4 wires


SLM module

H -~1motor I 2

Drivelink 2 power 3 Drivelink 3



SLM [-'[

Figure 6.39 Single-axis (analogue reference); total cables: 8 power + 8 control (c. f. traditional 8 power + 30 control)

~lmtr3 I
*1 SLM I
I encder ]

i j k l m n

Up to 16-bit analogue reference with very good linearity and low zero position offset and deadband (< 150 ItV). Touch trigger response within 50 Its. All-pluggable terminals. Control terminals standard D type. Comprehensive cable management system for power and motor cables ensuring good earth bonding. Excellent EMC characteristics, e.g. immunity tested to: + + + + 4 kV 8 kV 5 kV 10 V ESD contact ESD discharge fast transient burst (5 ns rise time) 80 per cent AM RF immunity.

4 wire

4 wire

Figure 6.40 Three-axis (digital SLM reference); total cables: 16 power + 24 control (c.f. traditional 24 power + 90 control)

axes drives in all types of CNC machine woodworking machines embroidery machines cut-to-length lines

o p

Very compact size in book format. Matched to motion controller, CNC and motor products.

robotics dynamic pick and place applications



In terms of energy efficiency, standard A.C. induction motors are good at full speed and full load, but at standstill they offer low impedance to the supply and thus, at full voltage, draw a high current. Starting current remains substantially constant at this high level, falling only slightly, until the motor is close to full speed. Power consumption during starting is therefore high. The condition is made worse by power factor. At the instant of starting, power factor is very low, usually less than 0.2, and rises slowly in relation to speed. This can incur a high energy cost, as the low power-factor tariff on industrial applications is high. Another consequence of the excessive current is the sudden application of torque to the driven load, imposing undesirable stresses on the motor and driven load. The losses in the

motor are also high and the starting duty needs to be carefully controlled and limited.

Direct-On-Line Starting (DOL)

Connecting the windings directly to the supply voltage when the motor is at standstill means that the flow of current is only limited by the system impedance. Starting current is usually of the order of five or six times the rated full-load current of the motor, Figure 6.41a, and may be more for larger machines. Inevitably there is a sudden high and undesirable surge of torque at the instant of switching on.

Star-Delta Starting
To limit the effect of full-line voltage at standstill, the windings are initially connected to the supply in a star


SOFT-START A.C. MOTOR CONTROL: C o n v e n t i o n a l S t a r t i n g

600 500
400 (5 300
_.1 1.1..

configuration; after a delay they are open circuited and reconnected in a delta configuration. Using this method the starting current is typically limited to twice the full load current. However, at the moment of reconnection in delta, a current surge occurs which may be as high as twenty times the motor full-load current, Figure 6.42. This peak is due to the residual back e.m.f, of the windings and its phase relationship to the supply voltage at the instant of reconnection. The duration of the peak is short, but the amplitude may be very high causing an impact torque in the mechanical system and a transient reduction of supply voltage.

2O0 100

l I I

20 300





Auto-Transformer Starting








Starting voltage can be controlled by using an auto-transformer and a series of tappings chosen to raise the voltage in stages during the run-up period. However, although starting current, at standstill, can be limited to the value of full-load current, the method still suffers from the disadvantage that the windings are open circuited several times during the run up, each time with a potentially high current and torque peak during reconnection, Figure 6.43.


6.41 Typical direct-on-line characteristics

a current b torque

(DOL) starting

Disadvantages of Conventional Starting

up to

600 500 400


"""" "-

& ~"~

I 20x FLC

There are two main disadvantages with the above-mentioned forms of motor starting; first, the current peaks, particularly those associated with star-delta and auto-transformer starting. When these current peaks occur, the supply system should be able to withstand them without any noticeable decrease in voltage. It is because of this that many electricity supply authorities limit DOL-started motors to a maximum of around 5.5 kW. In island installations where power is generated on site and there is no connection to a main power supply (a typical example being an offshore or desert-location oil or gas production installation) voltage regulation may be a serious problem. Such sites are extensively equipped with induction motors and starting of the very big motors has to be coordinated with generator availability. This is not usually a problem, given that procedures are effective and observed. The diverse duty cycles of the many medium-sized and smaller motors may, however, be capable of causing unpredictable system voltage regulation, all such motors commonly being DOL started. The second problem is that of mechanical shock due to the sudden torque stresses, caused by the current surges, which have to be absorbed by gearboxes, belts, drive shafts and driven equipment. Such stresses inevitably tend to reduce the serviceable life of these components.

200 100
0 ,

I I ,


40 n,%
/ i /







40 n, %




Figure 6.42 Star-delta starting

a current b torque

Difficulties can also be experienced with particular types of load such as cranes, where load oscillation can be started by the initial shock. Similarly, shock waves can be transmitted along hydraulic pipework, weakening joints in pumping systems. In conveyor systems, loads may be displaced or damaged on start up. What is needed, therefore, is a system

Chapter 6.4


" " " " " " ""' ~ "~' ~ ~ ~ % % % % O ,,.



04 d 300
._J LI-

0 200 100 0


i v



40 n,%




Figure 6.44 Electronic soft start

/ I

% \


200 -

~ . . . .


\ \ \

phase angle








100 ,,,'"


Figure 6.43 Autotransformer starting

a current b torque



that increases the voltage smoothly, so eliminating transient current and torque surges.


A commercial soft start consists of a power circuit containing six thyristors arranged in antiparallel pairs, and a control circuit which sequences the firing of these thyristors, Figure 6.44. The basic principle of operation is as follows. When power is applied to the thyristor bridge and a run command is given to the control circuit, the gate-pulse phase angle to each of the thyristors is gradually reduced at a rate set by the ramp time of the control circuit. The gate pulses allow the thyristors to conduct, and the decreasing phase angle smoothly increases the voltage to the motor windings as shown for a single phase in Figure 6.45. Because the voltage across the motor windings ramps up smoothly there are no current transients or torque surges.

t s S S ~

Figure 6.45 Output voltage control a 25 per cent b 50 per cent c 75 per cent

mixers screw and piston compressors centrifugal and piston pumps

fans circular saws stamping and cutting presses grinders gyrating crushers conveyors


APPLICATION BOARDS AND SOFTWARE: A p p l i c a t i o n s M o d u l e


Modern variable-speed drives commonly have the facility for the addition of option boards to bring extra functionality to the product. Many of these options have been described earlier in association with the particular products. One, however, deserves specific attention, as it not only has the power to transform the functionality of the drive product itself, but opens up the possibility of transforming the approach to designing an entire system. Control Techniques pioneered the use of applications boards in the 1980s allowing the user to program their own functions within the drive itself and with access to all drive parameters and inputs and outputs. The power and capability has advanced over the years and today users have the opportunity to realise complex algorithms using industry standard programming languages. Libraries of standard functions and applications are available, and connectivity to other field devices and factory automation systems has been enhanced by the development of fieldbus systems. Local intelligence of this form makes possible the realisation of powerful and dynamic distributed control architectures. Fully featured motion controllers, winders, indexers and many more functions can be realised, although the key benefit is that the power may be placed in the hands of the user. Drives engineers may think that they know the requirements of all applications but this is not true. The OEM or end user is the person who understands the application best, and is being provided with the tools to realise the optimum control strategy - optimum in terms of control

performance and cost. It is often possible to undertake entire machine control using this distributed architecture, eliminating local PLCs. The Unidrive applications module is centred around an Intel 960, 32-bit RISC processor. There is 96 kB of flash memory available for user programs, and 8kB of user RAM. The interface with the drive main processor is via dualport RAM providing bidirectional communications. The applications module can therefore not only read any of the drive's parameters, but can also write to those parameters allowing the set up/dynamics of the drive to be changed on the fly. As well as fieldbus interfaces, the unit has a fully configurable RS-485 port supporting: ANSI protocol as a slave or master controller, in twowire or four-wire mode at data rates from 300 bits/s to 19 200 bits/s MODBUS protocol (ASCII and RTU modes) as a slave only, at the above data rates the remote input-output unit high speed protocol at 38 400 bits/s

There is single character read/write access to the RS-485 port which allows other protocols to be simply implemented using the easy to use Control Techniques IEC 61131-3 ladder/function block programming tool. The tool is called SYPT (system programming tool), Figure 6.46. The external input-output unit, Figure 6.47, with modular analogue inputs, analogue outputs, digital inputs, digital




-.~~ |

Figure 6.46 SYPTprogramming tool

Chapter 6.5


Figure 6.47 Remote input-output module

outputs etc. provides the system builder with a highly efficient interface. An RS-232 port is provided for programming and debugging programs using the SYPT programming tool. High-speed, 345 ps or 460 ItS read/write access to the speed and torque reference within the drive allows highly dynamic control algorithms to be realised. Further, an internal singleaxis position controller is included which can be synchronised to the speed or encoder tasks. Full marker pulse and freeze support is implemented. Position control, speed control, digital lock and CAM profiling are all supported. 400 internal signed 32-bit registers are available for use with the SYPT programme, of which 200 are nonvolatile. Typical applications realised on an applications module include:

clutch-brake-based system. Eliminates overfilling to meet minimum weight requirements. Gap control - tail to head spacing, maintaining a constant gap between items on a conveyor regardless of product length. Photocells trigger monitoring of encoder input for product position, length and height. The applications module performs all the calculations required to deliver the product to the merge conveyor at the user-defined parameters. R e s u l t - easy configuration and control of product tail to head or head to head spacing onto a merge conveyor. Rotary cut o f f - to cut material to user-specified length and maintain cut on the registration marks. The applications module monitors the encoder position and product registration sensor to ensure that the product is cut in the correct position. Any changes to the drive parameters can be made on the fly. Result - a synchronised system providing fast and accurate cut lengths exactly placed on the registration marks. Positioning - proximity switches are placed at the desired position. The drive is configured to stop at an exact and consistent deceleration rate. Result - fast and accurate positioning of work piece with system cost minimisation.

More complex systems can be realised. Some further examples are given in Chapter 12.


As more and more capability is embedded into variablespeed drive products, it becomes necessary to provide software tools to allow the power to be utilised by users and not only the drives engineers. Control Techniques has developed a range of software tools for each range of digital drive products called Drivesoft. These software packages are again all very similar, but are tailored to each individual drive's needs. Drivesoft is a collection of WindowsXM-based set-up programs which allow the complete control and display of all parameters within a Control Techniques drive.

High-speed label printing (digital lock) - the applications module performs a complex CAM-type profile to ensure the placing arm and product are always at the same speed. The applications module compensates for small product registration shifts by using a product sensor on the master axis. R e s u l t - increased accuracy because placement follows product regardless of conveyor speed. Constant web speed unwind control with tension input the applications module controls the web speed and position based upon an encoder signal input. As the diameter of the take-up roll increases, the drive slows the speed of the motor. R e s u l t - precise speed regulation and the exact amount of material is wound onto the take up roll. Flying cut off, inline - when the correct product length passes, the cut bar is accelerated to match the speed of the product. When speed is matched, an output is activated sending the cutter head down. The operator is able to set the length using a Control Techniques CTIU operator interface. Result - easy data entry with fast and accurate cut cycles. Auger filler for dry material - the drive indexes exact revolutions for specific volume. The operator only needs to enter user units into the Control Techniques CTIU operator interface. Result - increased accuracy over a

Communications Modes
Drivesoft operates in two basic communication modes: online and offiine. In online mode the PC is connected via a serial cable to the serial port of the drive. Data from the drive may then be displayed, parameters written or read. All read/write parameters are available for alteration. In offiine mode Drivesoft requires no connection to the drive. Each parameter may be displayed and changed.

Drive Set-up Wizard

A drive set-up wizard guides the novice user in entering motor and application data. Help is provided for each step in the set-up wizard and, after the data is downloaded to the drive, a quick motor test can be performed.



Commissioning Screen
All operations within the commissioning screen are undertaken in the offline mode. To read the current status of a drive a read operation is required; to make any changes take effect within the drive a program operation has to be undertaken. The commissioning screen enables the simplistic setting of ramp data, for instance maximum and minimum operational frequencies, acceleration and deceleration rates together with autotuning.

start, stop, reverse, jog, reset and speed reference. The status of the inputs and motor parameters is displayed on panel meters for quick reference.

Parameter List
The parameter list allows the displaying of the complete list of parameters available within the drive. These tools make it possible for the nonspecialist to get the most out of even the most powerful drive. They also provide software management tools for version control etc.

Monitoring Screen
The monitoring screen allows control of the drive using computer control via serial communications. Controls are

Position and M o t i o n - C o n t r o l Systems



@ 2

180 187 188




The origins of motion controllers were closely related to textile machine developments. In the very early history of such machines the control of position and velocity was accomplished by elaborate, expensive and time-consuming solutions comprising cams, gears and shuttles. The automotive and machine tool industries were among those which saw the control of motion as a means of providing complex shapes and integrating complex operations. Being able to move heavy materials and process them in a repeatable and continuous manner added value and increased the productivity of their operations.

With the emergence of computers and microprocessor technology, other options became possible and modern motion controllers were born. A position or motion controller can be considered as a system for providing an outer position or path control at the front end of a drive which traditionally may be considered to be a speed controller. This can be represented as shown in Figure 7.1. Although position control is frequently used in single axes, the generic motion-control system is most frequently



A.C. supply

....................................................................................................................................1 [.....I............... ct o n ona0rve v e ......

position controller



"1 con,ro,,er I L


"1 con,ro,,er I
measured current

'vo a0e'

" conver,e,

I 0over


( measured position
Figure 7.1 Control loops within a position-control system

measured speed

speed demand


i drive 1

.... i

measured speed measured position motion controller speed demand

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .


A.C[ supply
drive 2


measured position

Figure 7.2 Two-axis system with a motion controller

associated with multiaxis applications. In such systems, the motion controller is often, but not always, used to coordinate the position or velocity of a number of drives. This form of application has been illustrated for a two-axis system in Figure 7.2. A more precise definition of a motion controller would be -

digital) which are fed to a drive (often referred to as an amplifier in motion applications) for controlling some type of actuator. Sometimes the controllers are designed for specialised applications such as: machine tools - computerised numerical controls (CNC) industrial robots transfer lines coordinate measuring systems laser welding and cutting

the application of programmable hardware and software (in conjunction with input sensory devices, actuators and other feedback devices)for the control of one or more linear or rotary motions.
The controller clearly includes a means of entering a set of instructions or code into its memory. These instructions are translated into a series of electrical signals (analogue and/or

In the 1990s general purpose standalone controllers became popular. This type of controller is typically more flexible

Chapter 7.2


than a dedicated controller and is adaptable to many different applications. The Control Techniques MC204 shown in Figure 7.3 is in this category. General-purpose motion controllers are used in a wide variety of applications, including such generic requirements as X-Y positioning, palletising etc. As well as being available as standalone units, motion controllers are also available as board-level components or can be integrated into a larger system. Various computerbased devices, such as programmable controllers, PCs, standalone industrial computers or remote mainframe computers serve to link and coordinate the motion-control function with other functions. In addition, an operator interface is present to input control logic, change existing programs or provide real-time modifications, such as system shut down or schedule changes. Increasingly, it is possible to incorporate position control into standalone drive products either through option modules or, for some simple motion functions, this may be built into the core product.

Figure 7.3 General-purpose motion controller- Control Techniques MC204


The requirement for a motion controller is essentially determined by the need to control a system or process beyond the capability of a single device. The control invariably requires the profiling of one or more axes within a machine. The properties to be profiled are usually either velocity (speed) or position. Before considering the applications further it is helpful to understand a few of the basic relationships associated with machine control, and how motion controllers can be used. First consider the relationships between the key motion parameters" distance (0) - velocity x time

acceleration (a) - velocity/time

= dv/dt (rate of change of velocity)

"~. dt (integral of jerk x time)

jerk (-y) - acceleration/time

- d a / d t (rate of change of acceleration)

Consider a motion profile, Figure 7.4, in which constant acceleration and deceleration is assumed; the position profile is also shown. If we analyse this simple motion in terms of acceleration and jerk, the latter being very important for the smooth transport of many items, including people in elevators, the results are seen in Figure 7.5. Figure 7.5 also includes the return motion to the original position. The high levels of jerk can be attenuated by changing the velocity/time profile. The implementation of a simple

= /v.

dt (integral of velocity x time)

velocity (v) - distance/time

= dO/dt (rate of change of distance)


[ ~. dt (integral of acceleration x time)





typical linear velocity

.9o.o_ =





state ,..I..,Vl..,




(/} .~

o o.o G) >

Figure 7.4 Motion profile showing constant acceleration and deceleration

if) e-

0 e-


return ~I position

if) o~


.m l-

~.z,-- t maximum ............................................................................. velocity

~- t

maximum acceleration

0 (.}

i !ilil ]ii i l


v t


Figure 7.5 Analysis of the motion profile of Figure 7.4

S-ramp function impacts the performance characteristics as shown in Figure 7.6. Further reductions in the jerk can be achieved by more complex speed ramps. Figure 7.7 shows the effect of a sine ramp.

Motion/position controllers are applied to many diverse applications. There does, however, exist a small number of functions which are common to many applications. A number of these are illustrated below.

C h a p t e r 7.2


target position

O t-



maximum velocity

4-, t-

J O >




. . . . . . . ~ . . . . . .

. . . . . . . . . . . . . . . . . . . acceleration

,L t v
(D O 0

I. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

.. maximum jerk


Figure 7.6 Simple motion profile incorporating S-ramp on velocity

t:3 o -

let position

. D



. m

position ~j.- t


. . . . . . . . . . . . . . . . .

~. . . . . .

maximum velocity

O >

% maximum acceleration

o O




maximum jerk

L ....~

Figure 7. 7 Simple motion profile incorporating sine ramp on velocity


TYPICAL MOTION FUNCTIONS" Position L o c k - E l e c t r o n i c G e a r b o x



This function is often used to provide a ratio between different drives. It can be used on machines which historically had a line shaft providing coordination. Three basic forms exist.

speed. In other words, during the acceleration period it is not important to maintain lock. Figure 7.9 illustrates the speed profile of such an application as well as showing the period when digital lock is applied. A typical implementation of such a system could be as shown in Figure 7.10.

Direct Positional Lock

In applications such as screw tapping (Figure 7.8), it is essential that two axes are directly locked in position. This lock, between spindle motion and the up/down motion, must be maintained at all times including acceleration and deceleration.

Ramped Rigid Lock

A third digital lock option is to allow the slave drive to lose synchronism with the master during acceleration, but to recover the position error when at the target speed. Figure 7.11 illustrates the speed profile of such an application as well as showing the period when digital lock is applied. A flying shear (rotary knife) is a good example of a system requiring this type of control. In this application only after synchronous lock is achieved can the shear, knife or punch be fired. When the shear, knife or punch is withdrawn the

Ramped Nonrigid Lock

In applications such as a take-off conveyor, it is only important to maintain lock once the drive has accelerated to

screw tapping spindle

up~down motion
~ili!i!i i i!i!i i i!i i !ili!i!i!i i!i i i!i!i i i!i!i!i!i~' i :i i i i i!i i i i i!i i i!!ili!i i i!i i !i !!ii i il ~!i~;iiii~!ii~ii~i,li~i!i~ii!i!~iii~;ili~!:

'~i i i i i i!i i i i i i i i~i i !i!ilili~!ii i ~i!~:~ii i~ iiiiiii!iii!ili!iiiiiiiiiiiTii!iii!iiii;!;ii!: :~i!iiiii!iiiiiiii!iiiiii!!iiiiiii!71i:ii ~

:i!i!i!iii!ii:i!i!i!i!ii!iil!:il .............................................. t

: i~ilii iiiii~: ~ i;~;!!! i

......................... :~i~i:iii}!!i!ii~!!!~!i!~!i!i!~!?

~i :

! ~

:~ili i i~
~iii)iiiTiiiil}iiiililiiiiiiilili;iTiil ~: :~iiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiii~' t

spindle motion

iilii....... iii!iiiiiiiiiiiliiiii::i::ii::i~ii::::=:::i~:~ i ..........................................................

Figure 7.8 Screw tapping

-(3 Q.

master speed

slave speed profile (x ratio)

system position locked

1 .

Figure 7.9 Ramped nonrigid lock speed profile

Chapter 7.3


master conveyor


pick and place

slave conveyor


line speed motion controller

. . . . . . . . . . . . . . . . . . . . . . . . . . . . .

"'1 I '

master position reference 2:1

. . . . . . . j

peed reference, slave position feedback

I ! !

Figure 7.10 Take-off conveyor- a ramped nonrigid lock system

master speed

recovery speed

slave speed profile (x ratio)

system position locked 1

Figure 7.11 Ramped rigid lock speed profile


.......... conveyorspeed

"13 c~


J, time

trigger point

.~ system locked

Figure 7.12 Flying-shear speed profile


TYPICAL MOTION FUNCTIONS: P o s i t i o n L o c k - E l e c t r o n i c


digital lock can be deselected and the position ramped back to the original datum. Figure 7.12 illustrates the speed profile of such an application as well as showing the period when digital lock is applied.

In all these examples of single-axis positioning the essence of the control system is the counting of pulses derived from the motor-mounted encoder or process-line-mounted feedback.


Single-axis positioning systems are characterised by the need to move from one position to another, usually in as short a time as possible and as accurately as possible. Further enhancements are often included to allow for position referencing, or homing. Figure 7.13 shows a typical system where the rotary motion of the motor shaft is translated into a linear motion v i a a ball screw. Alternatively in the case of an indexing application, a repetitive action is undertaken after a prescribed duration or travel, as in the example shown in Figure 7.14.

In many applications a slave-drive motion is required to follow the master in a nonlinear relationship. To illustrate this, consider the relationship between the slave and master in the following example of a packaging application (Figure 7.15). Such a profile has historically been achieved using a mechanical CAM, but has been substituted using electronic controls with the master-slave function embedded in look-up tables. The relationships between such tables and mechanical CAMs are of the form shown in Figure 7.16.

position control used once datum has been established any index position can be achieved within the full travel length

speed control for searching for home datum position


. _ .

proximity switch for home reference

i software and hardware limits can be implemented

--+ :+.)~
v v

!i:i:~,!! i

position and profile reference Figure 7.13 Simple positioning system

pneumatic knife unwind feed rolls take-off-conveyor

unwind software

-- -index position software

Figure 7.14 Simple indexing system

C h a p t e r 7.3




! ! | ! | | ! | | ! |

V (slave)

_ _

constant speed conveyor (master)

Figure 7.15 Packaging application showing nonlinear master-slave relationship

180 225

270 -I 315

slave axis

45 90

135 180

Figure 7.16 Mechanical CAM- master-slave relationship

The more CAM coordinates in the table, the smoother the motion between master and slave. This relationship is illustrated in Figure 7.17. The granularity/smoothness of motion can be improved by the use of interpolation techniques within the controller. In essence, interpolation calculates intermediate points between the data given in the table. The simplest form of interpolation is linear interpolation. Figure 7.18 shows the principle. As the name suggests, a linear motion between datum points is assumed. More complex interpolation algorithms can be found in many motion controllers including square interpolation (Figure 7.19), cosine interpolation (Figure 7.20) through to more complex helical and spline interpolation which will not be considered here. It is possible to combine different forms of interpolation to create very specific profiles. This flexibility allows complex CAM profiles to be effected with relatively small numbers of data points.

The power of motion controllers becomes of critical importance in multiaxis applications. It is all too easy to consider all multiaxis applications as highly demanding. This is not the case, indeed perhaps as many as 90 per cent of multiaxis systems demand little more than loose coordination of their respective motion. In such systems single-axis controllers can perform admirably with simple referencing between drives. It is in applications with precise contouring, such as axis drives on a machine tool, that true multiaxis control is demanded. In these applications it is not adequate to simply ensure that motion commands are accurately synchronised, but also that the motion of all drives is closely monitored; the actual position of each drive influences the trajectory of every drive. It is truly control of a point in space, and that space can have more than three dimensions! Multiaxis positioning systems also bring a requirement for multiaxis interpolation, which will not be discussed here.



reduced number of coordinates


c 0

required profile

(/) 0 Q. > (/)

increased number of coordinates

master position

>,, ,.i-, o 0

> (1) >



J ..i J

Figure 7.17 Effect o f granularity on master-slave motion smoothness





I x2

Xl dY/dX





Figure 7.18 Linear interpolation

Chapter 7.3






~X ~X









t m

----~ X









Figure 7.19 Square interpolation

a linear acceleration ramp in velocity b linear deceleration ramp in velocity


















---~ X

X 1 X'



I X2




Figure 7.20 Cosine interpolation

a sine acceleration ramp in velocity b sine deceleration ramp in velocity

Chapter 7.4



The power and flexibility of motion controllers demands powerful, easy to use programming tools. Today many proprietary programming tools are available. For generalpurpose controllers WindowsTM-based systems predominate. IEC 61131-3 provides a standard framework around which many systems have been successfully developed. Many standalone systems include the ability to import data from mechanical CAD packages providing a direct CAD/ CAM link. Text editors are incorporated to ensure good documentation and version control of all software. Further, performance can be monitored using features such as oscilloscope functions. CNC controllers tend to be programmed utilising broadly recognised codes for specific functions. The most widely

used listing of codes are G codes. Examples of G codes for various functions are given in Table 7.1.

Table 7.1 Examples of G codes for CNC controllers

G 00 G 01 G 02 G 04 G 06 G 17 G 21 G 45 G 94

rapidtraverse linearinterpolation circularinterpolation (clockwise) helical interpolation (clockwise) dwell splineinterpolation planeselect XY metricprogramming tool offset increase feed per minute

iii! ii i!i ii i i i!
iiiiiii!lll !!iiii iiiii iiiiill~~l~ iiiiii iiiii ili!iiiiiiiiiii i


]i!i i!i !i i i :i i
i !i i i i i i i i i i i~



~ :,Oh: ~m.tO ,~


iiiiiiiiiii liii!iiiii
iLii!i!iiiii!ii!ilili ~ iiliiiiiii~iiii!ii!ii

!iiiii iiiiii iiliii!i!iii!iiiiii ililiiiiiiii!iiiiiii il ii i

iit! Jiiii iiiii iiiiii iii

Figure 7.21 CAD to motion screen with automatic motion code generation




Position and motion-control systems are not necessarily complex. Most applications, when considered carefully, will involve the requirement for motion coordination rather than a rigid link. The recognition of this can greatly simplify the control system required and introduce the possibility of

radical fully distributed control topologies. The growth of PC-based programming tools provides great scope for future development relating to both centralised and distributed solutions.

Communications Systems

iiii'iiii iii~iiiiI


189 191 196 197

iiiiiii !iiii!!ii!ii!iiiii




In a modem digital drive, adjustments such as PID gains and acceleration ramp times are software settings rather than the potentiometers and DIP switches found in their analogue predecessors. To permit adjustment of these software parameters, most drives provide a human-machine interface (HMI) consisting of an onboard display and buttons. Users manipulate the buttons to scroll to the parameter of interest and to alter its value. The problem is that many sophisticated drives may have several hundred parameters and the built-in HMI system can be cumbersome to use. Compounding this problem is the fact that modem automation applications

usually involve multiple drives working in unison with each other. A field engineer may be faced with the daunting task of setting up the parameters for thirty drives, each one having over four-hundred parameters. To address this challenge, most modem digital drives have a built-in communications port using the RS-232 or RS-485 serial standard. There are two principal reasons for the inclusion of a communications port on a digital drive: set up of the drive's parameters and real-time control of a number of drives in an automation application.



Figure 8.1 Screen of typical drive set-up program



~ ~ :~::~

~:~i~ ~



referenCerun reference run


reverse reference run stop__ reverse direct wiring of signals

Figure 8.2 Traditional discrete interconnection wiring of a drive system

Many drive manufacturers provide WindowsTM-based set up and maintenance programs to acquire and alter the drive's parameters. These configuration programs can operate on a laptop computer and attach to the drive via the serial communications port. The standard PC COM port uses RS-232 which permits a single drive to be connected; if an RS-485 converter is used then multiple drives can be connected through a multidrop cable. Control Techniques' drives can be configured and maintained by PC applications such as Unisoft (for the Unidrive product range) which supports single and multidrop connections (Figure 8.1).

Chapter 8.2




serial network

Figure 8.3 Serial network interconnection of a drive system

Many applications require a number of drives to be integrated with I/O and PLC devices. The traditional approach is to use discrete wiring for the interconnections (Figure 8.2). Replacing the unwieldy wiring looms with a single digital serial network is an attractive alternative, as it will result in a dramatic reduction in the wiring and cost of installation (Figure 8.3). Provided that installation guidelines are met (i.e. maximum number nodes, trunk length and correct terminations etc.) reliability will be improved simply due to the reduced number of connections. The data remains in the digital domain with many analogue components and their associated conversion, repeatability and drift errors being eliminated. The flexibility offered by a serial network also allows applications to be configured for specific end user needs without adding extra cable connections. Internal parameters of remote nodes can easily be accessed for remote supervision and data logging. The speed of standard PC serial communications may be adequate for set up and configuration activities, but is usually too slow for this purpose. Consequently, a number of high-speed industrial networks have evolved to meet this need. They are generally known as fieldbus systems.

In conclusion, simple serial communication portals built into drives greatly enhance the task of set up and maintenance of the drive's adjustment parameters. For real-time control of the drive in an industrial process, a fieldbus offers the advantages of higher speed, simplified installation and enhanced reliability. The benefits in replacing discrete interconnections with a fieldbus are summarised below: wiring costs are dramatically reduced especially if distributed I/O is utilised data remains in the digital domain and most analogue components can be eliminated; no conversion errors, no repeatability and drift errors and substantial cost savings provided that the communication system is suitably robust the overall system becomes less susceptible to EM noise and ground-loop problems remote supervision and set up of the drives over the network if the selected factory network is one that has wide industry support, many sensors, actuators, HMI systems and controllers can be directly connected to it without special interfacing.


Communications networks have so much in common that it has become standard practice to relate their features and design elements to an internationally agreed model. The Open Systems Interconnection (OSI) Reference Model, developed by the International Standards Organisation (ISO),

is used as a framework for organising the various data communications functions occurring between disparate devices which communicate. The complete OSI model defines seven component parts or layers; however only three of these layers are usually developed for industrial communications.


required for the nodes to be connected in a logical bus structure.

Fieldbus systems often include special features related to device interoperability and the real-time nature of industrial automation applications. The device profile defines device interoperability features such as electronic data sheets which allow devices from different manufacturers to interoperate without complex configuration or custom software. Cyclic data is network data that bypasses parts of the software for efficiency.

Interface Circuits
All PCs support the RS-232 data transmission system. The signal appears as a single-ended voltage with reference to a signal ground. The voltage swing of the RS-232 circuit shown in Figure 8.5 is -+-12 volts ( - 1 2 V for logic one, + 12 V for logic zero). When not transmitting the signal is held at logic one ( - 1 2 V). Observing that the logic zero/one detection thresholds are typically +3 and - 3 volts in most implementations, it is clear that a modest spike could trick the receiver into making a false bit determination, as shown below in Figure 8.6. For this reason, RS-232 signalling circuits are almost never used in modern factory communications systems. Adoption of differential signalling solves most of the problems inherent in single-ended RS-232 communications. In the RS-485 standard, two conductors are used to represent the bit: one carrying the original bit and the other carrying its logical inverse. The differential receiver at the receiving end subtracts the two signals to recover the original data. Any noise induced into one conductor is induced in the other conductor and the subtraction operation will thus cancel out the common-mode noise. Note: the signal levels are 0 V and 5 V. RS-485 also supports operation with multiple nodes on the same cable. When a transmitter is not actually transmitting it is disabled and presents a high impedance to the bus. Typically, all receivers are enabled and see every message, but the protocol allows them to discard messages not intended for them.

Starting at the lowest layer in the OSI model (Figure 8.4), the physical layer is concerned with the actual transmission of raw bits. In a factory network, analogue quantities, switches, command codes and textual data are all converted into numeric information and transmitted as a stream of binary bits from the source node to the destination node. Typically, the bit stream is logically grouped into octets (or bytes). The following sections describe the key facets of the physical layer.

Network Cables and Connectors

The physical layer requires a transmission medium for the data signals to flow. The principal transmission medium of factory networks today is either copper cable or fibre-optic cable. Twisted-pair copper cable is preferred for cost and ease of installation. Fibre-optic cables are less susceptible to EM fields and offer higher bandwidths; however they are more expensive and difficult to install. Also, an expensive hub is

Data Encoding
I~~..............~Ye~i~i..~ii.~:~:.~i~~iP~P:,l:,i~~~i~i~~~i.~.i~.ilaye:..i.....~........~..~..~......... layer..7:..application layer.. . . .....~ .. . . .. .. ...................... .. layer 2: data link layer cyclic data
The differential signalling described in the previous section allows us to reliably send a single bit down a network cable. Since the network is typically used for numeric data or characters, multiple bits are used to convey the information. These multiple bits must be encoded either asynchronously or synchronously. In asynchronous encoding it is up to the receiver to properly sample and detect the multiple bits. The standard PC COM port uses a method of encoding called NRZ (non return to zero). In NRZ encoding, the voltage level determines the bit

Figure 8.4 Open Systems Interconnection

RS-232 __~~/interface


] ~

+ 12 volts logic zero ................ + 3 volts Ovolts

m m

m n

Figure 8.5 RS-232 interface

3 volts logic one - 12volts






+ 25 volts

logic zero . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

+ 3 volts

0 volts



1"l"--t ......... v'*s

........................................ 25 volts

Figure 8.6 RS-232- effect of noise


>> Rx

+5 ~tr data ansmitted

!]'Tx >>
m m





m B


note: the spike has been cancelled by the differential amplifier

kS )
f 0a*a


Figure 8. 7 RS-485

value (one or zero). In Figure 8.8, the character Hex 1C is transmitted. The receiving electronics must detect the change from line idle state to the start of a bit pattern (called the start bit). Once the start bit has been detected, each data bit must be sampled in the middle of its respective bit period (Figure 8.9). This assumes that the receiver has its own sample clock to do this and that the incoming bit rate (baud rate) is known. The disadvantages of NRZ encoding is that the timing of the sampling is independent of the transmitted signal. Also, the start bit and stop bit carry no information and thus waste throughput. Another encoding method called NRZI (nonreturn to zero inverted) uses transitions to determine the bit values (zero is no transitions, one is a transition). Finally, the Manchester encoding system allows the clock signal to be recovered from the transmitted data. In

Manchester encoding, there is always a transition in the middle of a bit period. Logical zero is a downward transition and logic one is an upward transition. Since there is always a transition, a phased-lock loop circuit can be used to extract the clock signal. This makes it easy for the receiver circuit to sample and detect the bits.

Network Topology
The topology of a network describes how the nodes are connected together. The main topologies are the bus topology where all nodes connect together onto a common medium, and a ring topology where nodes are interconnected in a unidirectional loop. Some networks are wired in a star topology which requires the use of a multioutput repeater called a hub. Ethernet 10baseT is an example of this approach.


logic 1 0 0 1 1 1 0 0 0

start bit

bit 0

bit 1

bit 2

bit 3

bit 4

bit 5

bit 6

bit 7

stop bit

Figure 8.8 Transmission of Hex 1C 0 0


Figure 8.9 Receiving Hex 1C

The data-link layer is responsible for encapsulating the digital information into message frames and for the reliable transfer of frames over the network.

A node waits until the bus is idle and then transmits its message. While transmitting, the node senses its own transmission to determine whether a collision with another node occurs. If a collision is detected then several schemes can be employed to arbitrate. Ethernet specifies that the node jams the network and then backs off for a random time interval before trying again. As network loading increases the probability of collisions increases and the network ceases to operate efficiently. This stochastic behaviour is not suitable for the majority of automation applications. CAN is also CSMA but the collision is resolved using bitwise arbitration according to the priority of the message. Significantly, this arbitration is accomplished on the physical layer itself using dominant bit signalling and does not result in wasted bandwidth. This signalling does have a bit rate and trunk length limitation of 4 0 m at 1 Mbit/s due to the finite propagation speed of the signal down the wire. CAN is deterministic for the highest priority message with a worst-case delay time of 130 gs (time for the maximum eight-byte message at 1Mbit/s). Performance can be improved further by using the hardware transmission timestamp to measure the delay and then transmitting this value to the slaves so that they can compensate for the jitter.

An example of a data-link frame is shown in Figure 8.10.

Data Model
All industrial fieldbuses work by connecting devices onto a shared medium and multiplexing data onto this medium in a serial fashion. The data model describes how messages or data are routed and identified on the network.

Messages are identified by a single node destination address. Most source/destination protocols also have the possibility for global addressing so that all nodes receive the message. However, it is not possible to select a group of nodes to receive the message (called multicast).

A message produced by a node is identified by its content (i.e. data identification) rather than a specific node destination. Any node may consume this message if it detects that the data is required. Clearly, this is very powerful and makes best use of available bandwidth but does require a complex configuration phase when data identifications are allocated. CAN, FIP and ControlNet can successfully operate a producer-consumer model although CAN only provides for a maximum of 2047 data objects.
Media Access Control

A node may only transmit a message if it is in possession of the token. Once the node has sent the message it must relinquish the token to its neighbour. This fair-share scheme provides a very flexible protocol for peer-to-peer communication without the need for a master or arbiter node. However, the worst-case time window for a node to receive the token and transmit a message is large and occurs when, in one token pass, every node transmits a message of maximum length. ARCNET is the best example of an industrial network which uses token ring.

The media access control protocol defines how access to the shared medium is arbitrated so that reliable data exchange can occur.

C h a p t e r 8.2




data I this is the data we want to transfer I identifies the node that is to receive the message start of frame bit pattern helps receiver circuits detect start of valid message number of bytes in the data field to follow

preamble is a burst of ones and zeros to enable phaselock loops at the receiver to lock in

cyclic redundancy check

bits help receiver verify that the message was received without error end of frame bit pattern helps receiver circuits detect end of message

Figure 8.10 A data-link frame


the retransmission is attempted too many times, then the message is discarded.

As the name suggests one node is designated the master usually, this is a unique node and also corresponds to the central logic controller of the system. The master controls all communication activity and the slaves only respond to a request from the master. This leads to deterministic behaviour but without any peer-to-peer communications. Much of the fieldbus installed base is PLC systems with a centralised architecture and a master-slave protocol. The two dominant networks in Europe are Profibus-DP followed by Interbus S, both firmly entrenched in the PLC-distributed slave architecture.

The producer-consumer model is very efficient if more than one node requires an item of data. However, in many automation applications the source-destination model with broadcast is adequate. Producer-consumer networks also need complex set up to allocate data IDs although if tools are provided by the vendor this can be relatively painless. Master-slave protocols only support centralised structures although slower peer-to-peer connections for nonreal-time data are possible on some networks. The producer-consumer capabilities of CAN without using a master arbiter make it an attractive solution for small networks. Token ring is the most flexible protocol for peer-to-peer communications but the deterministic behaviour is not as controllable.

Each node is allocated a time slot on the network when it is permitted to initiate a transaction with a peer node. In the WorldFIP system the time slots are controlled by the bus arbiter node, which stores a predefined list of data objects to be produced (note: FIP uses a producer-consumer model). The bus arbiter cycles through the list and broadcasts the data identification onto the network - the node that is set up to produce this data object recognises the ID and then produces the value on the network- any number of nodes then consume this data. At the time of producing data the node may request a slot for a noncyclic request. Once all the cyclic transmissions are complete the arbiter then cycles through all the cached noncyclic requests. ControlNet uses similar media access control.

The application layer defines and implements the services that the network offers each device. The most common services are read and write. The flexibility of the application layer generally incurs large overheads and consequently slower execution. Indeed, if the dynamic performance is critical then many networks bypass this layer for real-time data. Network data handled in this way is often called cyclic data.


In a factory environment, nearby lightning strikes, contact closures, power dips and other events may cause a transmitted message to be corrupted. A CRC polynomial inserted at the end of the message is used at the receiving end to determine if the message was corrupted. The data-link layer detects this and schedules the packet for retransmission. If The communication system, layers 1 through 7, manages the transfer of data between nodes. The profile or companion standard is a detailed specification of how this data is interpreted or mapped onto device functions. A common misconception is to assume that, with compatible communications, devices are interoperable: interoperability is only truly achieved if the profile layers are implemented.




There are a number of protocols designed for RS-232 or RS485 communications which are still widespread in automation applications.


All Control Techniques' products offer a simple ASCII-based protocol which allows parameters to be read and written (Figure 8.3). The response is a single character A C K (hex code 06) or N A K (hex code 15).

The Modbus serial communications protocol is a de facto standard designed to integrate PLCs, computers, terminals, sensors and actuators. Modbus is a master-slave system meaning that one device, the master node, controls all serial activity by selectively polling the slave devices. Modbus supports one master device and up to 247 slave devices. Each device is assigned a unique node address. There are two variants of Modbus: ASCII and RTU. ASCII mode uses a message format that is printable, messages start with a colon and end with a carriage return. RTU mode uses binary and is therefore not printable. Eightbit characters are sent as a continuous burst and the end of the message is denoted by three and a half character times of silence. RTU mode messages use half the characters of an equivalent ASCII message. Only the master initiates a transaction. The master is usually a host PC or HMI device since most Modicon PLCs are slaves and cannot initiate a Modbus transaction (the new Quantum PLCs can act as a Modbus master). Typically, the host master will read or write registers to a slave. In each case, the slave will retum a response message. For a read operation, the response will carry the requested data. For a write operation, the response is used to verify acceptance of the write command. A special case is the broadcast operation where a write operation can be directed to all slaves. In this case, no response message is forthcoming. The eight-bit address field is the first element of the message (one byte for RTU, or two characters for ASCII). This field indicates the address of the destination slave device that should respond to the message - all slaves receive the message but only the addressed slave will actually act upon it. The function code field tells the addressed slave which function to perform. Modbus function codes are specifically designed for interacting with a PLC on the Modbus industrial communications system. Two error check bytes are added to the end of each message: ASCII mode uses a longitudinal redundancy check (LRC) and RTU mode uses a 16-bit CRC check. In the example in Tables 8.1 and 8.2, the host PC is initiating a read request of three parameters starting with 1.08 from drive address 06. The starting holding register is 40108 but the 4 is dropped in the message string and the rest of the register address is entered as one less (0108 becomes 0107, 0107 is entered as 006B in hexadecimal). The response repeats the address and function code, but includes the values read from the drive.

Table 8.1 Query

Field name Header Slave address Function Starting address hi Starting address lo No. of registers hi No. of registers lo Error check Trailer
Total bytes

RTU (hex) none 06 03 00 6B 00 03 CRC (2 bytes) none


ASCII characters :(colon) 06 03 00 6B 00 03 LRC (2 characters) CRLF


Table 8.2 Response

Field name Header Slave address Function Byte count Data hi Data lo Data hi Data lo Data hi Data lo Error check Trailer
Total bytes

RTU (hex) none 06 03 06 02 2B 00 00 00 63 CRC (2 bytes) none


ASCII characters :(colon) 06 03 06 02 2B 00 00 00 63 LRC (2 characters) none


Table 8.3 Query

Field name EOT Slave address Example (hex) 04 31 31 34 34 02 31 38 32 33 33 32 2E 37 39 3 1 char

ASCII 1 1 4 4 1 8 2 3 3 2 7 9

STX Parameter


ETX Checksum
Total bytes

Chapter 8.4




Drive applications dictate specific fieldbus requirements. The following section reviews these requirements.

effects. Some applications will require close coordination of a number of drives (e.g. each axis in a CNC machine). In these applications set points and measured values in all drives must be consumed and sampled synchronously across the network. To achieve these dynamic requirements the real-time or cyclic data is usually handled differently, bypassing the cumbersome application layer, and presented directly to the drive. Moreover the transactions and mapping of the data is generally predetermined during system initialisation to avoid the overhead of sending addressing information.

Physical Layer
Contactors, power switching devices and switch gear all contribute to a high level of ambient EM noise. The transmission media must be immune to this and also be mechanically robust enough to survive the rigours of installation. The type of media and signalling method will dictate the limits for the maximum number of nodes on a network and the distance between them (Table 8.4). This should be checked against the application requirements. Screened twisted pair is generally preferred for cost reasons and ease of installation.

General Message Services

Other general nontime-critical access to the network must be supported. For example, operator interaction, drive set up (PID gains, ramp times etc.), downloading application software, general diagnostics and data logging. These messages are often referred to as noncyclic communication.

Error Detection
The consequence of accepting corrupted data will be catastrophic in any application and a reliable error-detection algorithm must be implemented, which will discard corrupted data. Error-recovery schemes are generally too complex for small embedded systems and would also require redundant data to be transmitted. Many systems include built-in, high-level error-handling schemes which call for retries ifa frame is corrupted. However, for real-time data, which has a limited lifetime, the system is best waiting for the next available sampled data.

Centralised v e r s u s Distributed Intelligence

If a communication system with the above dynamic characteristics is available it may seem desirable to locate all the system control functions in the central controller and simply transfer demands and feedback cyclically with the drives over the network. However, if the drive is capable of executing some application functions locally then an alternative approach is to use a distributed architecture. In this scenario some control loops are implemented locally utilising feedback available from the local I/O on the drive. This reduces the load on the network, which is then only called upon to transfer slower outer-loop references which are less sensitive to variations in the deterministic response. Although the dynamic requirements of the network are less, flexible peer-to-peer communication is needed to allow data to be shared easily between distributed processes. A distributed architecture has some advantages over a centralised structure: a powerful controller necessary to cope with large levels of computation for the whole system is expensive and the software will invariably be unwieldy and difficult to maintain. Also, the standard I/O which most drives provide can be utilised very efficiently by a local process without loading the network. Clearly, it may be difficult to partition and distribute some systems and a centralised approach is best (e.g. the interpolation functions of a multiaxis CNC machine).

Dynamic Performance
The bandwidth of a drive varies according to the motor and drive technology used. A servodrive system may have a bandwidth of 1 or 2 kHz. At the other extreme, an open-loop drive coupled to a large fan or pump may have a system time constant measured in seconds. To accommodate the high-performance servo applications the data will need to be delivered every 1 ms or less. Only a small set of real-time data items needs to be transferred at this rate, e.g. speed and torque, control and status. Moreover, the updates must be periodic and deterministic; in other words the data must arrive at regular guaranteed time intervals. Ideally the period or cycle of the updates should be synchronised with the digital control loops in the drives and controller to eliminate any undesirable beating

Table 8.4 Typical data rates for transmission media

Medium Screened twisted pair Coaxial Optical fibre

Typical data rate <_5 Mbit/s < 10 Mbit/s _< 100 Mbit/s

Profibus DP is a mature fieldbus system that is primarily used to connect distributed slave devices (I/O, drives, sensors, actuators etc.) to a PLC. Profibus DP is optimised for



this master-slave architecture and can provide high performance updates when used at the maximum 12 Mbit/s data rate. The dominant physical layer is a multidrop RS-485 bus using shielded twisted-pair cable. Secondary masters (class 2) can be added to the network and can be used for SCADA-type functions only. A master token is used to control the master accesses. There is a wide range of manufacturers in the Profibus user group and a number of device profiles (including a drive profile) have been defined. Development work on DP to address the synchronisation and peer-to-peer limitations is ongoing. These enhancements are both needed for motion control applications. To synchronise the slave drives, the master broadcasts a cyclic and equidistant clock telegram which all slaves use to synchronise internal clocks. The peer-to-peer data is realised using a producer-consumer data model.

capability of CAN. A number of device profiles have been defined including the A.C./D.C. drives profile. Assembly objects - cyclic control data is efficiently packed into the eight-byte CAN data frame without any protocol information. Assembly objects are usually configured for synchronous exchange with the master synchronisation telegram. Explicit messaging - this noncyclic service allows a single object attribute to be read or written. The CAN message carries the protocol information necessary to identify the object attribute and the value.

This is an open-system standard developed by a multivendor working group. A number of device profiles have been defined including the drive speed and position profile. Process data objects (PDOs) - control data is efficiently packed into the eight-byte CAN data frame without any protocol information- the 11-bit COB ID (CAN object identifier) is used to identify the data. The COB IDs must be allocated during system initialisation. PDOs can be transmitted cyclically in response to the synchronisation telegram, or as unsolicited event-based transmission. Service data objects (SDOs) - discrete messaging to read or write an object.

Interbus S is a master-slave fieldbus system optimised for connecting slave devices to a central controller (typically a PLC). Although the data rate is a modest 500 kbit/s, the useful data throughput is high due to the low protocol overhead. Slave devices typically provide between four and ten cyclic transmit and receive data words in conjunction with a noncyclic channel (PCP). The network uses a ring topology with the master coordinating all network activity. There are two physical-layer options: the remote bus variant uses RS-485 twisted pair with a maximum range of 400m between each node; the lower cost local bus uses TTL levels for short spurs.

Control Techniques developed CTNet as part of the distributed control product strategy. The objectives for CTNet were low cost, easy to install and set up, peer-to-peer exchanges to allow data to be shared between drives, and high performance. CTNet leverages proven token-ring data-link technology with a unique protocol stack having dual cyclic data channels to provide efficient synchronous transfers for real-time control data. In addition, the general-purpose channel provides occasional or event-driven transactions suitable for data logging, diagnostics and set up. A bus topology is used which, with the use of repeaters, can support up to a maximum of 255 nodes. The network uses a transformer isolated physical layer, which improves noise immunity and makes installation simpler. The default data rate is 2.5 Mbit/s. When equipped with CTNet, the drive application modules can be used to build distributed control systems which offer many performance and cost benefits over an equivalent PLC-based system.

CAN was originally developed by Intel and Bosch for automotive applications. Its performance, simplicity and low cost have led to its adoption in many industrial automation applications. The main CAN-based standards are DeviceNet and CANopen. The CAN physical layer is typically very robust but does have a baud rate v e r s u s distance trade off due to the bit-wise arbitration scheme.

Originally a Rockwell system, this is now a standard managed by the ODVA. Most DeviceNet systems operate under a master-slave scheme with a master PLC scanner. However, this structure does not fully leverage the peer-to-peer


Supply Harmonics due to Drives



~:~:.i " '..".~

201 203 204 206

4 @ @ 5 6



Supply harmonics are caused by the A.C. input current to load equipment departing from the ideal sinusoidal waveshape. They are produced by any nonlinear circuit, but most commonly by rectifiers. The supply current waveform is generally measured in terms of the harmonics of the supply frequency which it contains. The harmonic current causes harmonic voltage to be experienced by other equipment connected to the

same supply. Because harmonic voltage can cause disturbance or stress to other electrical equipment connected to the same supply system, it is controlled by regulations applying to public supply systems. If installations contain a high proportion of variable-speed drives and/ or other power electronic equipment such as UPS, then they may have to be shown to satisfy the supply authorities' harmonic guidelines before permission to connect is granted.



As well as obeying regulations, users of drives need to ensure that the harmonic levels within their own plant are not excessive. Some of the practical problems which may arise from excessive harmonic levels are: poor power factor, i.e. high current for a given power interference to equipment which is sensitive to voltage waveform excessive heating of neutral conductors (single-phase loads only) excessive heating of induction motors high acoustic noise from transformers, busbars, switchgear etc. abnormal heating of transformers and associated equipment damage to power-factor correction capacitors

An important property of harmonics is that they tend to be cumulative on a power system, i.e. the contributions of the various harmonic sources add up to some degree. It is worth emphasising this difference from high-frequency electromagnetic compatibility (EMC) effects, which may cause interference in sensitive data and measuring circuits through stray coupling paths. High-frequency effects tend to be localised and not significantly cumulative. It is important to be clear that, with few exceptions, if harmonics cause disturbance it is through direct electrical connection and not through stray paths. Screening is rarely a remedial measure for harmonic problems (telephone interference is a possible exception).


There are two kinds of regulation which may need to be considered.


Installation regulations are imposed by the electricity supply authority to protect other electricity consumers from the effects of excessive harmonics. They are usually based on an agreed level of voltage distortion which can be tolerated by correctly designed equipment. This is specified in terms of a total harmonic distortion (THD) (see definitions). The internationally accepted maximum THD compatibility level in a low-voltage system is 8 per cent, and to achieve this with a high degree of confidence it is usual to aim for a rather lower level as the planning level, typically 5 per cent. Individual harmonics are also subject to limits. Some relevant standards and regulations are given in the Appendices. From the point of view of the supply authority, the relevant harmonic voltage is at the point of common coupling (PCC) with other power consumers. The harmonic levels within the consumer's premises may be higher because of the impedance of cables and transformers. In large installations measures may be necessary to prevent harmonic problems within a site. Since there are no statutory requirements, a relaxed version of the authority limits can be applied internally. It is not advisable to allow the 8 per cent THD compatibility level to be exceeded, because the majority of equipment will have been designed to be immune only up to this level. Calculating the voltage distortion can be an expensive undertaking, because it requires existing harmonics to be

measured over a period of time, the system parameters such as source impedances to be derived and the effect of the planned new load to be estimated. For a large installation with a high proportion of the load comprising electronic equipment, it is cost effective to complete this exercise in order to avoid the application of unnecessary remedial measures. For simpler cases a full analysis would be burdensome. Regulations such as the UK G5/3 (to be replaced by G5/4, probably during 2001) provide simplified staged procedures to permit connection based only on harmonic current data, which can be obtained readily from the manufacturers' technical data. This involves making simplifying assumptions which are biased in a cautious direction. If the simplified stage does not permit connection, the full calculation procedure has to be applied.


A further simplification of the guidelines can be made if a product conforms to a relevant harmonic standard, when it can be connected without reference to the supply authority. The international standard for equipment rated at less than 16 A is IEC61000-3-2, the corresponding CENELEC standard being EN61000-3-2. These are applied to consumer products and similar equipment used in very large numbers, where individual permission to connect would not be practical. In the EU, EN61000-3-2 is mandatory for equipment within its scope from 2001. Small variable-speed drives rated at less than about 650 W shaft power fall within the scope of this standard, and can be made to conform to it by

C h a p t e r 9.3


the application of suitable measures. However, where they are used in large quantities in a single installation it may be more cost effective to assess their total current and obtain permission to connect from the supply authority.

In the future there will be a further standard, IEC61000-3-12 (EN61000-3-12) covering equipment rated up to 75 A (see IEC61000-3-4).


Harmonic current is generated by the input rectifier of an A.C. drive. The only exception is for an active input stage, where PWM is used to create a sinusoidal back e.m.f., and there is, in principle, no harmonic current. The only unwanted current is at the PWM carrier frequency, which is high enough to be relatively easy to filter. This arrangement is discussed later. The essential circuit for a typical A.C. VSD is shown in Figure 9.1. The input is rectified by the diode bridge, and the resulting D.C. is smoothed by the capacitor and, for drives rated typically at over 2.2kW, the inductor. It is then chopped up in the inverter stage which uses PWM to create a sinusoidal output voltage of adjustable voltage and frequency. Supply harmonics do not, however, originate in the inverter stage or its controller, but in the input rectifier. The input can be single or three phase. For simplicity the single-phase case is covered first. Current flows into the rectifier in pulses at the peaks of the supply voltage as shown in Figure 9.2. Figure 9.3 shows the Fourier analysis of the waveform in Figure 9.2. (Note: all currents shown in spectra are in peak values, i.e. x/2 times their r.m.s, values.) The figure comprises lines at multiples of 50 Hz, and because the waveform is symmetrical in the positive and negative half cycles, apart from imperfections, even-order harmonics are present only at a very low level. The odd-order harmonics are quite high,

but they diminish with increasing harmonic order. By the 25th harmonic the level is negligible. The frequency of this harmonic for a 50 Hz supply is 1250 Hz which is in the audio frequency part of the electromagnetic spectrum and well below radio frequency, which is generally considered to begin at 150 kHz. This is important, because it shows that supply harmonics are low-frequency effects, which are quite different from radio frequency EMC effects. They are not sensitive to fine details of layout and screening of circuits, and any remedial measures required use conventional electrical power techniques such as tuned power-factor capacitors and phase-shifting transformers. These should not be confused with the various techniques used to control electrical interference from fast switching devices, sparking electrical contacts etc. Three-phase drives cause less harmonic current for a given power than do single-phase drives. Figure 9.4 shows the input current waveform for a 1.5 kW three-phase drive. The line current is less in any case, and there are two peaks in each mains cycle each of about 20 per cent of the peaks in the single-phase drive. Figure 9.5 shows the corresponding spectrum. Compared with the single-phase case the levels are generally lower, and the triplens (multiples of three) are absent. The actual magnitudes of the current harmonics depend on the detailed design of the drive, specifically the values of D.C. link capacitance and inductance. Therefore the supplier must be relied on to provide harmonic data.

1 or 3 ~supply

0 ..........................

Figure 9.1 Essential features of A.C. variable-speed drive





. supply

~o , / ~ o



/ ~





> >




Figure 9.2 Typical input current waveform for a 1.5 kW single-phase drive (with supply voltage)





frequency, kHz




Figure 9.3 Corresponding spectrum for Figure 9.2

2 0


go ,_
~ .
-10 "


bme ms


Figure 9.4 Typical input current waveform for a 1.5 kW three-phase drive

Chapter 9.4


O .



08 12 1.6



frequency kHz

Figure 9.5 Spectrum corresponding to Figure 9.4

There is no difference in principle between the harmonic behaviour of A.C. and D.C. drives, but the following aspects of D.C. drives are relevant: The current waveform is not affected by the choice of design parameters (inductance and capacitance) in the drive. It does not therefore vary between drive manufacturers and can be calculated from knowledge of the motor armature inductance, source inductance and pulse number. The phase angles of all harmonics change with the rectifier firing angle, so the harmonics from multiple drives do not add up arithmetically. D.C. drives tend to be most often used at high power levels, and often a dedicated transformer is provided, so 12-pulse and higher pulse numbers are more readily provided.

Effect of Loading
In the case of an A.C. drive, the input current is proportional to the load. As the load falls, all of the main harmonics also fall, but not as rapidly as does the fundamental. In other words, the THD deteriorates as the load falls. This applies whether the power reduction is through reduced speed or torque or both. In the case of a D.C. drive, the above applies for output current, and hence motor shaft torque. However, the current does not fall significantly as the speed falls. At light load the waveform may improve somewhat at low speed if the current becomes continuous, but at full torque low speed the harmonic structure is much the same as at maximum speed. The highest harmonic current for a given drive invariably occurs at maximum load, but in a system with multiple drives it may be necessary to look in detail at the effect of load combinations.


Some of the effects of harmonics were summarised in Chapter 9.1. Figure 9.6 shows a voltage waveform where a distribution transformer is loaded to 50 per cent of its capacity with single-phase rectifiers. It shows the characteristic flat-top effect. Although this waveform looks alarming, in fact it would not affect most modem electronic equipment. However, the

harmonic content can cause excessive stress in components, especially capacitors, connected directly to the supply. The diode bridge input circuit in a single-phase A.C. drive is the same as that used in a very wide range of electronic equipment such as personal computers and domestic appliances. All of these cause similar current harmonics. Their effect is cumulative if they are all connected at the same lowvoltage (e.g. 400V) supply system. This means that to






o >


-400 J

Figure 9.6 Supply voltage waveform with single-phase loads of 50 per cent supply capacity

estimate the total harmonic current in an installation of single-phase units, the harmonics have to be added directly. Phase-controlled equipment such as lamp dimmers and regulated battery chargers cause phase-shifted harmonics which can be added by root-sum-squares to allow for their diverse phase angles. In a mixture of single and three-phase loads, some of the important harmonics such as the fifth and seventh are 180 out of phase and actually mutually cancel. Sometimes this information can be very helpful even if there is no certainty that the loads will be operated simultaneously- for example, in an office building which is near to its limit for fifth and

seventh harmonic because of the large number of singlephase computer loads, the installation of three-phase variable-speed drives will certainly not worsen the fifth and seventh harmonics and may well reduce them. Overloading of neutral conductors is a serious concern in buildings containing a high density of PCs and similar IT equipment. It is caused by the summation of triplen harmonics in the neutral conductor-the neutral current can equal or even exceed the individual phase currents, whereas it is common for the conductor to be of reduced area. Singlephase A.C. drives would have a similar effect, but it is unusual for them to be used at such a high density.



The calculation of input current for a controlled rectifier is covered in most of the standard textbooks on power electronics. A particularly clear and comprehensive account is given in IEEE Standard 519-1992. The following is a summary of the technique and the results. The basic theory of the controlled rectifier assumes an infinite inductance load. Then for a p-pulse rectifier the input current has a stepped waveshape with p regularly spaced steps in each cycle. This can readily be shown to contain no even harmonics, and only odd harmonics of the order n = kp 4- 1 where k is any integer.

The amplitudes of the harmonics follow the simple rule for a rectangular wave: I1
In z fl

For some purposes this simple calculation is sufficient, but the influence of finite inductance on the D.C. side should be taken into account, and the A.C. side inductance may also have a significant effect. Figure 9.7 shows the effect of D.C. current ripple on the four dominant harmonics of a six-pulse drive. The fifth harmonic increases steadily with increasing ripple, whereas for moderate ripple levels the other harmonics fall.

C h a p t e r 9.5


typical practical values



I I I I ! ! I I I I .

With large drives the IEC61800-3 fault level is unrealistic, and a fault level such as 16 kA, which corresponds with the capability of widely available switchgear, may be used.


The impact of a harmonic current on the power system can be estimated by calculating the resulting harmonic voltage at a point in the supply system shared with other equipment. The power supply companies have a duty to control the quality of the power delivered to consumers, so their interest is at the point where the supply is shared with another consumer- the point of common coupling (PCC). The basic equivalent circuit for this calculation is shown in Figure 9.9. For the study of harmonics, the principle of superposition is used which means that the mains source is turned off and the consumer being studied is considered as a source of harmonic current, as shown in Figure 9.10. Each harmonic is considered in turn. The voltage is simply the product of the current and the impedance of the supply system upstream of the PCC. The impedance at 50 Hz (or other mains frequency) can be found from the declared fault level of the supply, which should be available from the supply company. If it is expressed in MVA then the impedance in ohms at mains frequency can be calculated as follows:

-'- 3 0
0 0 .m tO


~ 2o



i i i i .'~








peak to peak current, % of IPc (av.)

Figure 9.7 Six-pulse converter-variation of line-current harmonic content with ripple current

o~ 2 0 _

0 0 -



g ~


Zs = MVA x 106





where V is the voltage between lines, and Zs is the fault impedance of one line. The impedance is assumed to be predominantly inductive, as is the case with high-power circuits, so that for a harmonic of order n the impedance is nZs. This calculation is required for assessment against stage 3 of the UK Electricity Association recommendation G5/3. It is widely accepted as giving a reliable basis for assessment of harmonic penetration. The presence of power-factor correction capacitors causes a more complex situation where resonance causes the impedance to rise at certain frequencies. If these coincide with odd harmonics where substantial currents exist, a higher harmonic voltage than estimated can occur. Fortunately this is an unusual situation, as it can be expensive to cure. Note that in Figure 9.10 the harmonic voltage within the premises of consumer 1 will be higher than that at the PCC, because of the voltage drop in Zc~. Meeting G5/3 at the PCC is no guarantee of tolerable harmonic levels within the system of the consumer generating the harmonics. In order to analyse a practical system, the known harmonic data for all the rectifiers and other distorting loads must be combined to predict a total current. In general, each harmonic from each unit is a vector quantity which can only be added to the others through vector addition. Usually the phase angle is unknown, and in the case of phase-angle controllers it varies with the operating condition. For uncontrolled rectifiers, the phase angles of the dominant harmonics will be similar, and the amplitudes add directly.

line impedance, %

Figure 9.8 Variation of line harmonic content with line impedance

Figure 9.8 shows the effect of supply inductance on these harmonics, all of which fall with increasing inductance, particularly the l lth and 13th. However, the benefit is reduced at large firing angles, so for operation at high torque and low speed (low voltage) the benefit of supply inductance may be minimal. Further information is given in IEEE 519.

I N D I V I D U A L D R I V E S - A.C.
It is not usually practical for the user to calculate the harmonic current for an A.C. drive, since the circuit parameters will be unknown, and in any case the values of D.C. inductance and capacitance used are such that a full nonlinear circuit analysis is required. The drive manufacturer should at least provide the harmonic current values at full load, and preferably at part load also. Linear interpolation of the harmonic currents as a proportion of the fundamental can then be used to estimate other loadings. For small A.C. drives with D.C. inductance, the supply impedance has a considerable influence. IEC61800-3 recommends that the fault level be assumed to be 250 times the drive input current rating. Data should also be provided for where line reactors are fitted.





fault level at PCC determined by Zs




consumer 1 consumer 2

point of common
coupling (PCC)

consumer 3

Figure 9.9 Supply system, showing point of common coupling with fault level


harmonic voltage at PCC

I "!



harmonic current
consumer 1

Zc~ Zc~



consumer 3

Figure 9.10 Supply system arranged for harmonic analysis

G5/3 permits the application of a coincidence factor of 0.9 to reflect the fact that perfect addition is not possible. Where phase-angle-controlled loads are added to one another or to a group of uncontrolled loads, the random phase angles mean that addition by square root of sum of squares is appropriate, whereas G5/3 suggests a coincidence factor of 0.75. Diversity of loading is also an important issue. In some installations only a small part of the possible load on each drive can occur simultaneously. This must be considered to avoid an over estimate of the harmonic loading.

If the system is supplied by isolated generators not connected to a grid, the impedance of the generators must be determined. The relevant parameter is the direct-axis subtransient reactance, x~ (strictly the quadrature-axis impedance should also be considered, depending on the load angle; in practice they are usually similar). Typical values are between 14 and 20 per cent, compared with the 5 per cent of a typical distribution transformer, so generators are less able to tolerate harmonic current than the public supply network.


The first point to make is that harmonics problems are unusual, although with the steady increase in the use of electronic equipment problems may be more common in future. The situations where problems have occurred most frequently are in office buildings with a very high density of personal computers, and in cases where most of the supply capacity is used by electronic equipment such as drives, converters and UPS. As a general rule, if the total rectifier loading (i.e. variablespeed drives, UPS, PCs etc.) on a power system comprises less

than 20 per cent of its current capacity then harmonics are unlikely to be a limiting factor. In many industrial installations the capacity of the supply considerably exceeds the installed load, and a large proportion of the load is not a significant generator of harmonics - uncontrolled A.C. induction motors and resistive heating elements generate minimal harmonics. If rectifier loading exceeds 20 per cent then a harmonic control plan should be in place. This requires that existing levels be assessed, and a harmonic budget allocated to new equipment.

Chapter 9.6


Calculations using the techniques described in Chapter 9.5 may be required to predict the effect on harmonic voltage from connecting additional equipment. The following measures can be used to reduce the harmonic level.

those shown in Figures 9.2 and 9.3 to those shown in Figures 9.11 and 9.12. It is particularly beneficial for the higherorder harmonics, but the fifth and seventh are reduced by a useful degree. Only the third harmonic is little improved. Since the three-phase rectifier has no third-harmonic current, the A.C. inductor is even more beneficial, as shown in Figures 9.13 and 9.14. In these examples the value of the A.C. inductor is 2 per cent (i.e. 0.02 p.u.). This is the highest value recommended where full torque at base speed is required, since the drive output voltage at full load begins to be reduced significantly for higher values.

When planning a new installation, there is often a choice of connection point. The harmonic voltage caused by a given harmonic current is proportional to the system source impedance (inversely proportional to fault level). For example, distorting loads can be connected to main busbars rather than downstream of long cables shared with other equipment.

Additional D.C. Inductance

Drives rated at 4 kW or more usually have three-phase input and include inductance in the D.C. link. This gives the improved waveform and spectrum shown in Figures 9.15 and 9.16, which are for a hypothetical 1.5 kW drive for ease of comparison with the previous illustrations. Further improvement is possible by adding A.C. inductance as well as D.C., as shown in Figures 9.17 and 9.18. This represents the limit of what can be practically achieved by very simple low-cost measures.


As shown above, harmonic current for a three-phase drive of given power rating is about 30 per cent of that for a singlephase drive; and there is no neutral current. If the existing harmonics are primarily caused by single-phase loads, the dominant fifth and seventh harmonics are also reduced by three-phase drives.


Standard three-phase drives rated up to about 200 kW use six-pulse rectifiers. A 12-pulse rectifier eliminates the crucial fifth and seventh harmonics (except for a small residue caused by imperfect balance of the rectifier groups). Higher pulse numbers are possible if necessary, the lowest harmonic for a pulse number p being ( p - 1). Individual A.C. drives may be supplied with D.C. from a single bulk 12-pulse rectifier, or where the loading on drives is known to be reasonably well balanced individual six-pulse drives may be supplied from the two phase-shifted supplies.


Series inductance at the drive input gives a useful reduction in harmonic current. The benefit is greatest for small drives where there is no D.C. inductance internally, but useful reductions can also be obtained with large drives.

Additional A.C. Supply-line Inductance

The addition of A.C. input inductance to the single-phase drive improves the current waveform and spectrum from








" " "









Figure 9.11 Input current waveform as Figure 9.2 but with 2 per cent input inductor


REMEDIAL TECHNIQUES: Use a Higher Pulse Number (12 Pulse or Higher)




~....._..~ ~






frequency, kHz

Figure 9.12 Input current spectrum for Figure 9.11


0 I

20 .
time ms



-8 Figure 9.13 Input current waveform as Figure 9.4 but with 2 per cent input inductors


0 . . . . . . 0 . . . . . .

. . . . . . .

. O . . . . .

0 . 0 .

< "= 2 c

. . . . . .

0 . . . . . . .

. . .

0 j
o., o.~ ~.~ ~.o ~.o

frequency, kHz

Figure 9.14 Input current spectrum for Figure 9.13




4 < 0



lo I
o o J

o o


time, ms

i i


Figure 9.15 Input current waveform for 1.5 kW drive with D. C. inductance




018 frequency, kHz




Figure 9.16 Spectrum for Figure 9.15

.,...; cOD s..

o o o

0 O


Figure 9.17 Input current waveform for 1.5 kW drive with D.C. and input inductors



Pulse N u m b e r

( 1 2 Pulse o r H i g h e r )


~ii_ 0

I I l

0 0




. . . .




frequency, kHz
Figure 9.18 Spectrum for Figure 9.17

phase-shifting transformer output I

Figure 9.19 Basic twelve-pulse rectifier arrangement



. . . . . .

= ~.






Figure 9.20 Input current waveform for 150 kW drive with 12-pulse rectifier

Chapter 9.6




< ~- 200


o O


o Figure 9.21 Spectrum for Figure 9.20


frequency, kHz




If the transformer rating matches the total drive rating reasonably closely then its inductance gives a very useful additional reduction of the higher-order harmonics. For ratings up to about 1 MW it is unusual to require pulse numbers greater than 12. The 12-pulse system is illustrated in Figure 9.19. The star and delta windings (or zig-zag windings) have a relative 30 phase shift, which translates to 180 at the fifth and seventh harmonics (as well as 17, 19, 29, 31 etc.), so that current at these harmonics cancels in the transformer. The transformer input current waveform and spectrum are shown in Figures 9.20 and 9.21, respectively.


Harmonic filters are available to attenuate specific harmonics. Most commonly they are passive circuits based on the tuning of power-factor compensation capacitors by series inductors. They can be very effective but there are potential difficulties and a specialist supplier should be consulted. Active harmonic filters are also available and avoid many of the difficulties of passive filters. These are generally rather expensive, but their increased use can be expected to lead to a price reduction in future. The simplest harmonic filter is a power-factor correction capacitor tuned by series inductance to the harmonic of interest so that the impedance is a minimum at that frequency. There are potential disadvantages to this arrangement: The filter will absorb harmonics existing on the power system, and must be rated for this duty. If inadequately rated it will trip and refuse to operate. The filter has a leading phase angle. With D.C. and similar thyristor drives this provides useful displacement-factor compensation, but with A.C. drives having negligible phase angle the leading current may have to be cancelled by a parallel inductance. The filter will have at least one resonant frequency where it magnifies harmonics. This has to be adjusted to avoid any odd harmonic frequencies. Multiple filters on a single supply system may interact and cause troublesome resonances.


The Control Techniques regenerative Unidrive has an active input stage which generates negligible harmonic current, as well as permitting the return of braking power to the supply. The input current for an active input stage contains negligible harmonic current if the supply voltage is sinusoidal. There are two side effects which must be considered: The input stage PWM frequency causes input current, which may have to be filtered. This is an optional extra in addition to the radio frequency filter. Existing voltage harmonics in the supply will cause some harmonic current to flow in to the drive. This should not be mistaken for harmonic emission.

For these reasons advice should be taken from an experienced supplier.



Electromagnetic C o m p a t i b i l i t y (EMC)

@ @ @ @ @

1 2 3 4 5


213 214 215 217 220



The purpose of this Chapter is to set out the necessary considerations for system designers and others when incorporating electronic variable-speed drives into complete machines and systems without encountering problems with

electromagnetic interference, and in compliance with relevant regulations. Of necessity, only general guidelines have been provided, but since real installations have a wide variety of detailed requirements, explanation of the underlying principles is given to allow the designer to cope with specific situations.



All electrical equipment generates some degree of electromagnetic emission as a side effect of its operation. It also has the potential to be affected by incident electromagnetic energy. Equipment using radiocommunication contains intentional emitters and sensitive receivers. The basic principle of EMC is that electromagnetic emission of electrical equipment, whether intentional or unintentional, must not exceed the immunity of associated equipment. This means that controls must be in place on both emission and immunity. Given the variety and uncertainty of effects and situations, some margin of safety must be provided between these two factors. Although all equipment exhibits some degree of emission and susceptibility, the limiting factors in most common environments tend to be related to radio equipment, with its powerful transmitters and sensitive receivers. Therefore, the majority of EMC standards are related to the requirements of radiocommunications systems. In principle, EMC covers phenomena over an unlimited range of frequencies and wavelengths. The EU EMC directive limits itself to a range of 0 to 400 GHz. This range is so wide that a perplexing number of different effects can occur, and there is a risk that all electrical phenomena become included in the scope of EMC. This is unhelpful. With current industrial electronic techniques, no significant effects occur above 2 GHz. Below 2 GHz, it is convenient to separate out effects very crudely into high-frequency effects,

which correspond to radiofrequencies beginning at about 100 kHz, and low-frequency effects. Broadly speaking, lowfrequency effects operate only by electrical conduction, whereas high-frequency effects may be induced and operate at a distance without a physical connection. Of course there is no precise dividing line between the two, and the larger the geometry of the system, the lower the frequency at which induction becomes effective. However, this division is helpful in understanding the principles.

Regulations exist throughout the world to control intentional and unintentional electromagnetic emission, in order to prevent interference with communications services. The authorities generally have the power to close down any equipment which interferes with such services. Many countries have regulations requiring consumer and other equipment to be tested or certified to meet these requirements - for example, the FCC rules in the USA and the C-tick system in Australia. The EMC directive of the European Union is unusual in requiting immunity as well as emission to be certified. It is not possible in a short section to explain all of these regulations. Most emission regulations are based on international standards produced by CISPR, and the three basic standards CISPR11, CISPR14 and CISPR22 underlie most other emission standards.


The underlying principle of all EMC regulations is that equipment should not cause interference to other equipment, and especially to communications systems. In addition, in many countries there is a requirement that equipment must be certified in some way to show that it meets specific technical standards, which are generally accepted as being sufficient to show that it is unlikely to cause interference. Equipment standards are primarily written to specify test methods and emission limits for self-contained products such as electrical consumer goods and office equipment which are basically free-standing units, even if they have the capability to interconnect with peripherals and networks. The EU EMC directive currently applies to 'a product with an intrinsic function intended for the end user' (UK EMC

regulations 1992 no. 2372). Precise interpretation of this in the realm of industrial products such as variable-speed drive modules has caused difficulty, since it is clear that some drive modules are used as virtually self-contained units, whereas others are built in to other end-user equipment. The module cannot meaningfully be tested without its associated motor, cables and other peripherals. Large fixed installations may contain numerous drives and other electronic products, and cannot practically be tested against standards which were primarily intended for compact freestanding consumer products. The European Commission published further guidelines to the application of the EMC directive in 1997, which clarify the intention for CE marking under the EMC directive. It applies to complete equipment and also to subassemblies which are intended for direct use or installation by the end user, but not to equipment which is exclusively

Chapter 10.3
for incorporation by a manufacturer into another product. These guidelines have no formal legal impact, although they are influential. The EMC directive is being revised during 2000-2001 under the SLIM initiative, and will in future incorporate the main thrust of the guidelines. Most drive manufacturers have chosen to test their products in representative arrangements, against harmonised European standards, and to apply the CE mark. It is, however, equally valid to offer a drive module without CE marking under the EMC directive (although it will normally carry the CE mark under the low-voltage directive - electrical safety), provided it is stipulated that it is intended solely for incorporation into other equipment, and that the installer takes responsibility for the EMC compliance of the end product. The purchaser should expect to be provided with comprehensive EMC data on the module, and clear guidelines on how it should be installed. The EMC compliance of the end product cannot be taken for granted even where all of the constituent parts are CE marked under the EMC directive or otherwise shown to meet relevant standards. There is always the possibility of summation of emissions, or other kinds of interaction. However, it may be possible to obtain certification for the end product through the technical construction file route on the basis of conformity of the subassemblies to specific standards. In practical terms, if the end product is a fixed installation where the legal requirement is no more than to meet the essential requirements of the EMC directive, that the equipment should neither cause nor suffer from electromagnetic interference, then a combination of compliant subassemblies is most unlikely to cause interference and therefore very likely to meet the requirements.


Emission standards work by specifying a limit curve for the emission as a function of frequency. A measuring receiver is used with a coupling unit and antenna to measure voltage or electric field. The receiver is a standardised calibrated device, which simulates a conventional radio receiver. Immunity standards are rather diverse because of the many different electromagnetic phenomena which can cause interference. The main phenomena tested for are: electrostatic discharge (human body discharge) radiofrequency field (radio transmitter) fast transient burst (electric spark effect) surge (lightning induced)

There are very many more tests available; those listed are the required tests under the CENELEC generic standards. The most important standards for drive applications are the following: IEC61800-3 and EN61800-3 - power drive systems(contains emission and immunity requirements) EN50081-2 - generic emission standard for the industrial environment EN50082-2 - Genetic immunity standard for the industrial environment (to be replaced by IEC and EN61000-6-2) The product standard IEC61800-3 applies in principle to variable-speed drive modules where they are sold as end products. There are however many cases where the drive will be incorporated into an end product which is not in itself a power drive system and is more likely to fall into the scope of the generic standards. In this case it is the generic standards which are of interest. The permitted levels are generally similar, except that IEC61800-3 defines a special environment where the low-voltage supply network is dedicated to nonresidential power users, in which case relaxed emission limits apply. This can permit useful economies in input filters.

Standards with worldwide acceptance are produced by the International Electrotechnical Commission (IEC). Standards for application under the EU EMC Directive are European harmonised standards (EN) produced by CENELEC. Every effort is made to keep these two families of standards in line, and most of them have the same number and identical technical requirements. There are some exceptions.


Most drives can be expected to meet the immunity requirements of the CENELEC generic standard EN50082-2.

Control Techniques' drives meet them without any special precautions such as screened signal wires or filters except for particularly fast-responding inputs such as data links and incremental encoder ports.



Table 10.1 Interference effects of extreme circumstances

Situation Very inductive D.C. loads such as electromagnetic brakes, without suppression, and with wiring running parallel to drive control wiring High radiofrequency field from powerful radio transmitter (e.g. adjacent to aircraft nose) Severe lightning surges due to exposed low-voltage power lines

Effect spurious drive trip when brake released or applied drive malfunction when transmitter operates drive trip or damage from overvoltage

Cure fit suppression to brake coil, or move wiring away from drive wiring

provide RF screening, or move to a location further from the transmitter antenna provide additional high-level surge suppression upstream of drive

The standard sets levels corresponding to a reasonably harsh industrial environment. However, there are some occasions where actual levels exceed the standard levels, and interference may result. Specific situations that have been encountered are given in Table 10.1.

Drives generate supply frequency harmonics in the same way as any equipment with a rectifier input stage. Supply harmonics are discussed in detail in Chapter 9. Harmonics generated by an individual drive are most unlikely to cause interference, but they are cumulative so that an installation containing a high proportion of drive loads may cause difficulties. Apart from supply harmonics, emission also occurs as a result of the switching of the power output stage over a wide range of frequencies which are harmonics of the basic switching f r e q u e n c y - that is, size times the supply frequency for a six-pulse D.C. drive, and the PWM carrier frequency for a PWM drive. This covers a range extending from 300 Hz, for D.C. drives, up to many MHz for A.C. drives. Unwanted electromagnetic coupling is relatively unusual at frequencies below about 100 kHz. Few standards set limits in that range, and interference problems are unusual.

The drive itself is not an important source of direct emission, because its dimensions are much less than a half wavelength over the relevant frequency range. There may be strong electric and magnetic fields close to the drive housing, but they diminish rapidly, by an inverse cube law, with increased distance from the drive. However, the wiring connected to the drive can be widespread and is likely to be long enough to form an effective antenna for the frequencies generated by the drive. The power output connections of a drive carry the highest level of high-frequency voltage. They can be a powerful source of electromagnetic emission. Since the cable connecting the drive to the motor is a dedicated part of the installation, its route can be controlled to avoid sensitive circuits, and it can be screened. Provided the screen is connected correctly at both ends, emission from this route is then minimised. The power input connections of a drive carry a highfrequency potential which is mainly caused by the current flowing from the drive output terminals to earth through the capacitance of the motor cable and motor windings to earth. Although the voltage level here is rather lower than at the output, control measures may be needed because these terminals are connected to the widespread mains supply network. Most commonly a radiofrequency filter of some kind is installed here. The control terminals of the drive carry some highfrequency potential because of stray capacitance coupling within the drive. This is usually of no consequence, but screening of control wires may be required for conformity with some emission standards. Figure 10.1 summarises the main emission routes for highfrequency emission. Note that the current paths are in the common mode, i.e. the current flows in the power conductors and returns through the earth. Series-mode paths are relatively unimportant in high-frequency EMC. Since the return currents in the common mode all flow in the earth wiring, earthing details are particularly important for good EMC. Much of the installation detail is involved with controlling the earth return paths and minimising mutual inductances in the earth system, which cause unwanted coupling.

The power stage of a variable-speed drive is a potentially powerful source of electromagnetic emission (noise) because of the high voltage and current which is subject to rapid switching. Thyristors are relatively slow switching devices, and this limits the extent of the emission spectrum to about 1 MHz; with IGBTs the spectrum may extend to about 50 MHz; If attention is not paid to installation guidelines then interference is likely to occur in the 100 kHz to 10 MHz range where emission is strongest. This frequency range is lower than that associated with personal computers and other IT equipment, which tend to cause direct radiated emission associated with the internal microprocessor clock and fast digital logic circuits.

C h a p t e r 10.4


direct emission power line current


earth current
,..n Vl


! !

power line potential earth current zapacl[anc;e of windings to frame

parasitic earth potential stray earth current

Figure 10.1 High-frequency emission routes



When a drive is to be installed, a cost-effective approach to EMC is to initially assess the risk of interference problems arising. This is in addition to considering any legal constraints on emission levels. Most industrial electronic instrumentation and data-processing equipment has good immunity, and can operate with drives with only modest precautions to control emission. Some specific types of equipment have been found to be susceptible to interference from drives. The list shows some product families which call for special attention: Any equipment which uses radiocommunication at frequencies up to about 30 MHz. (Note that this includes AM broadcast and short-wave radio, but not FM, TV or modern communications services which operate at much higher frequencies.) Analogue instrumentation using very low signal levels, such as thermocouples, resistance sensors, strain gauges and pH sensors. Other analogue instrumentation using higher levels (e.g. 0-10 V or 4 - 2 0 m A ) - only if very high resolution is required or cable runs are long. Wideband/fast-responding analogue circuits such as audio or video systems (most industrial control systems are intentionally slow acting and therefore less susceptible to high-frequency disturbance). Digital data links, only if the screening is impaired, or not correctly terminated, or if there are unscreened runs such as rail pick-up systems.

Proximity sensors using high-frequency techniques, such as capacitance proximity sensors.

For installations where it is known that no particularly sensitive equipment is located nearby, and where no specific emission limits are in force, some simple rules can be applied to minimise the risk of interference caused by a drive. The aspects requiring attention are as follows.

The drive supply and output cables must be segregated from cables carrying small signals. Crossing at right angles is permitted, but no significant parallel runs should be allowed and cables should not share cable trays, trunking or conduits unless they are screened and the screens correctly terminated. A practical rule of thumb has been found to be" no parallel run to exceed 1 m in length if spacing is less than 300 mm.

Control of Return Paths, Minimising Loop Areas

The power cables should include their corresponding earth wire, which should be connected at both ends. This minimises the area of the loop comprising power conductors and earth return, which is primarily responsible for highfrequency emission.



The main drive power circuit earth loop carries a high level of radiofrequency current. As well as minimising its area as described above, these earth wires should not be shared with any signal-carrying functions. There are two possible methods for minimising shared earthing problems, depending on the nature of the installation: (i)
Multiple earthing to a g r o u n d p l a n e - if the installa-

installation, then more attention must be given to the allocation and arrangement of earth connections. The concept of separate power earth and signal earth has been discredited in EMC circles recently, but is valid in widely spread installations where a good equipotential earth structure is not available. Figure 10.2 illustrates how two earthed circuits in a system may have different noise potentials. Local circuits earthed to either point will work correctly (circuits 1 and 2), but if a single circuit is earthed at both points then it will experience a noise potential which might cause disturbance (circuit 3). The solution is to nominate one earth point as the signal earth and use it as the sole reference point for shared signal circuits, as illustrated in Figure 10.3. This prevents the creation of loops for the noise current. The disadvantage is that this situation is difficult to manage in a large complex installation, and sneak paths can easily arise which cause problems and are difficult to trace. Alternatively, circuit 3 must be designed to be able to accept a high earth potential difference, for example by using optical isolation between the circuits associated with the motor and the drive.

tion comprises a large mass of metallic structure then this can be used to provide a ground plane. All circuit items requiring earth are connected immediately to the metal structure by short conductors with large crosssectional area, preferably fiat, or by direct metal-tometal assembly. Screened cables have their screens clamped directly to the structure at both ends. Safety earth connections are still provided by copper wire where required by safety regulations, but this is in addition to the EMC ground plane. (ii)
Dedicated earth points, earth segregation - if no

earthed metallic


exists throughout the


earth potential earth current --"--local earth 1 stray earth current local" earth 2

circuit 1

circuit 2
| L,,, ,,

Figure 10.2 Earth potentials and their effect on signal circuits


earth potential

earth current

signal earth
.... ,. . . . . iu ........

local earth

circuit 1


,~,circuit 3
,,, ! ~

i circuit2

J Figure 10.3 Use of single signal earth

! ' '']





There are some simple techniques, which can be used to reduce high-frequency emission from a drive at modest cost. These techniques should preferably be applied in conjunction with the basic rules given above, but they may also be useful as a retrospective cure for an interference problem. The single most effective measure that can be taken is to fit capacitors between the power input lines and earth, as illustrated in Figure 10.4. This forms a simple RFI filter, giving a reduction of typically 30 dB in overall emission into the supply network, sufficient to cure most practical problems unless exceptionally sensitive equipment is involved. Emission from the motor cable is not affected by this measure, so strict cable segregation must still be observed. The capacitors must be safety types with voltage rating suited to the supply voltage with respect to earth. Earth leakage current will be high, so a fixed earth connection must be provided. The values shown represent a compromise between effectiveness at lower frequencies and earth leakage current. Values in the range 100 nF to 2.2 laF can be used. The length of the motor cable affects emission into the power line, because of its capacitance to earth. If the motor cable length exceeds about 50 m then it is strongly recommended that these capacitors are fitted as a minimum precaution. A further measure, which reduces emission into both supply and motor circuits, is to fit a ferrite ring around the output cable power conductors, also illustrated in Figure 10.4. The ring fits around the power cores but not the earth, and is most effective if the conductors pass through the ring three times (a single pass is shown, for clarity). The ferrite should be a manganese-zinc power grade. Care must be taken to allow for the temperature rise of the ferrite, which is a function of motor cable length; the surface temperature can reach 100C.

full precautions must be observed. The drive installation guide should give these precautions in full detail for specific drives. The following outlines the essential principles: A suitable input filter must be fitted. The filter specified by the drive manufacturer should be used, and any limits on motor cable length or capacitance and on PWM switching frequency adhered to. Many filters which are not specifically designed for this application have very little benefit when used with a drive. (ii) The filter must be mounted on the same metal plate as the drive, and make direct metal-to-metal contact, to minimise stray inductance. Any paint or passivation coating must be removed to ensure contact. A back plate of galvanised steel, or other corrosion-resistant bare metal, is strongly recommended. (iii) The motor cable must be screened. A copper braid screen with 100 per cent coverage works best, but steel wire armour is also very effective, and steel braid is adequate. (iv) The motor cable screen must be terminated to the drive heat sink or mounting plate, and to the motor frame, by a very low inductance arrangement. A gland giving 360 contact is ideal, a clamp is also effective and a very short pigtail is usually tolerable but the drive instructions must be adhered to. (v) The input connections to the filter must be segregated from the drive itself, the motor cable and any other power connections to the drive. (vi) Interruptions to the motor cable should be avoided if possible. If they are unavoidable then the screen connections should be made with glands or clamps to an earthed metal plate or bar to give a minimum inductance between screens. The unscreened wires should be kept as short as possible, and run close to the earthed plate. Figure 10.5 illustrates an example where an isolator switch has been incorporated. With some drives, the control wiring needs to be screened with the screen clamped to the heat sink or back plate. The installation instructions should be adhered to in this respect. Omitting this is unlikely to cause interference problems, but may cause standard limits for radiated emission to be exceeded. (i)

If there is known sensitive equipment in the vicinity of the drive or its connections (see list in Chapter 10.4), or if it is necessary to meet specific emission standards, then


L1 L2

ferrite rin-" ..... u

/ f

3 turns (1 shown for clarity)


l i T

T _1- T
a x 4 7 0 nF



Figure 10.4 Some low-cost emission reduction measures


THEORETICAL BACKGROUND" Emission M o d e s isolator

from the drive

to the motor

coupling bar or plate


(if required)

Figure 10.5 Managing interruptions to motor cable


Although the digital control circuits, switch-mode power supplies and other fast switching circuits in a modern digital drive can all contribute to radiofrequency emission, their suppression is a matter for the drive designer and suitable internal measures can keep such emission under control. It is the main power stage, especially the inverter of a PWM drive, which is an exceptionally strong source of emission because the fast changing PWM output is connected directly to the external environment (i.e. the motor and motor cable). This is also the reason for the installation details having a major effect on the overall EMC behaviour. Figure 10.6 shows the main circuit elements of an A.C. inverter variable-speed drive. The output PWM waveform has fast changing pulse edges with typical rise times of the order of 50-100ns, containing significant energy up to about 30 MHz. This voltage is present both between output phases and also as a common-mode voltage between phases and earth. It is the common-mode voltage which is primarily responsible for emission effects, because it results in high-frequency current flowing to earth through the stray capacitances of the motor windings to the motor flame, and the motor cable power cores to the earth core and/or screen. High-frequency current causes unexpected voltage drops in wiring because of the wiring self inductance. The significance of this can be illustrated by a simple example. A 1 m length of wire has a typical inductance of about 0.8 laH. The output current from a drive to charge the stray capacitance of a motor winding would be typically 2 A peak with a rise time of 100 ns. This current would cause a voltage pulse

of 16 V with duration lOOns in the 1 m of wire. Whether this causes interference with associated circuits depends on their design, but certainly a 16 V, 100 ns pulse is sufficient to cause a serious error in a digital circuit or a fast acting analogue one. Figure 10.1 shows the main emission paths. Because of its high voltage, the motor cable is the main potential source of emission and will be an effective transmitting aerial at a frequency where its length is an odd number of quarter wavelengths. For example, a 20 m cable will be particularly effective at about 3.75MHz and also at 11.25MHz and 18.75 MHz. This will be modified somewhat by the presence of the motor and by the distance of the cable from surrounding earthed objects. In order to prevent this emission, the cable is usually screened. Figure 10.1 also shows how the high-frequency voltage in the motor and cable causes current to flow into the earth, because of their capacitance. The capacitance of a motor winding to its flame may be in the range 1 nF to 100nF, depending on its rating, and the capacitance from the cable power cores to earth is generally between 100pF and 500pF per metre. These values are insignificant in normal sinusoidal supply applications, but cause significant current pulses at the edges of the PWM voltage wave. The current returns through a variety of paths, which are difficult to control. In particular, current may find its way from the motor flame back to the supply through any part of the machinery, and if it passes through earth wires in sensitive measuring circuits it may disturb them. Also, a major return route to the drive is through the supply wiring, so any equipment sharing the supply may be disturbed.

Chapter 10.5


from supply L1 L2 L3

to motor U v w

! I I I i




I ! I I I


rise time

Figure 10.6 Main elements of an A.C. inverter drive

return current circulates in filter

no emission from screened cable screened cable / ~/ _/ "N~

power line

current minimised

current returns alongscr~~

/.--.,J J _. L

stray earth current minimised

Figure 10.7 The effect of an RFI filter and screened motor cable

Figure 10.7 shows the effect of using a screened motor cable and an input filter. Fields emitted from the motor cable are suppressed by the screen. It is essential that both ends of the screen are correctly connected to earth at the motor and the drive, in order that the magnetic field cancellation property of the cable gives its benefit. The screened cable also minimises the earth current flowing from the motor frame into the machinery structure, because of its mutual inductance effect. This subject is generally not well understood outside the EMC profession, and the reader is referred to Chapter 10.6 for a fuller explanation. The input filter provides a low-impedance path from the earth to the drive input lines, so that the high-frequency current returning from the motor cable screen has an easy local return route and does not flow into the power network. The primary role of the filter is to suppress common-mode high-frequency emission from the drive. There is also some series-mode emission because of the nonzero impedance of

the D.C. smoothing capacitor in the drive. The filter provides some series-mode attenuation to control this.


Figure 10.8 shows the circuit of a typical input filter. The capacitors between lines provide the series-mode attenuation, and the capacitors to earth and the inductance provide the common-mode attenuation. The inductance is constructed as a common-mode component, which is not magnetised by the main power current, therefore minimising its physical dimensions. It uses a high permeability core, which can accept only a very limited unbalance (commonmode) current. Filters for drives are carefully optimised for the application. The drive presents an exceptionally low impedance source to the filter, which means that conventional generalpurpose filters may have little benefit. The usual method of


THEORETICALBACKGROUND"Principles of Input Filters



L2 supply L3

i .o

A w



1 TT T


Figure 10.8 Basic input filter

specifying a filter is in terms of its insertion loss in a test set up with 50 f~ source and load impedance. An alternative test attempts to be more realistic by using 0.1 f~ source and 100 f~ load. Neither of these tests correctly represents a drive application, and neither can be used as any more than a very rough guide to the suitability of a filter.


The screening capability of screened cable is generally measured by the parameter ZT, the transfer impedance per unit length. In an ideal cable, any current flowing in the internal circuit produces no voltage between the ends of the cable screen, and conversely current flowing in the screen from an external source produces no voltage in the inner circuit. These two aspects minimise the emission from the cable and the immunity of inner signal circuits to external disturbance, respectively. In practice, the resistance of the screen, its imperfect coverage and other details cause a departure from the ideal and a nonzero value of ZT. The transfer impedance is not however the only factor involved. The cable exhibits strong internal resonances, which cause high currents to flow internally. The current is limited by the natural damping caused by electrical losses in the cable. Steel sheaths have a higher resistance and therefore give better damping than copper sheaths. Steel gives an inferior transfer impedance to copper, but the two factors largely cancel so that a steel wire armoured cable gives no greater emission with a drive than a good quality copper braided screened cable.

surrounds a three-phase set then the magnetic field is only caused by the common-mode current, and saturation is avoided. The manganese-zinc ferrite exhibits high loss in the 1-10 MHz frequency range where motor cable resonance occurs, and this gives useful damping of the resonance and a substantial reduction in the peak current. The loss in the ferrite does cause a temperature rise, and with long motor cables the temperature of the ferrite rises until its losses stabilise, close to the Curie temperature.


Because of the low source impedance presented by the drive, suitable filters generally have unusually high values of capacitance between lines and earth. This results in a leakage current to earth exceeding the 3.5 mA which is generally accepted as permissible for equipment which derives its safety earth through a flexible connection and/or plug/ socket. Most filters require the provision of a permanent fixed earth connection with sufficient dimensions to make the risk of fracture negligible. Alternative versions with low leakage current may be available, which will have more severe restrictions on the permissible motor cable length.


With long motor cables the common-mode current in the filter rises to a level where the high-permeability core of the filter inductance becomes magnetically saturated. The filter then becomes largely ineffective. Filters for drive applications therefore have limits on motor cable length. The capacitance of the cable determines the additional current loading on the drive and the filter. Screened cables with an insulating jacket between the inner cores and the screen present a tolerable capacitance. Some cables have the screen directly wrapped around the inner cores, which causes abnormally high capacitance and therefore reduces the permissible cable length. This also applies to mineral insulated copper-clad cables.


The use of a ferrite ring as an output suppressor was introduced in Chapter 10.4. The ferrite ring introduces impedance at radiofrequencies into the circuit which it surrounds, thereby reducing the current. Because of its high permeability the ring will not work if it surrounds a conductor carrying power current, due to magnetic saturation, but if it





The subject of signal circuit cable screening is often misunderstood and it is quite common for such circuits to be incorrectly installed. This applies to critical signal circuits for drives, such as analogue speed references and position feedback encoders, and also to circuits in other equipment in the same installation as the drive. This section outlines the principles, in order to assist readers in avoiding and troubleshooting EMC problems in complete installations.

E (inner and outer)


magnetic field Figure 10.9 Magnetic-field induction in screened cable


Correctly used, a cable screen provides protection against both electric and magnetic fields, i.e. against disturbance from both induced current and induced voltage. Electric-field screening is relatively easy to understand. The screen forms an equipotential surface connected to earth, which drains away incident charge and prevents current from being induced in the inner conductor. Magnetic-field screening is more subtle. An incident alternating magnetic field, which corresponds to a potential difference between the cable ends, causes EMF to be induced in both the screen and the inner conductor. Because the screen totally surrounds the inner conductor, any magnetic field linking the screen also links the inner conductor, so an identical EMF is induced in the inner. The voltage differential between the inner and the screen is then zero. This is illustrated in Figure 10.9. In order for this benefit to be realised, it is essential that the screen be connected at both ends. Although high-frequency engineers routinely observe this practice, it is common in industrial control applications for the screen to be left unconnected at one end. The reason for this is to prevent the screen from creating an earth loop, or an altemative earth path for power-frequency current. The problem of the earth loop is specifically a lowfrequency effect. If the impedance of the cable screen is predominantly resistance, as is the case at low frequencies, then any unwanted current flowing in the screen causes a voltage drop which appears in series with the wanted signal. This is illustrated in Figure 10.10. At higher frequencies the cable screen impedance is predominantly inductive. Then the mutual inductance effect takes over from resistance, and the voltage induced in the internal circuit falls. Further, the skin effect in the screen causes the extemal current to flow in the outer surface so that the mutual resistance between inner and outer circuits falls. The net result is that at high frequencies the cable screen is highly effective. A cut-off frequency is defined at the point where the injected voltage is 3 dB less than at D.C., and is typically in the order of 1-10 kHz. Where disturbing

E + IR (outer) E (inner)


Figure 10.10 Resistance coupling in screened cable



OV(1) optional earth


0 V (2)

~, -....


~ , optional earth

Figure 10.11 Correct screened cable connection for highfrequency screening effect

frequencies exceed the cable cut-off frequency, the screen should be connected at both ends.


The conclusion of this is that for all but low-frequency interference, the screen should be used as the return path for data, as shown in Figure 10.11. Whether the screen is connected to earth at each end, or to the equipment metalwork, is less important than that it be connected to the circuit common terminals. The recommendations of the equipment manufacturer should be followed. It is usual to clamp the screen to a metallic structural part because this gives the least parasitic common inductance in the connection. A pigtail causes a loss of screening benefit, but a short-tail (up to 20 ram) may be acceptable for drive applications where screening is not critical.


ADDITIONAL GUIDANCE ON CABLE SCREENING" C a b l e S c r e e n C o n n e c t i o n s

The screened cable should ideally not be interrupted throughout its run. If intermediate terminal arrangements are included with pigtails for the screen connections, every pigtail will contribute an additional injection of electrical noise into the signal circuit. If interruptions are inevitable, either a suitable connector with surrounding screening shell should be used or a low-inductance bar or plate should be used for the screen connection as in Figure 10.5. Suitable hardware is available from suppliers of terminal blocks. Low-frequency interference associated with earth loops is not important for digital data networks, digital encoder signals or similar arrangements using large, coarsely quantised signals. It is an issue with analogue circuits if the bandwidth is wide enough for errors to be significant at the relevant frequencies, which are primarily the 50/60 Hz power-line frequency. Many industrial control systems have much lower bandwidths than this and are not affected by power frequency disturbance. Servodrives however do respond at power-line frequency, and can suffer from noise and vibration as a result of power frequency pick up. The cable screen should not be used as the signal return conductor in this case. The correct solution for wideband systems is to use a differential input. Analogue differential inputs give very good rejection of moderate levels of common-mode voltage at power-line frequency. This rejection falls off with increasing frequency, but then the screening effect of the cable takes over. The combination of differential connection and correct cable screening gives good immunity over the entire frequency range. A typical arrangement is illustrated in Figure 10.12. There are electrical safety issues associated with earthing decisions. A galvanically isolated port with the screen connected only to the isolated common rail prevents lowfrequency circulating current, but carries the risk that a fault elsewhere might make it electrically live and a hazard to maintenance staff. Cable screens should be earthed in at least one place for every disconnectable length, to prevent a length becoming isolated and live. This approach is used in the Interbus industrial data network, where each link is earthed at one end and isolated at the other. Earthing at both ends carries the risk that an electrical fault might cause excessive power current to flow in the cable screen and overheat the cable. This is only a realistic risk in large-scale plant where earth impedances limit power-fault current levels. The correct solution is to provide a parallel power earth cable rated for the prospective fault current. An alternative is to provide galvanic isolation, although this

earth 1 _.


+ beed

earth 2

Figure 10.13 Use of capacitor for high-frequency earthing while blocking power-fault current



=OV 0 V to other circuits

I --




optional earth


(note-multiple earthing may cause disturbance at power frequency)

+ signal I I I I

! I ! ! !

optional earth

, '

, --

optional earth

D /


differential . amplifier input

optional earth

I --"

, --"

optional earth

0 V (1) optional earth



0 V (2) optional earth


i , -I I

. . . .

Figure 10.14 Preferred connection methods a l o w - s p e e d d i g i t a l c i r c u i t - no special precautions, o p e n w i r i n g b low-bandwidth l o w - p r e c i s i o n analogue c i r c u i t - screened
c wide bandwidth and/or high-precision analogue c i r c u i t - screened, differential d w i d e b a n d w i d t h d i g i t a l data c i r c u i t - screened, differential

Figure 10.12 Correct screened cable connection for high-frequency screening and low-frequency interference rejection

Chapter 10.6


carries the risk of a transiently high touch potential at the isolated end during a fault. Some galvanically isolated inputs include a capacitor to earth, which provides a high-frequency return but blocks power-frequency fault current. This is actually a requirement of certain bus systems. In principle such a capacitor should not be necessary, but it may be required to ensure immunity of the isolated input to very fast transients, or to suppress radiofrequency emission from microprocessors

etc. The capacitor must be rated at the mains voltage. It is usual to provide a parallel bleed resistor to prevent accumulated static charge. Figure 10.13 illustrates this capacitor arrangement.


Figure 10.14 summarises the recommended connection methods for several cases.

C H A P T E R 11

Systems Design

@ @ @ @ @ @ @ @ @ @ @
6 4


228 228 228 233 234 236 243 244

8 9

246 250

10 11





The successful integration of electronic variable-speed drives into a system depends upon knowledge of a number of key characteristics of the application and the site where the system will be used. This chapter contains basic considerations, which can help a user or system designer to avoid

some of the most common problems. First, advice is given on selecting the most appropriate drive type. Then guidance is provided on basic requirements of the installation to help ensure trouble-free operation.


A design matrix for electronic variable-speed drives is given in Table 11.1.


Electric motors of most kinds have the ability to behave as generators, i.e. they will deliver current if the shaft is rotated (and excitation is present), thus it is possible to obtain a braking effect by delivering the generated current to a load. In many applications, the use of the motor for braking enables separate braking apparatus to be dispensed with altogether. By controlling the generated current, the braking effort can be varied; with an appropriate choice of motor and control strategy, precisely profiled deceleration curves can be achieved. Limitations upon braking performance are imposed by the type of motor, type of electronic drive and selected means of absorbing the braking energy; for the latter there are two choices - a fixed resistor or a back feed into the supply system. These two alternatives are known, respectively, as dynamic and regenerative braking. Regenerative braking always means the return of power to the supply system; dynamic braking involves the use of a resistor to absorb the energy, and for that reason the term resistive braking is preferable and is used in this book.

Useful economic advantages may be gained by regeneration into the main power supply. Also, given that braking loads are large enough and that the duty cycle is appropriate, resistive braking may be exploited for energy recovery in the form of heat. From the point of view of the motor, there is no fundamental difference between resistive and regenerative braking: the motor delivers electrical energy, which is absorbed by one means or the other. But there are important implications for the configuration of the electronic drive equipment. Certain types only can accommodate full braking facility; others have inherent limitations, which are noted as appropriate in the following discussion. General capabilities are pictorially summarised in Figure 11.1.


The principles of single (1Q) and four-quadrant (4Q) D.C. motor operation have been discussed earlier; however, it is useful to refer to that explanation as a preliminary to the discussion here. As indicated in Figure 11.1, a 1Q drive has limited braking capability; for braking both forward and in reverse a 4Q

Chapter 11.3



o o~.~

"-' Ob


n~ o

0 r.~ ~

o" o

xl ~

o o o c~
O "~ n~ O O

o o
0 0



o o~-~





~ o



o o

o o


o 0~.~



.~ o


o o


~ ",1

~ xl ~ o ~



t~ E


~ ~ ~.~

~oo ~



0 '~ 0

~ .~ o .~

T ~ g

~ o~


~" ~ ~


v ~.
~ ~' ~ ~ ~

o~.~ ~.~ o ~ ~ ~ ~ ~ .~ ~ . ~ ~ ~ ~

O ~0

0~ O O O






yes reverse l I I

resistive regenerative



0 0


resistive regenerative
yes yes

w w

t I rN,-T I@ G
resistive regenerative
yes i



Figure 11.2 Typical arrangement for dynamic braking


resistive regenerative

yes I

,~ dynamicbraking ' i ~ here I---1

"0 (D d.) 3L

Figure 11.1 Summary of the main A.C and D.C drive configurations and braking capability

drive is essential. Nevertheless, the 1Q drive offers useful braking performance for certain applications. As regenerative braking depends on the return of power from the motor to the power supply, it cannot operate if the mains supply fails due to a blown fuse or a power cut. Resistive braking of 4Q drives is sometimes encountered as a fail-safe means of stopping the motor and its load, and as the only means of braking single-ended drives. A rotating motor can act as a generator and, if it delivers current, torque is produced which opposes the rotation of the machine. The principle of resistive braking is to connect a resistive load across the motor. Current flowing through the resistance creates braking torque, and the kinetic energy extracted from the rotating machine is dissipated as heat in the braking resistor. Figure 11.2 shows a typical arrangement for resistive braking of a D.C. motor. A single contactor is used for both the armature and the dynamic braking resistor (DBR), ensuring correct sequencing and fail-safe operation. The rotating motor generates a voltage, or counter electromotive force (CEMF) as it is sometimes called, the value of which is proportional to the speed of rotation so long as field flux remains constant. It is this voltage, applied to the braking resistance, which causes (braking) current to flow. Therefore the braking current (and hence braking torque) is proportional to speed or, expressed in a different way, the braking torque reduces as the motor decelerates, giving an exponential deceleration curve as shown in Figure 11.3.

actual stopping time is reduced due to friction

Figure 11.3 Speed versus time during dynamic braking

The motor CEMF, Eo, is calculated as follows"

Ea Va-laRa

where V a - rated armature voltage, I a - rated armature current and R a - armature circuit resistance. The total resistance in the motor circuit during dynamic braking is given by:
Rb -- RDBR q- Ra

Therefore, the initial or peak braking current (typically chosen to be 150 or 200 per cent of IA) can be calculated using the following equation, derived from Ohm's law:
Ip -- g a / R b -- ( Va - IaRa ) / (RDBR + Ra)


RDBR -- ((Va - IaRa)//Ip) - Ra

Braking resistors of small ratings are often of wire-wound construction on a ceramic former, and those of higher ratings are usually fabricated steel stampings, coiled or convoluted strip, expanded metal, cast iron, or liquid, all these giving large surface areas for dissipation of the heat produced during braking.




Since the kinetic energy of the motor and its load is converted into heat by the braking resistor, it is important to rate the resistor correctly for the duty it is expected to perform, taking account of load inertia and the number of stops per hour.

The nearest standard value is 12 ohms. In this case the value is slightly above the calculated minimum value because there may be commutation problems if the current is allowed to grow too large and there is no switching device to control the average current. The resistor overload factor, which relates overload time, cooling time to an overload factor is provided by power resistor manufacturers. Figure 11.4 shows typical characteristics. If this type of information is not available, and quite often only the 20 or 30-minute cooling time information is published, then either the resistor manufacturer must be contacted, or a generous safety factor must be allowed. The resistor itself will not be a significant factor in the overall system cost. For an overload time of ten seconds and a cooling time of 30 minutes the overload factor derived from Figure 11.4 is 6.0. resistor continuous power rating - regenerated power/overload factor = 15/6 - 2.5 kW

Example Calculation of a Brake Resistor of a D.C. Motor


7.5 kW motor, efficiency 85 per cent, armature voltage 420 V, armature current 21 A maximum armature current during braking = twice fullload current brake operation time, ten seconds repetition time is intermittent, therefore allow 30 minutes assume motor is working into the field-weakening range


The kinetic energy contained in the rotating motor and load is 0.5Jcd 2 where J = moment of inertia and a; - angular velocity. As most of the energy is concentrated above base speed, the resistor does most of its work above this speed. As this is the constant-power region, for most of the braking cycle the braking resistor also operates with constant power. With a twice full-load braking current the peak motor regenerated power is: 7.52-15kW for a short period of time. As the motor is 85 per cent efficient, its losses will be: (15 kW/0.85 - 15 kW) - 2.65 kW Therefore, the peak braking power at the shaft is: 15 + 2.65 - 17.65 kW This means that it should be possible to decelerate the motor and load faster than it is possible to accelerate it, and this without taking into account any load losses.


The principle of four-quadrant operation of an A.C. drive has been discussed earlier. It is important here to consider dynamic/resistor braking only. The basic arrangement is shown in Figure 11.5. The key difference between the A.C. braking circuit and the D.C. circuit of Figure 11.1 is the inclusion of a transistor in series with the braking resistor. This braking transistor is capable of controlling the current into the resistor. In fact, the transistor is controlled to limit the voltage in the D.C. link as, when the transistor is switched off, energy can still be fed into the D.C. link and results in increasing the D.C. link (capacitor) voltage. The upper switching level for the brake transistor is typically chosen to lie between the bus level when the mains supply is


15 sec -30 sec cooling time 1 min -5 min ; 30 min

I i


For a resistor: power P or:

1~ v=/P
-- V 2/ R



-o 5
0 > 0

\\\ \\

\'\\, \ \ .

In the example, V is the D.C. armature voltage - 420 V. P is the calculated maximum power r e g e n e r a t e d - 1 5 kW, so that: R4202/(15 103 ) - 11.76f~

-, .,
500 1000


10 20 50 100200 overload time, s

Figure 11.4 Typical power-resistor overload factors


DYNAMIC/RESISTIVEBRAKING:A.C. Regeneration and Braking





Figure 11.5 Typical PWM A.C. drive with dynamic braking resistor Rb
on top tolerance, and the maximum safe continuous operating voltage of the devices in the inverter bridge. It is common that this switching level is determined during operation or initialisation through monitoring of the mains supply (directly or indirectly via the D.C. link voltage). The hysteresis between the upper and lower switching levels will be about 20 or 30 V. In the example, V is the value of the D.C. bus voltage when regeneration is taking place, i.e. 750 V. P is the calculated maximum power regenerated back to the D.C. bus, 11.25 kW, so that: R(750)2/(11.25 x 103 ) - 5 0 ~

This is the maximum value that the braking resistor can have. If the resistor is given a slightly lower value then the braking transistor controls the power by switching on and off. For this example the next lowest standard value would be chosen, which is 47 ohms.

Example Calculation of a Brake Resistor of a PWM A.C. Induction Motor Drive System
7.5 kW motor, efficiency 83 per cent, 400 V 400 V, three-phase PWM inverter with a dynamic brake operating voltage of 750 V and 150 per cent overload rating brake operation time five seconds repetition time: one minute minimum maximum speed = twice base speed


As the braking resistor is normal to use a resistor of than the calculated average D.C. bus and to rely on the used only intermittently, it is lower continuous power rating power regenerated back to the overload rating of the resistor.


With a 150 per cent overload rating the 7.5 kW inverter can absorb: 7.5 1.5=11.25kW for a short period of time. As the motor is 83 per cent efficient, its losses will be 2.3 kW. In fact, with the increase in D.C. bus voltage from a nominal 565 V to 750 V, and with more than full load current, the motor losses will probably be even more than 2.3 kW. However, the actual braking power available at the motor shaft is:
11.25 + 2.3 = 13.55 k W

Refer again to the power resistor overload factor characteristic, a typical example of which is given in Figure 11.4. In this example, the five-second overload time/one-minute cooling time curve applies, and this gives an overload factor of 3.8. The power rating of the 47 ohm resistor is given by: resistor continuous power rating = 11.25/3.8 = 2.96 kW

The resistor value should not be chosen to be much smaller than the value calculated from R = vZ/P, otherwise the braking transistor has to handle a high peak current and has to switch more frequently; both factors cause high power dissipation in the transistor. Also, there is a high peak-power dissipation in the resistor, which it may not be able to accept. A resistor of the rating calculated will probably be a bank made up of individual resistors, and if there is to be no additional derating the individual resistors should not be mounted too close to each other. Typically, a spacing of at least twice their diameter is sufficient when the resistors are mounted vertically in a free flow of air. The resistor data sheet should be consulted for specific recommendations. Pulse-rated resistors should be used. These are often wound from corrugated tape or resistance strip on a helical former for higher powers. The rating of a resistor may be increased by forced cooling.

This means that it should be possible to decelerate the motor and load faster than it is possible to accelerate it, and this without taking into account any load losses.

For a resistor, as for the D.C. example:
R = v:/P

Chapter 11.4



Fuses are the most common, but perhaps one of the least understood forms of, protective device. In a variable-speed drive installation, fuses are principally used to protect the mains-to-converter cable and the converter-to-motor cable under fault conditions only. Fuses may also be used to protect components within the drive or, in the event of a failure within the drive converter, fusing may be used to limit the energy fed into the fault. Fuses are not required to protect for overload conditions as this is undertaken by the drive converter protection. It is important to retain focus on the specific purpose of each fuse in the circuit if the correct fuse type and the correct positioning of the fuse is to be assured. It is also critical to consider the protection circuit as a whole to ensure that the correct discrimination between protective devices exists. For example, when two fuses are connected in series, and the upstream fuse is feeding the downstream fuse and other loads, then it is important that, in the event of a fault in the circuit of the downstream fuse, the downstream fuse operates rather than the upstream fuse. In this way the other loads are unaffected. Discrimination is achieved by ensuring that the total I2t let through by the downstream fuse is less than the pre-arcing IZt of the upstream fuse, and the time/current characteristic of the upstream fuse should lie to the right of that of the downstream fuse. It is not practicable to provide comprehensive details of fuse technology here. For this reference should be made to specialised books (See Appendix E). We will consider fusing with reference to the specific functions within a drive system.

Table 11.2 Fuse recommendations for Control Techniques, Unidrive, type gG HRC industrial fuses to IEC 269, or RK1 600 V A. C.

Drive rating Power (kW) 0.75 1.1 1.5 2.2 2 5.5 7.5 11 15

Recommended A . C . supply & motor cable mm2 1.5 2.5 2.5 2.5 2.5 2.5 4 4 6

Recommended fuse rating (A)

Rated current (A) 2.1 2.8 3.8 5.6 9.5 12 16 25 34 40 46 60 70 96 124 156 180 202 300

AWG 16 14 14 14 14 14 10 10 8

22 30 37 45 55 75 90 110 160

10 16 25 35 35 50 70 95 95

6 4 4 2 2 2/0 2/0 3/0 3/0

6 10 10 10 16 16 20 35 40 50 60 70 80 100 125 160 200 250 450


In an A.C. drive the failure of the input diode bridge is associated with either high transient or continuous supply voltages or with a primary failure within the drive. Fuse protection of input components within an A.C. drive is very difficult to achieve and is rarely recommended. A D.C. drive is very different and fuse protection of internal converter components against a range of internal and external faults is possible and advantageous. The positioning of fuses within a D.C. drive system is of interest: a b c in the A.C. supply lines in series with the thyristors within the bridge in the D.C. motor armature lines


As the drive controller regulates current flowing in the system, the fusing needs to be designed to cater for a catastrophic failure within the drive or, more likely, a short circuit between cables. High rupturing capacity (HRC) fuses act as cleating devices for sustained high currents and are consequently well suited to this type of duty, and are commonly recommended by most drive manufacturers. It is important to note that drive products listed by UL (Underwriters Laboratories Inc.) are often tested together with their recommended fuses. Protection against fire and unacceptable earth currents must be proven under all conceivable failure modes. Drive manufacturers should make clear recommendations relating to fuse selection. Typical fuse recommendations for three-phase A.C. drives are given in Table 11.2. It can be seen that the fuse rating is not simply related to the cable rating but is influenced also by the drive rating.

For unity D.C. current, the relative r.m.s, currents are a - 0.82, b = 0.58 and c = 1.0. Fuse and thyristor coordination can be achieved by ensuring that the let-through IZt of the fuse in each of the above referenced positions is less than the device IZt. Semiconductor fuses must be used in all positions to afford protection. Fuses (b) within the converter bridge, have the advantage of protecting D.C./load faults as well as removing failed devices, which have failed to block. The disadvantage is that the location of these fuses in the power circuit is difficult to facilitate in a compact design which allows for easy fuse replacement.


FUSING: Protection of Drive Components

Table 11.3 Control Techniques, Mentor II, type: semiconductor; for 400 V supply rated 500 V D.C., for 480 V supply rated 700 V D.C.
Drive Recommended rated A.C. supply and current (A) D.C. motor cable mm2 25 45 75 105 155 210 350 420 550 700 825 900 1200 1850 4 6 25 35 50 95 150 185 300 2 x 185 2 x 240 2 x 240 2 x 400 3 x 400 AWG 10 6 2 1/0 3/0 300MCM 35 60 100 125 175 250 400 500 700 900 1000 1200 2 x 700 2 x 1200 40 70 125 175 250 300 550 700 900 1000 1200 2 x 700 2 x 900 2 x 1000 Recommended Recommended A.C. fuse rating D.C. fuse rating (A) (A)

It is more common, therefore, for drive manufacturers to specify A.C. line and, in the case of four-quadrant drives where the load can feed energy into a converter fault, D.C. line fuses. Fuses in the A.C. line remove the entire arm from the circuit if one thyristor fails to block voltage. If both seriesconnected devices in an arm fail, then the motor armature is short circuited and the D.C. line fuses are relied upon to clear the fault current. Fuses in the D.C. circuit must be rated for that duty, D.C. fault current in an inductive circuit being difficult to clear. It is important to use fuses in the positive as well as the negative line as they then share the duty, although it is unwise to assume a better than 80/20 sharing of voltage. Typical fuse recommendations for three-phase D.C. drives are given in Table 11.3. These guidelines are typical data only. Local regulations must of course be complied with in full.


Whether operating directly on the mains supply or under the control of a variable-speed drive, a motor can be thermally overloaded by any of the following: 1 2 3 increased ambient temperature obstruction of the coolant flow increased losses in the machine

current in the supply line is no greater than the rated value. The losses in the rotor are however considerably higher than during normal three-phase operation. The change in the balance of losses in larger motors needs to be considered and advice is available from motor manufacturers on this matter. Variations in air-gap flux (supply voltage/frequency variations in fixed-speed motors)- variations in air-gap flux, which are a result of voltage or frequency variations, can, over prolonged periods, increase both iron and copper losses in A.C. motors. Again, the impact is dependent upon the distribution of these losses within the motor. e Starting into a mechanically-locked load - if a motor is started into a mechanically-locked load, the starting current of the motor (typically of the order of six-times rated current) would flow continuously and the motor would reach its temperature limit, typically, within a few seconds. f Stalling the motor - if a motor is stalled, the current rises to approximately the starting current, and again the motor would reach its temperature limit, typically, within a few seconds. g Control faults - a control fault, such as a star-delta changeover switch failing to operate leaving a fixedspeed motor in star configuration, would clearly cause overloads. d

Considering the causes of increased losses in the machine further, and for the moment only looking at fixed-speed (direct-on-line) A.C. induction motors: a Overloading in continuous operation-this condition is largely self explanatory, and is often associated with changes to the working conditions. This may be due to the user unwittingly overloading the machine, or a blockage or failure in the system increasing the load. Excessive starting/braking duty c y c l e - a s a. Single phasing on A.C. machines-this condition is invariably due to a break in the wire, unreliable contacts or one of the supply fuses operating. For induction motors of smaller ratings, it is important to note that losses are generally lower than during normal three-phase operation providing the current in the supply line is no greater than the rated value. Torque is of course reduced. For larger induction motors, stator winding losses are lower with single-phase operation providing that the

b c

Chapter 11.5


It is interesting to review these conditions when the motor is fed from an electrical variable-speed drive controller, Table 11.4. Additional motor losses associated with the nonideal supply waveforms of variable-speed drives have been discussed earlier.

as the converter itself. They also contain inputs to which motor-embedded thermistors may be connected. This direct monitoring of motor temperature provides back-up protection, and is to be recommended where possible.

What Cannot be Used

It is important to recognise that the nonsinusoidal waveforms, and variable frequency, associated with variablespeed drives, invalidate the basis for protection afforded by most electronic overload relays. The use of such devices on the mains supply of a drive is also invalid. Any such devices should only be used after clear and detailed discussion and approval with the manufacturer.


M o d e m electrical variable-speed drive controllers contain comprehensive thermal protection for the motor as well

Table 11.4 Motor overtemperature protection Fault

Overloading in continuous operation Excessive starting/ braking duty cycle

D.C. drive


open-loop induction motor drive

A.C. closed-loop induction/ PM servomotor drive

as the drive is inherently monitoring and controlling the current fed to the motor, the protection of the motor against constant overload is invariably built into the design of the drive starting in a variable-speed drive does not result in excess losses in the motor; the basis for limiting starts necessary for fixed-speed motors does not therefore apply there may be a need to limit starts for a variable-speed A.C. drive controller if the d.c. link is charged via a resistor, such a resistor will have a maximum starting duty, which will be specified by the manufacturer, if frequent start/stop duty is required then electronic control only should be used i.e. do not remove the mains supply from the drive, in this way the starting resistor, if present, will not be stressed not applicable unless a drive-motor cable becomes detached or broken, it is unusual for a motor phase supply to be lost; some drives are able to detect imbalance in the motor phase current and offer protection it is not recommended to use contactors in between the drive and motor as a means of control; they may be used as part of an isolation or bypass strategy

Single phasing on A.C. machines

Variations in air-gap air-gap flux is controlled a flux controller in the drive many open loop-drives do not flux (supply voltage/frequency by field current control controls air-gap flux; incorporate a flux controller; variations in fixed-speed motors) by means of a field controller the actual flux is estimated and flux variations in supply voltage have no impact, other than control is dependent upon defining the maximum the quality of the flux estimation; as the user is primarily interested in operating speed at which full flux can be torque quality, drive manufacturers provided usually provide data on torque linearity rather than flux Starting into a mechanically-locked load as the drive is inherently monitoring and controlling the current fed to the motor, the protection of the motor against overload at start is invariably built into the design of the drive; closed-loop drives detect and many open-loop drives estimate speed and can therefore protect against a locked rotor condition the thermal motor model within drive controllers can simulate hot spots in the motor, but does not always take account of reduced cooling effect, it is possible however for drives to detect a stall condition; in closed-loop drives this is automatic; for open loop this can be seen as a due to high current/low speed condition control faults of the type described above (e.g. star/delta changeover failures) are either not possible or inherently monitored and protected against in variable-speed drives

Stalling the motor

Control faults




When an A.C. drive and machine are connected via long cables the performance of the drive system is affected by the cable resistance and the current pulses due to high rates of change of voltage applied to the cable when the inverter IGBTs switch. This rate of change of voltage results in pulses of current to charge the capacitance of the cable. The effects of operation with long cables are described below, and are different for different forms of control/motor used. The following descriptions, and performance figures, relate to the Control Techniques Unidrive A.C. drive, operating in the following modes: closed l o o p - induction motor closed l o o p - PM servomotor open loop-induction motor

Cable length-the length of the pulses increases with cable length. The drive overcurrent protection system trips the drive if the current from the drive exceeds the trip level. Some filtering is provided in the trip system, but it cannot remove all of the cable-charging current effects and still provide adequate drive protection. As the current pulses increase in magnitude and length they have more effect on the overcurrent protection system. Therefore, as the D.C. link voltage and cable length are increased the likelihood of unwanted overcurrent trips is greater. As the cable-charging currents are not related to the size of the drive, the problem of unwanted trips due to these currents becomes less as the drive rating increases. In the following discussion the rated current level from the drive, IR, is referred to as a proportion of the trip level current, ITmp, of the drive in the form (IR = 0.83 Ire, lp). The trip level is the motor phase current magnitude that would cause the drive to trip. The drive current demand can be up to 0.83 x Ire, iP. The 0.17 IrmP margin allows for current-loop overshoot and cable-charging currents. Table 11.5 0.75 kW Control Techniques Unidrive Supply voltage 420 V 528 V Cable length 50 m 50 m Measured charging current O.18 x ITme 0.22 Ire,i,,


Cable resistance results in a voltage drop between the drive output and the machine terminals, which depends on the level of current flowing in the cable. If a drive is operated in closed-loop flux vector mode the cable resistance does not affect the performance because the closed-loop current controllers will automatically compensate for the voltage drop until the drive output voltage approaches its maximum limit. This limit, which is normally at a level approaching that of the supply voltage, will be reduced by the cable voltage drop. In closed-loop vector mode the maximum voltage applied to the machine cannot be higher than 95 per cent of available drive output voltage under steady-state conditions (although this may be further limited if the rated voltage parameter of the drive is set at a lower level). Once the drive output has reached this limit the flux in the machine is reduced (field-weakening or constant-power range) and more current is required to produce a given amount of torque. Therefore the cable resistance reduces the speed at which field weakening occurs, and in this range limits the maximum torque available and increases the drive and motor current.

Table 11.6 1.5 kW Control Techniques Unidrive Supply voltage 420 V 528 V 420 V 528 V Cable length 50 m 50m 100 m 100 m Measured charging current 0.14 Irlcle 0.16 x IrRIp 0.22 x Irmp 0.30 x Irme

Table 11.7 2.2 kW Control Techniques Unidrive Supply voltage 420 V 528 V 420 V 528 V Cable length 100 m 100 m 150 m 150 m Measured charging current 0.16 x ITRIP 0.22 x I7-1~12, 0.26 Irmp 0.30 IrRiP

Cable-Charging Currents
The current pulses due to the switching action of the inverter IGBTs vary in magnitude and length depending on the following: 1 Type of c a b l e - t h e pulses with armoured or screened cable are higher than those for cable without a conductive sheath. Only armoured cable has been used in producing test results, but results with screened cable would be similar. D.C. link voltage - the magnitude of the pulses increases with drive D.C. link voltage.

Table 11.8 4 kW Control Techniques Unidrive Supply voltage 420 V 528 V 420 V 528 V Cable length 150m 150m 200m 200 m Measured charging current 0.12 x Irmp 0.16 x Irmp 0.18 Irme 0.22 x Ivmp

Chapter 11.6


Tables 11.5 to 11.8 show measured values of the maximum percentage of the trip-level current produced within the overcurrent trip system by cable-charging currents alone when all three output phases switch together. This is the worst case seen by the overcurrent protection system. If long cables are used, the current limits should be set at a level which allows for current controller overshoot and the charging currents within the maximum current magnitude available from the drive. It is possible for a Unidrive to produce more than 175 per cent torque-producing current in a machine, and so even when the controller overhead and an allowance for cable-charging currents are taken into account, it may still be possible for the machine to produce 175 per cent torque. The following example illustrates typical calculations: A 2.2 kW Unidrive operating with 150 m of armoured cable with a 420 V A.C. supply. over current trip level - 20.2 A r.m.s. assume current controller overshoot - 0.1 ITmp cable-charging current - 0.26 Ire,/p total current magnitude < (1 - 0.26 - 0.1) ITe, i e 0.64 x Ire,/p .'. maximum current m a g n i t u d e - 20.2 x 0 . 6 4 - 12.9 A r.m.s. If machine-rated c u r r e n t - 9 . 5 A and power factor-0.9, then rated magnetising current-4.1 A and rated torqueproducing c u r r e n t - 8.6 A. At current limit the machine c u r r e n t - v / ( 4 . 1 2 + ( 8 . 6 x overload)2)= 12.9 A, ". overload-1.42 (i.e. the machine can produce approximately 1.42 x rated torque). Practical tests with a 528 V A.C. supply (480V + 10 %) have shown that the drive will operate without tripping using cables longer than would be expected based on the above information with the current limits set at 175 per cent. The reason is because all three phases did not switch at the same time when the drive output currents were at their peak values.

is all torque-producing current any reduction in current limit reduces the machine torque below 175 per cent of rated with a matched drive and machine.


For open-loop drives which do not use closed-loop current control, it is common for voltage-drop compensation to be included in the control strategy. This compensation is for both machine stator and cable resistance. The compensation, however, does not always remove the entire voltage drop. The graphs in Figures 11.6 to 11.11 show some typical results giving the level of torque produced for various sizes of Control Techniques Unidrive with standard Leroy Somer 400 V machines of the same current rating. In each case the current limits are set to 150 per cent and the torque taken for a level just before the current limit causes the frequency to fold back.

Cable-Charging Currents
The cable-charging currents have the same effect as in closed-loop induction motor mode. Again, the current limits should be reduced if necessary. As well as the normal frequency-controlling current limits the open-loop drive has a voltage-based peak limit designed to provide fast current limitation during transients. The peak limit should hold the current magnitude to 0.75 x I r R i P . The margin allowed for overshoot and cable-charging currents is higher than for the closed-loop modes; however the control system is much slower and so a larger overshoot could occur during transients.


Because of charging current, long motor cable runs can cause a reduction in the torque available from the drive, and in extreme cases the drive may even trip due to excessive current (OIAC). The effect is more marked with the smaller drive ratings. The following guidelines give suggested limits for standard shielded cables. The capacitance of the cable is approximately 300pF/m measured from the three phases together to the sheath (and ground core if fitted). This is typical of steel wire armoured cable (SWA), steel braid sheathed cable (SY) and similar, where the individual phase conductors are surrounded by a further insulating medium before being covered by the screen. For other values of capacitance the length limits are approximately in inverse proportion to the capacitance. Note that individual separate phase conductors give lower capacitance, but may not be acceptable for EMC reasons. Cables where the screen is laid directly over the phase conductors and mineral insulated cables are known to have much higher capacitance and should be treated with caution.


The drive again operates with closed-loop current control, which automatically compensates for cable voltage drop up to the drive output voltage limit. In servo mode the machine-rated voltage is usually lower than the drive limit because the machine is not to be operated in the fieldweakening mode (unlike a drive in closed-loop induction motor mode). Some headroom between the machine and drive voltages must be allowed for the drive current controllers to operate. This headroom will be reduced by the cable voltage drop.

Cable-Charging Currents
The cable-charging currents have the same effect as in closed-loop induction motor mode. Again, the current limits should be reduced if necessary. Because the machine current


A . C . DRIVE MOTOR CABLING: L i m i t s t o C a b l e L e n g t h

30 25
.:,.,~ ..,~., ~ "~~~ ~ _--.=-_ . _ ~ _ =

~ ~ . .


- -.~ ~ . . ~

E 20

~ . . . ~

=- 1 5 o o


100 m 150 m - - - 200 m ideal




25 frequency, Hz






Figure 11.6 0.75kW Unidrive-maximum torque production versus cable length (3kHz switching frequency, 2.5mm 2 armoured cable)

30 25
. . . . . . . . . . . . . . . . . . . .jp._'~,=,.~

E 20

=15 E


100 m 10 150 m 200 m ideal

| | u | u | n n n l




25 frequency, Hz






Figure 11.7 0.75kW Unidrive-maximum torque production versus cable length (12kHz switching frequency, 2.5mm 2 armoured cable)

30 25

E 2O

=15 E o 10


100 m 150 m - - - 200 m ideal




25 frequency, Hz






Figure 11.8 4kW Unidrive-maximum torque production versus cable length (3kHz switching frequency, 2.5mm 2 armoured cable)




30 25 E 20

= o-1 5 1o


100 m 150 m 200 m ideal





25 frequency, Hz






Figure 11.9 4kW Unidrive-maximum torque production versus cable length (12kHz switching frequency, 2.5mm 2 armoured cable)

80 70 60


E 50 = 40 E



20 10

150 m 200 m ideal










frequency demand, Hz

Figure 11.10 7.5kW Unidrive-maximum torque production versus cable length (3kHz switching frequency, 4mm 2 armoured cable)

80 70 60

E 50 = 40 E



20 10
0 | | | | I | | |

150 m 200 m ideal

| |










frequency demand, Hz

Figure 11.11 Z5 kW Unidrive- maximum torque production versus cable length (12 kHz switching frequency, 2.5 mm 2 armoured cable) Note 1: the compensation system in Unidrive is designed to operate fully only at frequencies below 1/4 rated frequency. This causes the droop in torque between 12.5 and 25 Hz. At higher frequencies the cable and machine resistance effects are proportionally less compared with the machine terminal voltage, and so the torque rises again as the frequency increases. Note 2: the ideal torque level shown is 150 per cent of the rated torque of the machine.


A.C. DRIVE MOTOR CABLING" Limits to Cable Length

Table 11.9 Maximum cable lengths at low and high supply voltage
Drive rating Cable length (m) at A.C. supply voltage and 3 kHz switching frequency 400 V 0.75 kW 1.1 kW
1.5 kW

which are affected by the high-frequency current in the cable capacitance.

Step 1 Estimate cable capacitance, from one line to all
others. Typical values: multicore cables, and screened/armoured cables where there is a plastic sheath between the phases and the screen: 130pF/m screened cables with no plastic sheath between cores and screen, mineral insulated cables" 300 pF/m Add an allowance for the motor capacitance. This depends on the motor size, but a value of 1 nF per motor is a reasonable estimate. This will usually be rather smaller than the cable capacitance.

528 V 50 75

65 100

2.2 kW 4 kW 5.5 kW-11 kW 15 kW-90kW 160 kW

200 300 300 200 300

150 250 300 124 300

Table 11.9 shows the maximum recommended cable lengths for Control Techniques Unidrive at low and high supply voltage. Select the lowest rated drive, which is shown in the table for the cable length, and supply voltage required. For other switching frequencies: maximum cable length = maximum cable length at 3 kHz (3 kHz/switching frequency)

Step 2 Decide on the available charging current from the

drive. This depends to some extent on the application and the drive; here is a typical procedure for Control Techniques Unidrive: Ich - 1.4 lln (2.1 - k) where" In is the nominal rated r.m.s, drive output current and k is the acceptable short-term overload factor. The drive is rated for 150 per cent, i.e. k - 1 . 5 , but increased cabledriving capability is achieved if this can be reduced. For many applications 1.25 is acceptable and for pump/fan applications 1.1 may be acceptable. In the case where the drive has been derated to allow for the long cable, i.e. a larger than normal drive is used, a lower value of k will be appropriate; it should be the ratio of the expected maximum short-term torque to the (theoretical) rated torque capability of the drive. The factor 2.1 in the expression is the ratio of drive instant trip current to nominal output current, for Unidrive. The correct factor must be applied for the drive being used.

15 kW Unidrive at 400 V, 9 kHz: maximum cable length - 200 (3/9) - 67 metres


The use of output chokes to facilitate the application of long motor cables is rare. It is usually simpler and more elegant to use a larger drive rating. Using output chokes allows drives to operate with long motor cables and/or multiple parallel cables. Calculation of the necessary inductance of the choke is complex, but the following guidelines are helpful. However, these are guidelines only and there is scope for error so wherever possible confirmation tests should be carried out in advance of installation.

Step 3 Decide the maximum D.C. link voltage, VD.c., where

the drive is required to produce full torque. This will normally be 1.41 times the highest r.m.s.A.C, supply voltage.

Step 4 The minimum value of inductance (per phase) is

given by the expression"

With single lengths of cable the current is limited by the self inductance of the cable, and the cable behaves as a transmission line with a Z0 typically in the range 15 f~ to 50 Q. The charging-current pulse magnitude is the D.C. link voltage divided by Zo; the duration is proportional to the cable length and is approximately 12 ns per metre of cable. For example, with 150 m of screened cable having Zo of 45 f~ and a D.C. link of 540 V the current is 12 A with a duration of 1.8 ~ts. For very long cables it is necessary to limit the current for drives in the lower power range (typically up to 11 kW). For multimotor applications where there are several cables in parallel, the effective Zo is very low and output inductors must be used. They also help with other problems such as the premature operation of thermal motor protection relays,

Lmin = T

L Ich J

If using standard iron-cored chokes, the inductance at the high frequencies involved will be rather lower than the specified 50/60 Hz inductance. A good rule of thumb is to specify an inductance of twice that determined by this calculation, i.e. 2Lmin.

Step 5 The maximum value of inductance is determined

by the acceptable voltage drop at the working frequency. Calculate this from the expression:

Lmax =

X VA.c. 2 71L V/3In

where x - a c c e p t a b l e voltage drop fraction, e.g. for 5 per cent use 0.05, VA.c.--motor voltage rating (line-to-line, r.m.s.) and f o - maximum drive output frequency.

Chapter 11.6
Since the voltage drop is inductive, it does not subtract directly from the motor terminal voltage. A value of x of 0.05 is generally acceptable. If the application is very critical with regard to obtaining full rated torque at full speed then it may be advisable to apply a lower value of x, e.g. 0.02.


Step 2 Using k - 1.25, as motor power is only 6 kW on a

7.5 kW drive, In - 16 A:

I c h - 1.41 1 6 ( 2 . 1 - 1.25)
= 19.2A

Step 6 If 2Lmin <Lmax then any value between these

limits can be used. If there is a need to minimise the high-frequency current, for example to prevent premature operation of thermal relays, then the highest value should be used. If 2Lmin > Zmax then the drive cannot operate with this length of cable and a higher rated drive must be used. Choose the value of L.

Step 3 Vo.c.- 1.41 400 x 1.1

= 622 V

Step 4
Lmin - ( 2 0.541 x 10-6/3) (622/19.2) 2
= 378 gH

Step 5 Allowing 5 per cent voltage drop, x - 0.05 Lmax - (0.05 400)/(2 7r 50 x v/3 16)
= 2.3mH

Step 7 Consideration must now be given to the highfrequency losses in the chokes. The loss in each choke can be estimated from the following expression:
P = o.8f cv

Step 6 2 378 gH < 2.3 mH so any value between these

limits can be used. Use 2 mH to minimise the high frequency current, to prevent premature operation of thermal relays. where fs is the drive switching frequency. The factor 0.8 is a rough estimate of the fraction of the total losses dissipated in the choke. Note that the loss is proportional to the switching frequency so the lowest acceptable frequency should be selected.

Step 7
P0.8 6 103 0.154 10 .6 6222 = 286 W This is excessive, so reduce the switching frequency to 3 kHz P0.8 3 103 0.154 10 -6 6222 = 143W

Step 8 It is now necessary to decide whether the choke is able

to tolerate this loss. This is a difficult judgement. As a crude rule, the loss should not exceed 0.1 of the VA in the choke at maximum speed, i.e.:
P ~ 0.27rfo(max)Zl2

where L(max) is the maximum output frequency. If the loss exceeds this limit, and it is not possible to reduce the switching frequency, then a resistor should be connected in parallel with each choke to extract some of the power. The resistor value is given by:

Step 8 This loss will be acceptable if:


10 -3 1 6 2 - 16.1W

The calculated loss exceeds this limit, so a resistor should be connected in parallel with each choke. The resistor value is given by: R - v/[2 x 2 x 10-3/(0.154 x 1 0 - 6 ) ] - 161 f~ The value is not critical so choose 150 f~ or 180 f~ with a power rating of 100 W.

The value is not critical and variations of +50 per cent are acceptable. The power rating of the resistor should be at least 0.8 P. Provision must be made for the resistor to dissipate this power without overheating itself or nearby equipment. Values of 100 W per phase are not uncommon.

Position of Chokes in Multiple Motor Configurations

Note that these guidelines apply to single continuous or multiple contiguous runs of cable. Multiple parallel runs - as might be used where several motors are driven from a single drive - cause much higher charging current than the equivalent contiguous run and should be avoided. This is because the inductance of the cable reduces the effect of the capacitance, but this benefit is lost for multiple parallel sections. If multiple cables are unavoidable then output line inductors should be used (Refer to Figure 11.12). It is not recommended that any drive be operated with more than 300m of cable. If this is necessary then it is recommended that an output filter be used to remove the switching frequency component of the output-sometimes referred to as a sinusoidal filter. These are available from specialist filter suppliers.

A Control Techniques 7.5 kW Unidrive with eight motors of 0.75 kW each connected by 140 m of multicore cable. 400 V A.C. supply, maximum output frequency 50 Hz, switching frequency 6 kHz. Thermal relays in all motor circuits.

Step 1
cable capacitance - 130 pF 140 8 - 0.146 laF motor capacitance -- 1 nF x 8 -- 0.008 l.tF C = 0.154 gF


A.C. DRIVE MOTOR CABLING: O u t p u t Chokes for Long M o t o r Cable Applications

Output Line Inductor

incorrect wiring for

multipleparallelcable run

~~" V

correct wiringfor multiple parallelcable run


preferred wiring for multiple motor connection

Figure 11.12 Position of chokes in multiple motor configurations. (Individual motor protection relays are also required- not shown)


The cross-sectional area of the cable conductors determines the resistance. The recommendations given in drive instruction manuals are minimum sizes based on ensuring safe operation without unacceptable heating of the cable. Where long cable runs are used it is advisable to check the voltage drop. Typical cable cross-section recommendations are as given in Table 11.10 for both A.C. supply cables and A.C. motor cables for three-phase applications. Standard codes of practice can be applied to ensure acceptable voltage drop. In most cases the current used for the voltage-drop calculation will be the maximum continuous motor current, but if in a particular application the overload capability is needed at or above base speed then the overload current should be used (e.g. 150 per cent of full load current).

Table 11.10 Cable cross-section recommendations

Drive current rating 2.1 2.8-12 16-25 34 40-46 6O 70 96-124 156 180 202-300 mm 2 1.5 2.5 4.0 6.0 10.0 16.0 25.0 35.0 50.0 7O.0 95.0 AWG 16 14 10 8 6 4 4 2 2/0 2/0 3/0

Chapter 11.7



The A.C. power source can have a critical impact on the performance and reliability of any power conversion equipment including drives. Most variable-speed drives, from leading suppliers, are designed to operate from typical A.C. power sources found in industrial plants and commercial installations throughout the world. Nonetheless it is vitally important that care is taken with all installations to ensure that the supply is understood and, should any abnormal conditions exist, that steps are taken to overcome possible difficulties. Three possible abnormal conditions need consideration.

A.C. drive systems without a D.C. link choke can have problems if the r.m.s, fault current is not limited to approximately 16 kA. This could result in damage to the drive. D.C. drives could be damaged if operated on low-impedance supplies. In such a case failure of the power semiconductors could occur due to excessive rate of change of current (di/dt). This characteristic is not unique to specific drive products but a limitation of the available power semiconductors. In cases where the supply impedance is too low, additional reactance (inductance) needs to be added in series in the line feeding the drive. This reactance may be in different forms: a b c iron-cored, nonsaturating reactor/choke air-cored reactor/choke isolating transformer


Drive products are now available to cater for wide, but clearly defined, spreads of supply voltage. Operation outside the defined voltage range is not permitted. Supplies, which fluctuate greatly outside these boundaries may well cause damage to the drives. If high or low-voltage conditions are suspected, it is recommended to take a chart recording of the line voltage to determine the extent of the problem. Ensure that this is done over a worst-case period which may for example be when large loads elsewhere in the installation are running. A transformer (isolation or auto) with taps will usually correct a high or low supply voltage condition.

Note: autotransformers are not recommended as they present very low inductance. A good rule of thumb is to ensure a total supply impedance of approximately 4 per cent reactance. As a rule of thumb, or where information about the supply is not known, it is good practice to fit line reactors of 2 per cent. Typical calculation: line voltage - 400 V line frequency = 50 Hz supply impedance = 1% drive rating = 100 kW


For almost every application served by an electricity supply utility, supply frequency variation will never be a problem. Care should be taken, however, if variable-speed drive equipment is operated from a generator set. In this case, for most A.C. drives, with an uncontrolled (diode) supply converter, there is no cause for concern. For D.C. drives, where the control has to synchronise to the supply, it is important that the supply frequency remains within the defined range (for the Control Techniques Mentor/Quantum this is 48-62 Hz) and the rate of change of frequency is below 5 Hz per second.

To calculate required inductance for total impedance of 4 per cent: 1 per unit impedance - (v/3 x (400)2)/100 x 103 = 2.77 ft 1 per unit inductace - 2.77/(2 x 7r 50) = 8.8mH required line r e a c t a n c e - 3% x 8.8 mH = 264 ~tH A suitable choice would be a 265 ~tH/200 A three-phase reactor/choke.


High-quality commercial variable-speed drives are designed to operate from typical industrial power distribution systems with a maximum fault level of ten to twenty times the drive rated power. Problems can also occur if a drive system is installed very close to the main power supply or powerfactor correction capacitors, both of which present low supply impedance to the drive. A.C. drives with D.C. link chokes are in general unaffected by low supply impedance.

High Supply Impedance

In cases where long power lines are feeding a drive, the supply can have a high impedance. Such installations are often prone to interruption and switching effects which require consideration in relation to supply voltage, as discussed earlier. Another very important effect of high supply impedance is associated with multiple drive installations. All power electronic drives comprise power semiconductor switches, which sequentially switch the supply lines to a load. As one



situations. Sites with high impedance supplies may require additional measures for reliable operation. There is no easy solution to overcome this situation. Power lines carrying drive current need to be oversized, as do any transformers used in order to minimise the impedance. This oversizing may need to be as high as five times that normally considered adequate.

drive switches its load onto the line, the voltage at the terminals of the drive dips as the current drawn flows through the high line impedance. This voltage dip, or notch, is not only seen by the instigating drive but by other drives as well. This notch may be so severe as to cause power semiconductors of an adjacent drive to falsely turn on by a process known as the Miller effect (high dV/dt charging internal parasitic capacitance within the power semiconductor itself). This notching effect is sometimes described by users as 'the drives talk to one another', or 'one drive interferes with another especially during periods of high acceleration or heavy load demands'. This is not so much of a problem with A.C. drives where controlled rectifiers are less common on the supply, but can be critical for many D.C. drive systems. High-quality drives contain snubber networks (resistor/ capacitor bypass networks) and transient voltage suppressors (MOVs-metal oxide varistors), which protect the power devices and ensure reliable operation in typical industrial

Multiple Drive Installations

The cross coupling of drives discussed above can be a very serious problem. It is good practice that drives which form part of multidrive systems should be fitted with independent line reactors. IEC 146 recommends the fitting of line reactors approximately equal to the supply reactance. This limits the notches seen by other drives to approximately 50 per cent of the magnitude seen without the reactors. This would normally only be a problem in cases with high-reactance supplies.


The reliable trouble-free operation of all industrial equipment is dependent upon operation in an environment for which that product was designed. The single most significant reason for the premature failure of a variable-speed drive controller is operation in excessive ambient temperature. The design of the enclosure in which the drive is housed is therefore of critical importance. The following guidance covers the basic calculations necessary to ensure that heat generated by a drive can be satisfactorily transferred to the air surrounding the cubicle. When making the calculation remember to take account of all power dissipated in the cubicle not simply that generated by the drive. Further, in the internal layout of the cubicle, where possible, avoid placing electronic components at the top (hot air rises!), and where possible provide fans to circulate internal air. Remember, as a rule of thumb, an electronic product's lifetime halves for every 7C temperature rise!

that only walls which are not obstructed (not in contact with walls, floor or another hot enclosure) can dissipate heat to the air. Calculate the minimum required unobstructed surface area Ae for the enclosure as follows:
Ae = P/[k(Ti - Tamb)]

where Ae = unobstructed surface area (m2), P = power dissipated by all heat sources in the enclosure (W), Tamb = maximum expected ambient temperature outside the enclosure (C), Ti=maximum permissible ambient temperature inside the enclosure (C) and k = heat transmission coefficient of the enclosure material (Wm -2 c - l ) .

To calculate the size of an enclosure to accommodate the following: two Control Techniques 4 kW Unidrive drives EMC filter for each drive braking resistors mounted outside the enclosure maximum ambient temperature inside the enclosure 40C maximum ambient temperature outside the enclosure 30C maximum dissipation of each drive = 190 W maximum dissipation of each EMC filter = 25 W total dissipation = 2 (190 + 25) = 430W


The enclosure itself transfers the internally generated heat into the surrounding air by natural convection, or external forced airflow. The greater the surface area of the enclosure walls, the better is the dissipation capability. Remember also

Chapter 11.8


The enclosure is to be made from painted 2 mm (3/32 inch) sheet steel having a heat transmission coefficient of 5.5 Wm-2C -1. Only the top, the front and two sides of the enclosure are free to dissipate heat. The minimum required unobstructed surface area Ae for the enclosure is as follows:



Ae - P/[k(Ti - Tamb)]
= 4 3 0 / [ 5 . 5 ( 4 0 - 30)] = 7.8m 2 If we select an enclosure with a height (H) of 2 m, a depth (D) of 0.6 m, and minimum width Wmin" dissipating surfaces top + front + (2 sides) > 7.8 m 2 > 7.8 m 2

Figure 11.13 Enclosure having front, sides and top pane/ free to dissipate heat

(Wmin x 0.6)+(Wmin X 2)+(2 X 0.6 2 ) > 7.8m 2 Wmin > ( 7 . 8 - 2.4)/2.6

>2.1m If the enclosure is too large for the available space it can be made smaller only by: reducing the power dissipation in the enclosure reducing the ambient temperature outside the enclosure increasing the permissible ambient temperature inside the cubicle if possible by derating equipment in line with manufacturer's recommendations increasing the number of unobstructed surfaces of the cubicle Ti = maximum permissible ambient temperature inside the enclosure (C) and k = r a t i o of Po/Pi where Po is the air pressure at sea level and pi is the air pressure at the installation. Typically, a factor of 1.2 to 1.3 can be used to allow for pressure drops in dirty air filters.

To calculate the size of an enclosure to accommodate the following" three Control Techniques 15 kW Unidrive drives EMC filter for each drive braking resistors mounted outside the enclosure maximum ambient temperature inside the enclosure 40C maximum ambient temperature outside the enclosure 30C maximum dissipation of each d r i v e - 570 W maximum dissipation of each EMC f i l t e r - 60 W total dissipation- 3 x (570 + 6 0 ) 1890 W


In this case the dimensions of the enclosure are determined only by the requirements to accommodate the equipment making sure to provide any recommended clearances. The equipment is cooled by forced air flow. This being the case it is important in such an arrangement to ensure that the air flows over the heat-generating components to avoid localised hot spots. The minimum required volume of ventilating air is given by:

Then the minimum required volume of ventilating air is given by:

v3 k P / ( V , - Ta,.b)

V -- 3kP/(Ti


3 hr -1 = (3 1.3 1890)/(40 - 30) = 737 m 3 hr -1

where V - cooling air flow (m3hr-1), P - - power dissipated by all heat sources in the enclosure (W), Tamb = maximum expected ambient temperature outside the enclosure (C),




MOTORS General
The installation, commissioning and maintenance of industrial power drive equipment requires continuous regard to the statutory requirements of safety in industry. Industrial power-supply voltages and high-speed high-torque drive systems, unless handled properly, can represent great danger. All equipment must be used in accordance with the duty, rating and conditions for which it is designed, and particularly the power supply must be in accordance with that shown on rating plates, subject to standard tolerances. The loading and speed of the driving motors must not exceed those of their rating plates or any overload ratings agreed formally with the makers. No attempt should be made to open inspection apertures or similar openings unless the drive is known to be fully isolated from the power supply and the motor cannot be rotated from the load side. All safety and protection guards and covers should be in place before motors are started. Power and control cables associated with the drive must be of adequate current-carrying capacity and voltage grade for the duty, and properly mounted and secured. The power supply system to which the drive is connected must have an adequate short-circuit fault clearance level. Motors must not be operated under ambient conditions for which they were not designed. Ambient air temperature should not exceed 40C or the temperature agreed nor should cooling air be contaminated in any way injurious to the machine, nor restricted in flow into ventilating inlet apertures.

excessive brush wear. Any deflection of the motor flame by bolting down upon an out-of-true base must be avoided by the use of shim washers. It is wise to blank off any open motor apertures during installation to avoid foreign material entering the motor. If wide temperature and humidity variations are possible, heat should be carefully applied to ensure that the motor is dry internally before starting up. A check of the insulation values of the windings to earth by the Megger test should indicate not less than 20 Mf~. If below, refer to the supplier who will advise suitable action. It is strongly recommended that during such insulation tests any electronic apparatus associated with the motor should first be disconnected, otherwise damage may occur. With brush gear machines, both D.C. and A.C., check that all brushes are in position, can be moved freely in their holders, press on to their running surface with equal pressure; that they are fully bedded on to this running surface, and that all connections are tight. If practicable, rotate the shaft by hand to check freedom and smoothness. If a speed/position feedback device is fitted, check any coupling for security and accuracy of fitting. If the feedback device is stub-shaft mounted, check for concentric running (0.05 mm or 0.002 in eccentricity at the stub shaft end may be regarded as maximum) and for correct axial positioning, with the brushes in the centre of the commutator width in the case of a D.C. tachogenerator. Ensure that there is no obstruction to the cooling air flow to the motor, and that the body/fin casing of a totally-enclosed fancooled machine is clean, free from debris and that the air inlet to the fan cowl is unobstructed. With screen-protected motors ensure that inlet and outlet ventilating apertures are clear. Check that the motor is not in the direct path of hot air flow from other machines or equipment, and that all guards and covers are in position. If the motor has a forced vent fan unit, ensure that the fan impeller rotates in the correct direction. This needs to be done by careful inspection, since the wrong direction of running does not necessarily reverse the airflow but can reduce the air volume below 60 per cent flow, with a risk of winding burnout. An arrow normally indicates the correct direction. If in doubt refer to the manufacturer. If this is not convenient, a good indication can be obtained by checking the airflow in both directions of impeller rotation. The largest volume and the highest air velocity usually indicate the correct direction of rotation. The fitting of pulleys and couplings to motor shafts calls for care if these are of interference fit, since excessive axial pressure or impact force can easily damage modem precision bearings. Heating the pulley or coupling prior to

If motors are stored prior to installation, storage conditions should be such as to avoid deterioration; otherwise the manufacturer's guarantees could be invalidated. Motors, either cased or not, should never be stored out of doors. Packing cases are invariably not weatherproof. Storage should be clean, dry and free of vibration which can cause brinelling damage to motor beatings. Extremes of temperature and humidity can cause injurious condensation.

Motors should be mounted on rigid, level foundations or flange mountings, which are horizontal or vertical or as specified. Mounting rigidity is important, particularly with brush gear machines, since motor frame vibration can cause

Chapter 11.9
mounting is good practice, as is supporting the motor shaft at the nondrive end before pressing on. The use of a tapped hole in the motor shaft end, now a standard with many manufacturers, to pull on a pulley or coupling is obviously a sound technique. For motors fitted with brush gear especially, it is good practice to specify a balanced pulley. Most manufacturers balance rotors to good standards; a seriously out-of-balance pulley can adversely affect brush gear performance, and in any case imposes excess load on bearings. Correct belt alignment and tensioning is required for best belt life as for best motor beating life. Correct coupling alignment is important in avoiding excessive radial and axial loading on the motor bearings. Perhaps the most common coupling between the motor and its load is the vee belt, which often allows the use of higher speed, more efficient, more economically priced driving motors. It also acts as a buffer against mechanical shock loading between the motor and load. The wedge action of the vee belt gives increased friction between belt and pulley for a given belt tension. This reduces belt slip, thus extending belt life and, by reducing motor bearing side loading by comparison with the equivalent fiat belt, extends beating life also. Another advantage of the vee belt is that wide drive ratios of up to four or five to one are practical, as are short centre drives when space is limited. The shortest spacing between centres approximates to the diameter of the larger vee pulley. Modem belts stretch in service initially, before stabilising. It is therefore most important to check vee belt tension after a few days' running and to readjust the tension to recommended values.


correct grease for the bearing type and speed, at intervals of approximately 2/4000 hours determined by running speed. Some large machines incorporate grease escape valves on the beating housings; these bearings should be relubricated in accordance with the lubrication plate fitted to such machines. Overlubrication is the most common cause of bearing failure. An overgreased beating can overheat, seriously damaging the beating and changing the lubricating properties of the grease, which in turn does further damage. Ball and roller beatings require little attention other than a periodic check while running for unusual noise, or signs of overheating. Whistling noises are usually caused by defective lubrication, and rumblings by contaminated grease or damaged surfaces. Bearing running temperature depends upon loading, speed, motor and ambient temperatures. Modem motors run quite warm and it is expected that bearing temperature will be slightly above that of the motor frame for typical speeds. For high-speed motors, beating temperature can be expected to be somewhat higher and bearings will operate quite satisfactorily at temperatures up to 100C, although this would be high for current designs. Any significant change in running temperature requires investigation, and if there are any signs of bearing damage heavy black staining or blueing, surface cracking, brinelling, track indentation or excessive w e a r - t h e beating should be replaced. Detailed lubrication instructions are generally provided by the manufacturer, and these should of course be followed.

Brush Gear Maintenance

For safety, isolate the drive from the electric power supply before attempting any work within the motor. The covers should be removed at regular intervals so that brush gear can be inspected and cleaned. Experience with the installation will suggest the frequency, determined by motor use, speed, vibration level and general cleanliness. After initial commissioning, it is wise to check brush gear at weekly intervals to obtain an appreciation of brush gear behaviour. When this is confirmed to be good, the intervals can be increased with confidence. In general, brush wear, after an initial bedding period of a month or so, should not be less than 8/10 000 hours per inch of brush length. Total permissible wear is to about 50 per cent of the original length. If upon examination the brushes have not worn excessively, have an even glaze (but not a polished shine) on the running face, are not scored or chipped or broken, slide freely in the brush boxes, are not discoloured and have flexible conductors correctly attached, then they can be returned to service. Any dust accumulation on the brush gear can be removed by a stiff brush or an air blower, and in replacing the brush ensure that the flexible connection is securely fixed and free of any obstruction. Brush contact pressure is fixed by the manufacturer in the case of tensator clock spring-type tensioning. Both this and spring-arm-type tensioning can be checked with a spring

Maintenance Guide
Maintenance involves much that is self evident yet is of increasing importance where emphasis is upon minimum downtime and production loss. Regular inspections ensure that motors remain clean and dry externally and internally. Oil particularly must not be allowed to accumulate on motor brush gear, commutator and slip tings. Ventilating grills and air filters should not be allowed to become obstructed. Compressed air for cleaning should be used with care, so that contaminants are not driven inside the motor or between winding coils. Hand bellows or a small centrifugal blower are more appropriate. The majority of motors below 100 kW rating have sealed bearings, which are capable of good performance over many years of service. In fact, many such bearings are shielded rather than sealed, for true seals at modem motor shaft speeds would produce significant additional bearing heating. Where after several years' running a motor is taken out for a service overhaul, it is customary to replace the bearings, or at least melt out the existing grease, clean up and replace with the correct specification of heated grease ensuring its penetration onto the bearing running tracks. Motors with provision for bearing regreasing should be given four or five strokes from a grease gun filled with the



balance. As a general rule, brush pressure should be around 0.2 kgcm -2 of brush area for speeds up to 2500min-1 0.25 kgcm -2 for higher speeds and for traction machines subject to high vibration levels, 0.3 kgcm -2. With spring-arm brush tensioning, adjustment can be made by using alternative slots on the arms. However, changes from those set by the manufacturer should only be made for good reason. In fitting new brushes, it is important that brushes of the originally fitted grade are used and carefully bedded in, one brush at a time. Lift each brush in the holder and place a long strip of grade 0 glass paper, abrasive side to the brush, between the brush and the running surface. Draw the glass paper backwards and forwards following the curved surface of the commutator or slip ring to ensure that the whole face of the brush is shaped to the correct curve. Repeat for each brush in turn, ensuring that carbon and glass dust is kept out of the motor. Never use carborundum paper, the particles of which can embed permanently in the brush face and thereafter score the commutator surface. Any disturbance of the brush box mounting arm holding ring position (rocker ring), requires that the ring be returned to the previous position precisely and clamped. Its position is determined for the best commutation during the manufacturer's testing and marked with paint or a stamped mark on the rocker ring. Brushes are reaching the end of their useful life when worn down to about 50 per cent of their new length. All brushes should be changed together as a set, even though some are less worn. If a commutator or slip-ring surface is marked, blackened or grooved, it can be corrected by the careful application of a commutator stone (glass/epoxy commstone). The stone should rest on a convenient brush arm, be kept in firm contact with the commutator or slip-ring surface and moved axially to ensure that the whole working surface is cleaned as the motor shaft is rotated briskly by hand. Never attempt to clean a commutator with power on the motor. Serious marking or scoring requires the commutator or slip tings to be skimmed on a lathe, to maintain concentricity. Thereafter the skimmed surface has to be lightly polished. Skimming will sometimes require commutator mica insulation between segments to be undercut, that is, cut back to avoid abrasive contact with the brush face. At the time of undercutting the segment, slot edges should be given a slight chamfer to ensure that carbon does not accumulate in the slot. This is skilled work and requires proper training. The running surfaces of commutators and slip rings should have a smooth chocolate-brown appearance after a few weeks of use. This patina indicates low friction and good electrical performance and should not be disturbed. Slight pinpoint white sparking at the brushes is generally noninjurious and in fact appears to assist in the early establishment of a good patina. In the event of marked sparking, blackening or scoring of the running surface, inspect for rough areas, eccentricity or flats developing. With commutators, inspect also for protruding or recessed copper segments or high-slot micas. Depending upon the severity of the cause, the damaged surface may be

cleaned with a strip of glass paper or a commutator stone. Again it is stressed that these operations should only be carried out by a skilled operator. If this correction proves unsuccessful, skimming between lathe centres, and mica undercutting in the case of the commutator, will be necessary. A common cause of poor commutation, excessive brush wear rates and commutator scoring or blackening is low brush current density for prolonged periods. Driving a D.C. motor mechanically with no current flow through the brushes would result in very high brush wear rates and damage to the commutator surface. Current flow through the brush/ commutator contact area and also the presence of some water vapour in the atmosphere are essential to good brush gear performance. It appears that a state of ionisation in this contact area, with carbon and water molecules present, provides the lubrication necessary and, as previously mentioned, a little pinpoint white sparking helps. In general, at least 60 per cent of the nameplate current should, on average duty, flow through the brushes. This is easy to assess in terms of current density, knowing that half the total number of brushes are connected to the positive and half to the negative armature terminals. By measuring the cross section of each brush and multiplying that area by half the brush number, the total brush area per terminal is known. The current density for standard motor designs should not be less than 77 mA/mm 2 (50 A/in2). Where the current density is below 60 per cent of full-load current continuously, it can be raised by removing one brush per brush arm from the commutator track, leaving an equal number of brushes on each brush arm. If in doubt, the motor manufacturer should be consulted.


It is essential that electrical equipment is isolated from the incoming supply and that sufficient time is allowed for any internal supplies to discharge fully before any work or internal adjustment is started. In general terms there is little need, if any, for routine maintenance of modern power electronic equipment since the solid-state technology involved simply has no components, beyond power fuse gear, which require planned replacement. However, there are occasions when work is required; for example, where an electronic control fault is suspected it may be useful to monitor important supply voltages within the equipment or to substitute a spare printed circuit board. The following points are pertinent.

Siting of Equipment
In both A.C. and D.C. applications, important considerations apply to the siting of the drive module. A primary consideration is ventilation, which is essential to allow the drive to perform to its full specification. Other vital considerations are ambient temperatures, humidity and purity of the cooling air. This is especially important in locations where carbon black or flammable

Chapter 11.9


solvents are present. Therefore the siting of the drive has to be carefully planned and in severe cases ducting may have to be arranged to carry air from outside the area or building. The ambient temperature of the air drawn into the ducting should also be taken into account, to ensure that the heatsink is able to dissipate the heat generated by the output devices. Normal limits of ambient temperature for electronic drive modules are - 1 0 C to +40C before alternative methods of cooling have to be considered. If ambient temperature falls below -10C, controlled heating may be necessary. Drive cubicles typically have separate cooling-path ventilation built into the panels. When siting these cubicles, care must be taken to ensure that the inlet and outlet vents are not obstructed in any way by other equipment (which in itself may generate heat) or by any other structure.

allowed to settle on the electronic equipment where high voltages are nearly always present.

Condensation and Humidity

Generally, equipment intended for use in areas of high humidity or condensation is designed to minimise the possible generation of water vapour and water droplets since the presence of water in any form is definitely undesirable in equipment of this type. Anticondensation heaters are commonly used and should be regularly checked for correct functioning. Early in the service of any installation it is advisable to watch for the formation of any unexpected condensation, especially in situations where the equipment is not powered-up continuously and where temperature cycling may be a contributory factor. In such cases anticondensation heaters should be retrofitted.

Ventilator Systems and Filters

Low power electronic equipment is often cooled by natural convection via an arrangement of electrically isolated heatsinks which themselves form part of the equipment structure. Higher power equipment is most often housed in wallmounting or freestanding cubicles and frequently has a ventilation system consistent with the environmental rating of the enclosure itself. This may in its simplest form be a totally enclosed cubicle of sufficient volume to ensure that the dissipation over the surface area of the exposed sides is adequate for the heat generation within. Next would be a simple louvre system which, beyond ensuring that the louvres are not restricted in any way, would present no reason for concern. Forced-air arrangements, utilising an air input and exhaust route, are very common and normally include air filtering of a conventional paper or fibre type. It is unusual, however, to have a system which warns of filter contamination, apart from for filters normally guarding the power electronics themselves against overtemperature, and consequently it is important to check such filters regularly. This service period is normally defined by experience of the actual environment. Filters are most commonly of the disposable or nonreusable kind. However, some may be reused and the recommendations of the supplier should be followed. Do not, ever, risk operating the equipment without the filter element; the atmosphere may carry electrically conductive particles, which will eventually cause malfunctioning if

Fuses have a finite life and may be expected to fail due to fuse-element ageing, especially if the normal current lies towards the upper end of the fuse rating. Many fuse-gear manufacturers quote life expectancy under stipulated conditions, but the operating conditions of most installed systems are seldom known and only an average life can be rightfully expected from any fuse. Consequently, when encountering an open-circuit fuse it is important not to immediately assume that there is a specific reason for failure other than ageing. However it must be stressed that, before a fuse is replaced without further question, it is most important that appropriate electrical tests are made to ensure that there are no obvious short circuits or overloads in the protected circuit. Extensive supplementary damage can be caused to both electronic equipment and associated electrical equipment by replacing fuses or resetting circuit breakers before conclusive tests. Modern fuse technology is extremely complex, with many special-purpose fuses being specified, especially so in the protection of semiconductor devices. It is not acceptable to simply fit a replacement fuse of the same current rating. The replacement fuse must be either a direct replacement of the original or an exactly comparable type of approved and listed characteristic demonstrating equivalent current, voltage and rupturing capability. If necessary, the fuse supplier or manufacturer should be asked for verification. Replacement of any fuse by one of a greater value is rarely necessary and any such decision should be carefully considered.





Connecting the D.C. links of several drives together allows regenerated/braking energy from one drive to be reused by another motoring drive. This improves the efficiency of the system, since the regenerated energy is not wasted in braking resistors and the motoring drive draws substantially less power from the mains. This can be particularly advantageous where one or more drives may be holding back a line to provide tension. It is also often applied in high-performance servodrive applications where substantial amounts of energy are used in accelerating and braking drives. As well as offering advantages in terms of simplifying energy management, a common D.C. bus system also has the potential to simplify the mains connection and protection, as usually only a single mains feed is needed. There are disadvantages, however, and care needs to be taken in the implementation of such a system. Direct connection of the D.C. links of A.C. drives usually entails the direct connection of the D.C. link capacitor banks of all inverters. These capacitors store substantial amounts of energy. Further, these electrolytic capacitors are subject to ageing, and failure can be rapid, unpredicted and usually to short circuit. In such an event, all the stored energy in all the D.C. link capacitors of the group of drives will be fed into the fault, causing substantial damage to the original failed drive as well as to others. Protection can be afforded through the provision of fusing between the linking D.C. busbar and the individual drives in both lines. Care needs also to be taken in respect of any requirement for isolation of individual drives, which would now need to be undertaken at the D.C. link. The availability of D.C. switchgear (and fuses) is much less than for A.C. equivalents. D.C. fuses do have to be rated at higher voltage levels than the standard ratings used for A.C. protection. The subject of fusing in common D.C. bus systems is somewhat confusing only in respect of the policy of some drive manufacturers which do not recommend fusing. Although in some cases the fuses are integral to the converter package and are therefore present but not discussed, in others reliance is clearly placed upon electronic monitoring and control to protect the drive. The merits of such a philosophy are unclear and it must be assumed are based upon a reliance on the integrity of the D.C. link components, notably the D.C. link capacitors. Control Techniques recommends suitably rated fuses in each D.C. connection. Having decided upon a common D.C. bus system, it is necessary to decide upon the form of power-supply conection. This is dependent upon many factors including:

whether all the drives to be connected to the bus are, or can be, of the same rating the peak current to be drawn from the system in relation to the individual ratings of the drives the location of the D.C. bus soft-start/charging circuit within individual drives

A number of mains supply converter configurations are possible including: a simple bulk uncontrolled rectifier utilising the mains supply converter in one of the drives to supply all drives feed mains supply converter of all drives and connect the D.C. buses of all drives; effectively hard paralleling of all drive input rectifiers a bulk four-quadrant controlled rectifier a bulk four-quadrant PWM converter

A simple chart showing linkage between requirements and the alternative solutions may be helpful in making the initial selection, Table 11.11.


The use of a bulk input converter is strongly preferred if the installation requires drives of different ratings to be connected together. The D.C. link chokes of standard drives tend not to be in circuit if the drive is supplied with D.C., and hence it is necessary to supply an external choke the specification of which depends on the total rating of the drives connected to the D.C. link. Splitting the inductance equally between the positive and negative D.C. link can provide some impedance to limit fault current if an earth fault occurs in either the positive or negative D.C. link. It is also necessary to provide an external rectifier module. Inrush current limiting is not required when the individual drives have their soft-start circuits (inrush resistors and relays/contactors) in circuit until the D.C. link is at the correct level. A high-voltage polypropylene capacitor should be fitted across the D.C. terminals of the rectifier module. This helps to reduce the reverse recovery voltage spikes, and can also help provide a path for RFI currents in applications where long motor cables are used. The varistor network on the input phases provides protection against line-line and line-earth voltage surges. This is required as the A.C. input stages of the drives are not being used. Such a configuration is shown in Figure 11.14. The D.C. link inductance should be selected to keep the D.C. link time reasonably constant. As the D.C. link capacitance is the arithmetic sum of the capacitance of all connected drives, the required D.C. link inductance can be calculated as

requirement to regenerate energy back into the mains supply fail-safe braking requirement supply harmonic limitations use of standard/commercial components

Chapter 1 1 . 1 0 Table 11.11 Linkage between requirements and alternative solutions

A simple bulk uncontrolled rectifier Regenerate energy back into the mains supply Fail-safe braking requirement Supply harmonic limitations no Supply converter in one drive to supply all drives no Hard paralleling of all drive input rectifiers no A bulk four-quadrant controlled rectifier yes


A bulk four-quadrant PWM converter yes

if braking resistor fitted 6 pulse 12 pulse

if braking resistor fitted

if braking resistor fitted

if braking resistor fitted 6 pulse 12 pulse

if braking resistor fitted

Use of standard/ commercial components All drives to be connected to the bus are, or can be, of the same rating The peak current drawn from the system in relation to the individual ratings of the drives The location of the D.C. bus soft start/ charging circuit within individual drives

6 pulse 12 pulse unimportant

6 pulse 12 pulse unimportant

unimportant but unlikely

unimportant to ensure good current sharing as current sharing will not be perfect even in this balanced system safety margins are needed


determines the rating of the rectifier

only possible if peak rating < rating of an individual drive rating

determines the rating of the rectifier

determines the rating of the rectifier

needs to be between the D.C. bus connection and the D.C. link capacitor in the drive

Note: where used, the more the better


Lb. C.
Ii | n

+ D. C. link drive A - D. C. link



+ D. C. link drive B

varistor network

0.47 IJF polypropylene (typical)



- D. C. link

+ D. C. link drive C

'I ,

l |

- D . C. link


D.C. bus

Figure 11.14 Common D.C. bus fed from a simple bulk uncontrolled rectifier for drives with internal current limits



follows: 1/LD.c. -- 1/Ldrive 1 + 1/Ldrive2 -Jr-''' where Ldrivel, Ldrive2 etc. are the design values of the chokes fitted in the standard drives.

current sharing is good. However, a 10 per cent derating is strongly recommended to allow for small imbalances which will exist. Such a configuration is shown in Figure 11.16. In the design of such systems a number of further issues need to be considered including: A.C. and D.C. power wiring should be of a star form and not daisy chain with efforts made to equalise cable length drives should be located close to each other the resistance of the A.C. fuses may help in sharing current so it is important to use the same fuse types in corresponding positions deliberately adding some external resistance could help with sharing A.C. line chokes can be used to help sharing as the brake threshold will be slightly different for each drive there should be only one main brake resistor for all the paralleled drives

This arrangement is possible only where the peak current drawn from the system/group of drives is lower than the rated current of the largest drive. This can be the case in applications such an unwinder-winder, or a machine tool where there is a large spindle and small axis drives. In applications where this condition is only marginally satisfied, use an overrated drive to facilitate this solution. In this case no additional D.C. inductance is required as that within the large drive is being utilised. Such a configuration is shown in Figure 11.15.

If A.C. supply and D.C. link connections of drives with different ratings are connected, unequal and disproportionate current sharing in the input diode bridges will result, due mainly to the different choke sizes presenting different impedances. If drives without a D.C. link choke are to be connected in this way then individual A.C. supply reactors are required. If the drives of equal rating are similarly connected, the impedances seen on the input stages are very similar and The attraction of such a system is that it can, in principle, be realised using a standard, four-quadrant, D.C. drive converter such as the Control Techniques Mentor. There are however limitations. First, standard D.C. drives can typically only regenerate up to 1.15 times the r.m.s, supply voltage, making this the upper limit for the D.C. link voltage. This link voltage is lower than the standard rectified supply voltage of v/2 times the r.m.s. supply voltage, and consequently the available base speed is reduced to 1.15/v/2 = 81% of normal. This can be addressed by feeding the converter from a step-up transformer.

braking resistor and thermal overload

largedriveA +D.C.


[--I I rq I Fq I {




I r



h I


~ d

II h




Figure 11.15 The mains converter in one drive (A) supplying all drives (B, C etc.) from its D.C. link

C h a p t e r 11.10




!--I !---]

I drive A I


I drive B I

I--I I drive C [ !---i

Figure 11.16 Hard paralleling of all drive input rectifiers

il il il




precharge resistor

IGBT bridge
internal D.C. link capacitor connected via +/terminals to D. C. bus


Lf Lin

7 i, l



Figure 11.17 A four-quadrant PWM converter feeding the D.C. bus (standard Unidrive parts are enclosed by the dotted line)

Selection of the required D.C. line inductance is important in order to ensure that the time constant of the D.C. link is within the control capability of the drive 'speed loop' and ensures stable operation under light load conditions. As the D.C. link capacitor current ripple is directly related to the lifetime of that capacitor, it is important to keep it within its design limits. It is easy to see that, what sounded like one of the simplest forms of implementing a common D.C. bus solution, requires a significant amount of detailed engineering with intimate knowledge of both the A.C. and D.C. drive designs. Such a solution must therefore be undertaken in collaboration with the drive manufacturer.


Figure 11.17 shows a PWM converter suitable for feeding a common D.C. bus (or a single drive). The standard Control Techniques Unidrive has been designed to operate in such an

arrangement. The mains supply is connected to what would normally be the motor terminals (UVW) through the input inductors, L;,, and filter, LUand CU,as shown. The D.C. bus connections are made to the +V and - V terminals. Another single inverter drive, or multiple inverter drives, can be connected across +V and - V to produce a four-quadrant drive system with control of one or more motors. Input inductors must be provided to give a known minimum amount of source impedance, to allow the inverter to operate as a boost converter and to limit the PWM switching frequency related currents to an acceptable level for the converter. Increasing the value of inductance will reduce the PWM ripple current but will also have the undesirable effect of reducing the current-loop bandwidth and the power factor. A supply filter (formed by LUand Cf) may be required to further attenuate switching-frequency-related distortion so as to meet any applicable standards on mains supply harmonics, or to prevent other equipment connected to the same supply from being affected by high-order harmonics. When A.C. power is first applied the D.C. link capacitance is charged through the inverter antiparallel diodes and precharge resistor. The resistor prevents excessive charging






OF A . C .





current which could damage the diodes and potentially trip any input protection. Once the D.C. link capacitor is charged the contactor is closed to short circuit the resistor for normal operation. Note that this contactor is for precharging and is needed in addition to standard control switchgear. This form of supply gives excellent regulation of the D.C. bus even in the event of a transient condition such as a highspeed motor reversal, as shown in Figure 11.18. The control system is seen to limit the change in D.C. link voltage to less than 5 per cent with very rapid changes in power flow. In overload conditions where the D.C. link power exceeds the maximum A.C. power, the PWM rectifier will be forced into current limit. Figure 11.19 shows transient overload operation where the PWM rectifier goes into current limit; however, the system remains stable and the D.C.

link voltage deviates from the set point until the load is reduced. As well as offering a very well regulated D.C. bus, which will provide long D.C. bus capacitor life, the PWM is an elegant solution for applications requiring regeneration of energy back into the supply. It benefits from low harmonic distortion on the mains, although careful system design is necessary to ensure that high (switching) frequency harmonics do not cause interference to other equipment.


If one EMC filter is used for the complete system it is important to note that it needs to be rated for the total drivemotor cable length.










..................................... . . , . ,~ ................. ....... ,................l............... , ...... ..................... t ............................... '...... ........................................................


. . . .

Figure 11.18 D.C. voltage regulation during a high-speed motor reversal; VD.c. (D.C. link voltage) = 180 Vldiv, time = 200 msldiv



i"'..... 1,./lift.,
700 V
_: ......... ....

t /







Figure 11.19 D.C link voltage during transient overload; I/D.c. (D.C. link voltage)= 100 V/div, time = 20 ms/div

Chapter 11.11




Many mechanical drive trains, both fixed speed and variable speed, experience vibration. As operating speeds and controller performance continue to increase and motor mass and inertia fall, the danger of resonance problems increases. This subject area is complex and this description will be limited to an overview of the principles, and identification of the key sources of excitation of mechanical resonances. The vibration level of a mechanical drive train (motor, coupling, load etc.) is the result of imposed cyclic forces either from residual imbalance of the rotor, or from some other cyclic force and the response of the system to these forces. Problems tend to exist in one of three categories of application: High-power, high-speed applications, where operation above the first critical speed is required. (ii) Applications where torque tipple excites a resonance in the mechanical system. (iii) High-performance closed-loop applications, where the change of motor torque can be very high, the shaft linking the mechanical parts twists and the control loop sustains the vibration. Although the torque tipple produced from modem variablespeed drives is small, in comparison with earlier technology harmonic torques are produced. It is natural, therefore, to initially conclude that, in a system employing an electrical variable-speed drive, torque ripple is the problem- this is in practice rarely the case. It is important to recognise that any system in which masses (or inertias) are coupled together via flexible elements is capable of vibrating. Even a simple motor and load can be seen to be a two-mass torsional system. Consider the following completely general case. For the two inertias shown in Figure 11.20 the equations of motion, if zero damping is assumed, are:
.i, ( d 2 0 , / d ? ) + x(O, = 0

Figure 11.20 Two-mass torsional system where a;,,tf=torsional natural frequency (rad -1) and fntf= torsional natural frequency (Hz), and: K - torsional stiffness
= aJp/L


where G - shear modulus of elasticity, J p - polar moment of inertia of the shaft = r4/2 for a circular shaft of radius r and L - length of shaft being twisted.

An A.C. motor of inertia 0.5 kgm 2 is coupled to a load of inertia (0.4 kgm 2) via a shaft with a torsional spring constant of 60 x 103 Nm rad- 1:
f~tf = (1/27r)v/[kt(J1 + Jz)/(J1 x J2)]

= (1/27r)v/[60 x 103(0.5 + 0.4)/(0.5 x 0.4)] = 82.7 Hz The system designer needs to be able to calculate the source and expected level of the forces in the system, and design each component accordingly. Some of the important principles and factors to be considered are detailed below.

J2(d20:/dt 2) + X(O

- 0,) = 0


Consider first the causes of vibration that are entirely independent of a variable-speed drive. It is helpful to consider these in the following categories" subsynchronous vibrations (vibration frequency below the shaft rotational frequency) synchronous vibrations (vibration frequency at the shaft rotational frequency) super-synchronous vibrations (vibration frequency above the shaft rotational frequency) critical speeds

Eliminating 02 or 01 gives the general solution:

0 -- A + Bt + C COS(03ntf -J- (/))

(.tint -- 4 [ g ( J 1 f or --~ J2)/(J1 x J2)]

fntf = (1/27r)v/[K(J1 + J2)/(J1 J2)]



Subsynchronous Vibrations
The most common cause of subsynchronous vibration is in induction motor systems where beating at slip frequency, or multiples thereof, can occur. This is due to the fact that all electromagnetic forces in an induction motor occur at frequencies equal to or a multiple of the supply frequency. The rotational speed, however, is slightly less and mechanical imbalance forces will be cyclic at this reduced frequency. This is a classic example of two cyclic forces at relatively close frequencies combining to give a low-frequency beat. In the fault condition of an induction motor with a broken rotor bar, the vibration at slip frequency will dramatically increase.

require careful consideration and detailed analysis to avoid problems. The critical speed of a shaft is not solely dependent upon the characteristics of the shaft alone, but is greatly affected by the stiffness of the bearing supports. The actual magnitude of the shaft and bearing housing vibration is dependent upon the resonance curve for the shaft; the closer it is running to the critical speed the higher the vibration level. Reference to the motor manufacturer must be made if there is any concern about operating near the critical speed of the motor. It is not common for manufacturers to publish critical speed data. It is common for maximum motor speeds to be published, and they provide comfort for the vast majority of applications (Table 11.12).

Synchronous Vibrations
The most likely cause of vibrations at the shaft rotational frequency is mechanical imbalance or shaft misalignment. Mechanical imbalance may be simply a specification/manufacturing quality issue. It is surprising how common it is for confusion as to whether the motor is balanced with the shaft key fit or not or a half key. ISO 8821 has led to a broad acceptance of the half-key convention. The definition of a half key recognises that a key profiled to fill the whole volume of the keyway is often impractical and allows a fulllength rectangular key of half height or a half-length key of full height, the latter to be centred axially in the keyway. Imbalance may also be due, in larger motors, to shaft bending owing to unequal cooling or heating of the rotor. Thermal problems are usually time dependent, and are characterised by a gradual change in vibration with time and/ or load. Offset rotor or an elliptical stator bore will both result in cyclic magnetic forces at shaft rotational frequency. If the system inertia is sufficient and it is possible to disconnect the motor supply, this problem will suddenly disappear.


Torque ripple is inherent in almost all electrical variablespeed drives. The frequency and magnitude is dependent upon the type of converter and control applied. Consider first the D.C. motor fed from a six-pulse converter (so named because the resultant motor armature current has ripple comprising six peaks for every cycle of mains frequency). On a 50 Hz mains supply this ripple has a fundamental frequency of 300Hz (60Hz=~360Hz). In a separately excited D.C. machine the torque is proportional to armature current, the torque ripple has a frequency of sixtimes mains frequency. The magnitude of this ripple is typically in the range 10-20 per cent rated torque. The frequency of the torque ripple is a function of the mains frequency and is independent of operating speed, so provided 300 Hz (360 Hz) is not close to the natural resonant frequency of the mechanical system no problems should result. Torque ripple due to the commutation process in the motor is also important where a small number of commutator segments are used. This component of torque ripple has a frequency which is proportional to speed. Consider now the situation with a PMW inverter system. Although the theory of torque ripple calculation is complex, torque ripple at six and twelve times the output frequency of the drive are of most practical importance. The magnitude of the torque ripple is dependent upon the magnitude of the current harmonics which is in turn very dependent upon the drive and control type and the demands of the application. It is not possible to give exact calculations of torque ripple but a general comparison by drive type is helpful. It is interesting but of limited value to consider the mathematics behind this form of resonance. Consider again the equations of motion for a two-mass system described in the introduction to this section; as we are

Super-Synchronous Vibrations
Super-synchronous vibrations tend to be associated with outof-roundness bearings or shaft asymmetry along its length due, for example, to a single large keyway. It is unusual for either of these issues to present a practical problem. Wear or damage to roller bearings could be a problem and should be considered. Initial bearing problems will result in only a small increase in bearing vibration, but as wear increases sudden catastrophic failure of the bearing would occur leading possibly to shaft or even stator core damage.

Critical Speeds
As the physical ratings and, through the use of variablespeed drive systems, operating speeds increase, motors are being designed for application above their first (and in some cases even their second) critical speed. Such applications

Table 11.12 Maximum motor speeds for the Leroy Somer MV induction motor (two, four and six pole)

Frame size Maximum speed (min-1)

80 15 000

90 12 000

100/112 10000

132 7500

160 6000

160LU/180 5600

200 4500

225/250 4100

280 3600

315 3000

Chapter 11.11
Table 11.13 Comparison of torque ripple by drive type
Drive/motor type D.C. drive Torque ripple level a high b very low PWM-fed induction/PM motors SR/stepper very low high Source of ripple a armature current ripple b commutator interaction of flux harmonics with fundamental current torque waveform inherently pulsed Frequency of ripple a 6 mains b speed segments rev - 1


subharmonics of synchronous speed harmonics of number of steps per revolution

considering torsional vibrations considering the relative displacement of one body to the other, we obtain:

System Control-Loop Instability

Instability of this type is where the change of motor torque can be very high, the shaft linking the mechanical parts twists and the control loop sustains the vibration i.e. the control loop acts as a positive feedback to the vibration rather than negative, damping feedback. As a further complication in high-performance closed-loop systems, it needs to be remembered that the speed/position feedback device constitutes a further third mass in the drive train. This is practically important in some applications and can limit torque bandwidth performance to below 1 kHz. This form of instability is highly complex in nature and, although practically important, is beyond the scope of this book.

J(dZO/dt 2) + KO = 0
where J is the total inertia of the system. If we have a driving force r(t), i.e. a force which is time variant, we can obtain the differential equation of the new system by adding the driving force to the above equation:

J(dZO/dt2) + klO

= r(t)

If we consider the simple case r(t)= F cos a;t, then:

O(t) - k2 cos(a;ot - 6) + {FIk~ [1 - (co/COo)2]}cos cot

The key issue to note is that this represents a superposition of two harmonic oscillations. The frequency of the first is the natural frequency of the system a;o/27r, the second a;, the frequency of the driving force. By inspection the amplitude, p, of the oscillation at the driving frequency depends upon a; and COo.As a; ~ aJo the amplitude, p, tends to infinity. This is as expected because, when the forcing frequency is the same as the natural frequency of the system, resonance occurs.


Some vibration of a motor drive train is inevitable, and provided that resonances, critical speeds etc. are avoided, it can be tolerated. In such systems steps can be taken to reduce vibration and the resulting noise. The primary measures fall into four categories: a Improve balancing and stiffening to reduce the amount of vibration generated and ensure proper alignment of all rotating parts. Resonance may also occur at certain oscillation frequencies in the surfaces of speed-controlled machines. An example of this may be a tie bar used to link the end frames of a motor. Simple measures of applying an intermediate node would resolve any possible problems. Use isolation to prevent vibrations being transmitted. A wide variety of isolators is available including simple rubber mats and custom machine shoes through to specialised dampers offering specific stiffness in various directions. Use of appropriate nonlinear, detuner-type torsional couplings Build robust foundations. The dimensioning of foundations is critical to ensuring that vibrations from the machine are not transmitted to the structure and that no resulting damage occurs. It is not possible to include the detailed design calculations for foundations here. It is important to recognise the vital nature of foundations to a successful installation.


A rule of thumb which is contained in EN 61800-2:1998/IEC 61800-2:1998 is that to avoid vibration during operation, the speed controller of a drive must be tuned to a value such that: b

fn >> 10/rR
where TR is the requested response time of the speed controller i.e. the time required following the initiation of a change in demand for an output going in the correct direction. This is a very global, somewhat imprecise and certainly conservative rule of thumb, and in that context it is reasonable to assume that TR is the time to reach 80 per cent of the demanded speed following a step change in demand. c d




m m /

1 2 3


259 264 288

It is not practical to describe all possible applications and/or characteristics for electrical variable-speed drives. This chapter aims to provide an insight into some of the possibilities/opportunities. Typical characteristics are covered and techniques applied in many different applications are described. Then, a large number of examples of actual

applications are covered in varying degrees of technical depth. This chapter is not intended as a summary of what is possible, rather a sample of what has been done and the fundamental applications techniques which have been applied.


In order to successfully select and apply the optimum drive system, it is necessary to understand the essential features of both the alternative drive technologies and the load to be

driven. The following listings of common loads could prove useful when selecting a drive.



See Table 12.1.

Table 12.1 Typical load characteristics and ratings for the metals industry
Drive duty Rolling mill Strip mill Slitters and perforators Wind/unwind Rating range up to 1000s of kW up to 100s of kW 50 to 150 kW Comments/drive type high-impact torque loading, constant kW speed range, steel works specification and difficult environment; closed loop induction motor drives and D.C. drives. normal torque loading (150% maximum) constant kW speed range, steel works specification and difficult environment; closed loop induction motor drives and D.C. drives. in metal finishers' plant, environment and specification easier - normally IP23 enclosure, forced ventilation with filter is acceptable; drive often integrated with wind/unwind stand drives; closed loop induction motor drives and D.C. drives. constant kW rating over build-up range; regenerative braking with four-quadrant operation; steel works specification and difficult environment; in metal finishers' plant easier conditions apply as above; closed loop induction motor drives and D.C. drives. constant kW rating over pipe diameter range; usually four-quadrant with regenerative braking; environment can be difficult with oil spray present; closed loop induction motor drives and D.C. drives. high values of acceleration and deceleration torque required; four-quadrant regenerative, typically four to six speeds required: clean, spray, fill, spin 1, spin 2; difficult environment; single pipe vent or box-enclosed motor with filtered air supply; closed loop induction motor drives and D.C. drives. mostly flange mounting and timer belt drive to 30 kW gearbox coupled above; always field range control, reversible, four-quadrant drive; often with encoder for spindle orientation; forced-vent with filter; often coaxial fan unit to 60 kW; closed loop induction motor drives; at low powers permanent magnet servo drives and at very high powers D.C. drives are used. foot or flange-mounting gearbox otherwise as above; permanent magnet servo drives; at higher powers closed loop induction motor drives and D.C. drives are used. constant kW speed range with individual motor field control on multiblock drives with single controller; progressive speed increase between heads as wire diameter reduces and speed increases; dancer arm