Documentos de Académico
Documentos de Profesional
Documentos de Cultura
Development Guide
A Comprehensive Handbook
for Test Engineers
Section 2. 13. Using SCPI and Direct I/O Section 3.
Networking Choices 93 vs. Drivers 117 LXI: The Future of Test 141
Introduction 117
Overview 93 Deciding how to communicate 117 Overview 141
Sketching the big picture 117
9. Using LAN in Test Systems: Achieving communication 118 16. Value, Performance and
The Basics 95 Exploring the application alternatives 120 Flexibility: The Promise of
Introduction 95 Maximizing performance and LXI 143
Coping with complexity 95 flexibility 121 Introduction 143
Setting the standard 96 Assessing I/O software alternatives 123 Why test managers are asking for a new
Using LAN in test systems 98 Conclusion 124 approach 143
Conclusion 100 Addressing the challenges with LXI 144
14. Using LAN in Test Systems: The advantages of LXI 146
10. Using LAN in Test Systems: Applications 125 A closer look at LXI 150
Network Configuration and Introduction 125 Exploring new possibilities with LXI 153
Basic Security 101 Scenario 1: Sharing instruments 125 Appendix 16A: Defining synthetic
Introduction 101 Scenario 2: Remote monitoring and data instruments 155
Creating a safe haven 101 acquisition 127 Appendix 16B: Creating cost-effective
Understanding the pitfalls 101 Scenario 3: Functional test systems 130 measurement solutions with Agilent
Designing the private, protected LAN 102 Configuring a VPN 131 Open to test your way 156
Conclusion 105 Comparing network performance 133
Conclusion 134 17. Transitioning from GPIB
11. Using LAN in Test Systems: to LXI 157
PC Configuration 107 15. Using LAN in Test Systems: Introduction 157
Introduction 107
Setting Up System I/O 135 Comparing system architectures 157
Creating the right environment 107 Introduction 135 Setting up an LXI system 159
Exploring network settings in Windows XP Simplifying LAN-based instrument Simplifying software changes 160
and Vista 107 connections 135 Conclusion 161
Using multiple network connections 108 Assessing the Agilent IO Libraries
Managing IP addresses 108 Suite 135 18. Creating Hybrid Test
Configuring LAN with Agilent IO Libraries Connecting instruments to LAN 137 Systems with PXI, VXI and
Suite 109 Conclusion 140 LXI 163
Conclusion 110 Introduction 163
Assessing modular systems 163
12. Using USB in the Exploring LAN-based hybrid systems 165
Test and Measurement Going beyond hybrid to all-LXI 167
Environment 111 Conclusion 168
Introduction 111
USB in the PC universe 111
19. Assessing Synthetic
Agilent support for USB instrument Instruments 169
connectivity 112 Introduction 169
Setting up USB instruments with the Reviewing the roots of SI 169
Agilent IO Libraries 113 Putting SIs in perspective 170
Conclusion 116 Comparing present and future
approaches 170
Exploring the initial applications 173
Utilizing current SI devices 174
Conclusion 176
Contents
Section 4. 22. Calibrating Signal Paths
RF/Microwave Test in RF/Microwave Test
Systems 177 Systems 193
Introduction 193
Overview 177 Understanding vector and scalar
calibration 193
20. Optimizing the Elements Performing vector calibration of network-
of an RF/Microwave Test analyzer paths 195
System 179 Performing vector calibration of non-
Introduction 179 network-analyzer paths 196
Letting the DUT define “future” 179 Performing scalar calibration of non-
Reviewing some essential network-analyzer paths 197
considerations 180 Conclusion 198
Translating requirements into optimized
equipment choices 181
Glossary of Test-System
Pulling it all together 184 Development Terms 199
www.agilent.com/find/open
Introduction
Introduction
Section 1. Test System Design
Section 1. Test System Design
1. Introduction to Test-System Design
Introduction Test can be used simply as a gating Excel is the most common software
factor for “good” or “bad” devices, or analysis tool for the R&D engineer.
This chapter offers an overview it can be used to gain a competitive The connection is usually simple:
of the process of designing test advantage. This chapter offers an a PC connected via GPIB or USB
systems, beginning with a discussion overall view of how tests are made, to an instrument or a small set of
of how carefully designed systems techniques to optimize tests, and a instruments. Simple software, such
can transform test into a strategic number of methods you can use to as Agilent IntuiLink, finishes the
competitive advantage. The chapter your advantage. It covers the three connection.
then walks you through the key primary sectors of the product life
factors to consider when designing cycle that require test: R&D, design Once the design becomes more solid,
a test system, choosing the level of validation, and manufacturing. there is a need to find its limits and
automated control, and planning Other chapters cover such topics weaknesses. That’s where the design
for future needs. It concludes with as hardware architecture, choosing validation system comes in. To make
a comparative case study of testing instruments, software architecture, the results more repeatable and less
power supplies using manual, semi- computer I/O and connectivity, dependent upon operator expertise,
automated and automated control. assembling a test system, maximizing the test system is automated using a
throughput, and optimizing deploy- PC and some sort of graphical soft-
Transforming test into a ment and maintenance. ware such as Agilent VEE or National
Instruments LabVIEW.
strategic advantage A systematic test-system design
process as outlined in this guide Graphical software, often used for
Functional test is fundamental to will assist you to quickly design a design validation testing, gives the
the electronics world. In the past, test system that produces reliable engineer a more comprehensive set
test has been treated as a necessary and repeatable results, meets your of tools for control and analysis,
expense, but enlightened companies throughput requirements, and does while at the same time creating
have realized that test can be a signif- so within your budget. For further a more repeatable measurement
icant asset. A test system can be used information regarding test-system process that may include remote
for far more than simply verifying the design, you can refer to the book control of sources, measurements,
limits of the device under test (DUT). from which much of the informa- and system switching. The same
Consider these possibilities: tion in this chapter was derived: instruments used in the R&D bench
Test-System Design, A Systematic system are often used in design
• find the weaknesses of the device— validation. This gives continuity to
Approach by Tursky, Gordon, and
before your customers do the whole process, so that the initial
Cowie (Prentice Hall, 2001).
R&D measurements can be compared
• predict failures or out-of-spec
The earlier a product weakness is to those made for design validation.
trends in production
discovered, the less expensive the
consequences. That’s one reason Textual software generally provides
• search for the boundaries of the
why the role of test changes with an effective programming environ-
design—to stretch specifications
the stage of the product life cycle. ment for manufacturing test, as
or search for something you didn’t
When a product is first developed, it enables the engineer to extract
know the product could do
the role of test is to verify that the the highest throughput from the
• verify the long-term characteristics design concept is viable. This calls test system. In manufacturing,
of the product for quick measurements, usually with repeatability and reliability become
hands-on use of discrete test instru- paramount concerns. Again, if the
• optimize a production process ments. Sometimes there is a need to same equipment can be used for all
load measurement data into an Excel three test situations (R&D, design
• test for environmental limits validation, and manufacturing), then
spreadsheet for use in a lab report or
for further analysis. the R&D engineer can more readily
• find the weaknesses in a
assist with any problems that may
competitor’s product
arise during manufacturing test.
The process of designing and Planning your test system • physical constraints such as size,
integrating systems used for elec- environment, and available power
tronic test requires more than simply Creating a comprehensive test plan
coding instrument commands to allows you to take a big-picture • heat buildup and power dissipation
automate the measurements made view of the project and forces you to
• how the test system will be veri-
on the R&D bench. The instruments focus on meeting the objectives and
fied, maintained, and calibrated
are only one part of the complete test requirements for the test system. The
system; cables, software, test-plan result is a considerable time saving • RF environment
documentation, and fixturing are in the development process.
equally important. The latter are • accuracy and resolution
especially prevalent in a manufac- Even in the R&D environment, there requirements
turing environment. are times when it is useful to create
a test plan, so that you can docu- • throughput goals
Test-system ment and compare results after each
design cycle. You must also consider • development time constraints
considerations the future for any test system you
create today. It may be reasonable • software-development and runtime
There are many factors to consider to create a dedicated and somewhat environment
when developing a test system. The inflexible test system on some
three main driving factors are test • cost constraints
high-volume projects, but it is usually
requirements, development time, more appropriate to create a system • continuity constraints with
and test cost. The factor that is most that has the flexibility to adapt to existing legacy systems
important will drive the other two. future needs.
For example, if the test requirement Among the decisions involved in
is for a very accurate measurement, The test plan describes more than determining the design of a test
as in R&D or design validation, just the requirements of the DUT. It system, the most obvious is what it is
you must be willing to take a bit should also cover other areas of the you must test. This is usually defined
more time to achieve the required test such as the level of experience in a test specification. The test speci-
accuracy. On the other hand, the required of the test system operator, fication should include a complete
manufacturing manager would not calibration and maintenance require- list of the product functions to be
be pleased if the test system were to ments, physical limitations, and verified, operating parameters to
perform more tests than required, throughput requirements. meet, and any regulatory standards
or perform them at a higher-than to adhere to.
needed level of accuracy, due to the The first step in creating a test
obvious impacts on test-system cost system is to seek out and compile all
and throughput. the information needed to create an Accuracy
overall test plan. Important informa-
Before the process to design a test tion includes the following: System accuracy is a critical speci-
system can begin, you must have fication of any test system, and the
a good understanding of the test • functional and parametric tests to overall test plan should include both
application. This goes beyond simply be performed the accuracy requirements of the test
understanding the device you are and the recommended margin. As a
testing, as you must also be aware • DUT design validation criteria minimum, the test equipment should
of other factors such as the skill have twice the accuracy specified for
• format and usage of test results, the DUT. To maintain this margin
level of the test system operator,
including sharing data throughout requires that the operating tempera-
the operating environment, and any
the enterprise ture be maintained closely and
standards requirements.
• number of tests that calibration cycles be followed
faithfully.
• DUT pin counts
www.agilent.com/find/open
11
Control decisions Once the testing becomes more The type of control, either manual,
complex or repetitive, a fully auto- semi-automated, or fully automated,
A major consideration for a test mated test system is in order. A fully should be determined early as it
system is the level of automation to automated test system takes care will influence which instruments
build into the system to control the of signal switching, measurement, you select. As shown in Table 1.1,
test process. Manual control requires recording, and even analysis of the many factors influence which control
that a human operator make all of the results for pass/fail determination. method is most suitable for your
test connections, set the instruments, Once the DUT is in the test fixture, application.
and then record the data. Increasingly, the test system takes over and runs
even in simple R&D setups, most all of the tests. This is the ultimate in
engineers prefer to use instruments terms of test speed, reliability, and
under the control of a PC in order to repeatability, but it is also the most
have a record of the test. expensive and time consuming to
develop.
Table 1.1. Comparison of test system control options
www.agilent.com/find/open
13
Semi-automated control A semi-automated test system might Semi-automated control is often
require the operator to manually much faster than manual control
Semi-automated control is a common connect the DUT, provide instruc- and produces a more reliable and
type of control approach used for tions to the operator for the proce- repeatable result. This method
test systems, and is useful in R&D, dural steps, and automatically record of control can take advantage of
design validation, and manufacturing the results. For example, a semi- simplified software development with
test (Figure 1.2). Test systems using automated system might have an Agilent’s VEE or Visual Studio .NET
this control approach have manual oscilloscope and an RF source that for quickly creating the required
portions for flexibility where it is are under computer control, with a automation.
needed and automation where it power supply under manual control.
makes sense. Those sections of the The engineer would vary the voltage The most common type of test equip-
test system that are expected to to the DUT via the power supply, run ment includes a fully functional front
change often or would be too expen- a set of tests at this voltage level, and panel and a computer interface that
sive to automate can be manual. then manually change the voltage allows both manual and automated
Those sections that will not change and run another set of tests. use. This is a major benefit, even
or would benefit from automatic data when automating, as you can always
recording can be automated. go back to a manual approach if you
need to measure other parameters,
troubleshoot the system, or conduct
Figure 1.2. A test system using semi-automated control often uses a an experiment. These standalone
PC for the operator interface.
instruments are beneficial when
developing a fully automated test
system for manufacturing as it is
common to start with a semi-auto-
mated system and then increase the
level of automation as experience
and production volume increases.
www.agilent.com/find/open
15
Planning for the future A few moments considering the Conclusion
future can have a significant impact
When making test-system design on future options. For example, when Although test-system development is
decisions, you should keep future selecting instruments for a manual a complex task that can include many
needs in mind. Upgrades are a fact system, there is usually very little aspects of electronic and mechanical
of life for a test system. They can be added cost to select instruments design, following a systematic
very expensive and time consuming that have computer interfaces. You approach and partnering with quality
but are often unavoidable. Naturally, may not need the interface today, test equipment manufacturers will
any upgrades must justify the expense but computer control is not possible enable you to enhance your success
and effort required. Reasons for without it (and could be costly, while lowering the cost and time it
upgrades include difficult, or even impossible to add takes to create the test system.
at a later date).
• accommodate changes in design
of the DUT Using open standards will increase
Case study: testing power
the likelihood that test system supplies
• conduct additional tests components will be useable in
the future. Proprietary interfaces This case study is an example of how
• obtain higher accuracy a test system can evolve from R&D to
have a habit of disappearing or
not supplying the drivers you need design validation to manufacturing.
• obtain higher throughput Many of the same instruments are
for future software options. Using
proprietary measurements made by used in all three areas with the major
• eliminate redundant tests
specific equipment in a test system difference being the type of control
• rearrange the test sequence to from manufacturers that do not used. This is a common practice as
detect failures earlier supply future upgrade paths could the knowledge gained in each phase
make an entire test system obsolete of product development is trans-
• improve analysis if that exact instrument is no longer ferred to the next.
available.
• automate more of the test
Consider where the instrument
Manual control
• decrease the skill level required
architecture is in its lifecycle. For When developing a product such as
to operate the test system
instance, is it a cardcage design a power supply, the R&D engineer
• replace obsolete equipment based on a PC backplane that will will create a test system as required
soon be replaced? Are vendors to explore options and verify results.
• change reporting requirements designing new products to this archi- The test bench in Figure 1.1 is
tecture (or to its replacement)? typical of such use. Many instru-
• upgrade the operating system
ments are within reach and it is easy
Following proper software design
• conform to new standards to rearrange them as needed. All
techniques resulting in well-written
of the connections to the DUT are
software that is easily understood,
• add newly developed models made manually and each instrument
maintained, and modified is an
is manually operated. This is an
• repeatability is important obvious requirement for future
example of a test system with manual
upgrades. Good documentation is
control.
also critical to the future of a test
system: Chances are you will not be
the one that is tasked with future
modifications.
www.agilent.com/find/open
17
Automated control The block diagram in Figure 1.5 The speed, repeatability, and reli-
includes connections to the DUT ability of the fully automated system
The move to a fully automated test and measurements that test the can be significantly better than that
system may require additional instru- power supply in the manufacturing of other test system control methods.
ments. The computer now controls environment. The number of tests Also, the skill level of the operator
all of the instruments as well as the performed may approach those can be less. But the time and expense
reconfiguration of the interconnec- conducted during R&D and design to create the system and make any
tions for various tests. The digital validation but they are normally not changes usually makes automated
multimeter, scope, and loads are still as thorough. Manufacturing tests are test systems only feasible for manu-
used, but now switches are employed often performed only at one oper- facturing uses.
to connect the DUT to the instru- ating point that is considered to be a
ments. As the tests are performed, worst-case condition. This maximizes
the computer uses the switches as the amount of information gained
required. about the DUT in the minimum time.
Fixture
MUX
Switch DC/DC
matrix power
source
(DUT)
DMM
Function generator
Oscilloscope
AC source
DC source
DC loads
Power switch
19
GPIB interfaces To communicate with instruments loaded, the instrument will be ready
over GPIB, you need to install an to communicate on the bus. USB
GPIB is the most common interface I/O software package. Plug and devices also are “hot swappable,”
for programmable test-and-measure- Play drivers, IVI-COM drivers, and which means you don’t have to shut
ment equipment. It is still one of the VISA (Virtual Instrument Software down your PC to plug in or unplug an
best choices if you want to maximize Architecture) are examples. These instrument.
throughput for a variety of block packages support popular languages
sizes. GPIB is a parallel bus that such as C and C++, Microsoft With USB, the computer schedules
includes control lines, handshake Visual Basic 6.0, Visual Basic .NET, and initiates all transactions. If you
lines, and 8 bi-directional data MATLAB, Agilent VEE, LabVIEW, are using a Windows NT® operating
lines—specifically designed for instru- and others. system, you will find that it does not
ment communications and control. support USB connections.
GPIB supports up to 14 devices that USB interfaces
can be connected to your PC. You
can use either a star or a daisy-chain USB was originally intended as Figure 2.1. You can configure a GPIB bus in
configuration for connecting multiple an alternative to the RS-232 serial either a daisy-chain or star topology, or you can
intermix these two configurations.
instruments (see Figure 2.1), but interface and the Centronics parallel
cable length is limited to 2 meters interface, an older standard I/O
(times the number of devices) up to a interface for connecting printers and
maximum length of 20 meters. certain other devices to computers.
USB is suitable for a range of
You can achieve data transfer rates of computer peripherals, from slow Instrument 1 Instrument 2
more than 500 KB/s on a GPIB bus if devices, such as mice and keyboards,
you limit bus cable length to 1 meter to high-performance devices such
(times the total number of devices), as scanners, printers, and cameras.
up to a maximum length of 15 Now, USB is finding its way into Instrument n
meters. Longer cable lengths reduce test-and-measurement instrumenta- Daisy chain bus configuration
the maximum data transfer rate to tion, too.
less than 500 KB/s.
USB is a serial interface bus that Instrument 1 Instrument 2
When you use GPIB, configuring the includes two power wires and a
instrument I/O bus is a relatively twisted pair to carry data. USB is
easy task. However, each instrument capable of data transfer rates of
on the bus needs to have a unique about 12 Mb/s for v1.1, and up to
address. This requirement means 480 Mb/s for v2.0. In addition, v2.0
you may have to manually change is fully backward-compatible with
an instrument’s address when you v1.1. The main difference is the data
configure your system. transfer rate. Instrument n
GPIB has other drawbacks, too. GPIB USB is capable of supporting up to
cables and connectors are rather Star bus configuration
127 devices on a given interface.
large, bulky, and relatively expensive. If you use a GPIB-based system,
And because GPIB isn’t a standard you must ensure that instrument
built-in PC interface, you have to addresses are unique, but USB
open your PC housing and install an provides this function automatically.
interface card in one of your PC’s When USB devices are manufactured,
expansion slots. they are given unique identifiers
based on the manufacturer, the
instrument serial number, and the
product number. When the device
is powered up and connected to a
controller, the controller detects
its presence automatically, and if
the host-side software drivers are
www.agilent.com/find/open
21
Connection methods Remote access Figure 2.3. Single and multiple instrument configurations
can be connected to private LANs and site LANs.
You can connect LAN-enabled instru- A site LAN has the potential for
ments several different ways. They permitting any controller on the LAN
may be connected directly to a site to access instrumentation—either
LAN (a workgroup LAN, intranet, intentionally or unintentionally. If
or enterprise LAN), or they may be the site LAN can be accessed from Crossover cable
Instrument
connected to a private LAN. physical locations outside of your
facility, then others can access your
In private-LAN configurations, your Simply connected private LAN
instrumentation. This open access
PC and your test instruments are can be a valuable asset because it lets
connected to each other via a LAN, you remotely control instruments
but they are not connected to a site and systems almost as easily as if
LAN. The simplest private-LAN To site LAN
they were next door. You can use
configuration consists of a controller Instrument
remote access capability to diagnose
and only one instrument. See the first system and instrument faults from
illustration in Figure 2.3. You also Simply connected site LAN
faraway locations. Multiple engineers
can connect multiple instruments can share the expensive test instru-
in a private LAN, as shown in the ments and systems from remote
second illustration in Figure 2.3. locations.
If you plan to use your site LAN, However, this open access also can be
rather than a private LAN, you a disadvantage. For example, if the
need to be aware of two potential site LAN is connected to the outside Switch
drawbacks: world to provide Internet access, you
face a serious risk of exposure to
1. Traffic on your site LAN can slow
undesired system accesses. Firewall Instrument 1 Instrument 2
down your measurements.
software and/or using a router which
2. If you are using a LAN interface requires specific device addressing Multiple instrument private LAN
for controlling your test system, rather than a switch or hub can
it is possible that a faulty instru- provide protection.
ment could damage or disrupt
If you want remote access to your
the network, particularly when To site LAN
test equipment, but security and
the instrument is turned on ROUTER
controlled access are a system
and tested for the first time.
requirement, then you need a lockout
Controlling your test instruments
feature. Some instruments, such as
via a private LAN is the safest
the 33220A function/arbitrary wave-
approach, since it limits the range
form generator, provide this feature Instrument 1 Instrument 2
of potential disruption and access
via an Allow List. An Allow List is
and maximizes performance.
a list of remote LAN addresses that Multiple instrument site LAN
For all setups, you can connect are permitted to communicate with
instruments to the LAN either with the instrument. Any controller that
a conventional LAN cable or through attempts to access an instrument
a wireless adapter. Wireless routers whose address is not on the Allow
and hubs also are available, as are List is rejected. This feature provides
wireless USB-to-LAN interfaces. See a level of system security for those
Application Note 1909-3, Creating instances where your system is
a Wireless LAN Connection to a connected to a site LAN and is at risk
Measurement System. for inadvertent access.
www.agilent.com/find/open
23
These variables make it difficult The time to close the channel and Comparing costs
to predict the actual data transfer make the measurement dominates
rate that might be expected for any the total time. GPIB’s strong perfor- Today, many companies are looking
given system configuration. Table mance in this use model is one of the for ways to lower the cost of test.
2.1 shows a relative comparison of reasons it has lasted so long as an If this is true of your organization,
data transfer rates for several data interface. implementation cost will be an
block sizes among GPIB, USB v1.1, important consideration in selecting
USB v2.0, and LAN interfaces. These For a detailed look at data transfer an I/O interface for your test system.
data were compiled using the Agilent rates of two different block sizes
Model 33220A function/arbitrary over the various interfaces, see New PCs typically have a LAN and
waveform generator and a Hewlett- Application Note 1475-1, Modern several USB ports built in, but GPIB
Packard Kayak PC with an 800 MHz Connectivity—Using USB and LAN interfaces usually require a card that
processor running on a Windows XP Converters. This application note you must purchase separately. GPIB
operating system. compares the Agilent 82350B GPIB cards typically cost about US$500
PC card, the 82357A USB/GPIB and additional USB or LAN cards
For small data-block sizes of a few converter, and the E5810A LAN/ usually sell for US$10 to US$50.
hundred bytes, there is no appre- GPIB gateway in terms of controller
ciable difference in bus speed, but and operating system requirements, Also, if you plan to use USB or
the higher-performance buses (USB set-up steps, data transfer rates, LAN interfaces to connect multiple
v2.0 and LAN) demonstrate a marked allowable distances from instruments instruments in your system, you
improvement in the time required to to the PC, etc. These details will help will need switches or hubs. These
transfer large data blocks. you choose the best interconnection hubs can cost from US$25 to US$200
method for your application. One of each, depending on features and the
The differences in data transfer rates the benefits of having an instrument number of ports they support.
between small and large data blocks that supports multiple interfaces
for any given interface are largely You also need to consider the cost
is the ability to easily compare the
due to variations in the latency, or of the cables for your test system.
actual data transfer rate for each of
software overhead, required for each GPIB cables are relatively expensive,
the I/O interfaces in a given applica-
of the interfaces prior to the start of ranging in price from US$60 to
tion. This permits you to select the
the actual data transfer. US$150 each, depending on their
interface that offers the optimum
length. USB cables, on the other
performance.
If you’re looking for high throughput hand, range from US$8 to US$30.
in a test system, don’t be swayed If the application program’s LAN cables are usually the least
by the perception that high-speed I/O calls are written with a driver expensive, typically costing less than
interfaces will always get you there. interface that provides a common US$10. Some can be found for as low
In most test systems, the use model set of programming commands as US$3.
is one of “Close a channel; measure a independent of the interface, such as
point,” then “Close another channel; You can make useful cost compari-
Agilent’s VISACom, then it becomes
measure another point.” In this case, sons by assuming that all test
a simple matter to direct the I/O calls
block transfer rate is meaningless. instruments are able to support any
to any of the three interfaces.
of the three interfaces and computing
the interface cost for your proposed
test system. Today, few test instru-
Table 2.1. Relative I/O times from a PC to an Agilent 33220A (data taken with a 1-meter cable
ments actually do support all three,
on an HP Kayak XU800 with an 800 MHz processor running Windows XP)
since the industry is just beginning to
Interface Function change Frequency change 4K arb 64K arb provide instruments equipped with
multiple computer-industry-standard
LAN (socket) 100 ms 3 ms 8 ms 110 ms interfaces. However, the I/O inter-
USB 1.1 100 ms 4 ms 10 ms 185 ms face converters mentioned earlier
USB 2.0 99 ms 3 ms 8 ms 100 ms permit GPIB-only instruments to be
connected to USB- and LAN-based
GPIB 99 ms 2 ms 20 ms 340 ms interfaces. For example, the Agilent
82357B USB/GPIB interface enables
your PC to communicate with GPIB
devices via the PC’s USB port.
Table 2.2. Typical costs for LAN, GPIB and USB interfaces
www.agilent.com/find/open
25
Conclusion For design verification and manufac-
turing, USB and Ethernet-based LAN Get help configuring your
With the new generation of test are good choices, although LAN is I/O interfaces
instruments offering a choice of typically the better of the two alter-
interfaces, you need to decide which natives for larger systems because Configuring an interface to connect
interface is best suited for your of its data-throughput performance, your PC to an instrument or system
test system. Comparing costs, data cost, remote access, and ease of
can be a daunting task for someone
transfer rates and ease of imple- system assembly.
mentation will help you choose the who is not well versed in the
interface most appropriate for your The added flexibility, remote system intricacies of PCs, I/O technologies,
application (see Table 2.3). For R&D access and control, performance on a and I/O inter-face configuration.
applications, where the number of par with USB, captive cable connec-
In the past, this was especially so
instruments in a system is usually tors (which aren’t found on USB),
and the capability for wireless opera- for LAN-based I/O that required a
small and a quick and easy interface
tion offered by the LAN approach can system to be connected to a site
set-up is desired, USB is usually the
best choice. make LAN the most attractive choice LAN. Fortunately, step-by-step guides
for many systems applications. such as Agilent’s USB/LAN/GPIB
Interface Connectivity Guide are now
Table 2.3. Advantages and disadvantages of GPIB, USB, and LAN interfaces available to help you to configure
your I/O interfaces. The Connectivity
Interface Advantages Disadvantages Guide describes in detail how to
LAN • Good data-throughput performance • Requires LAN knowledge to
connect instruments to various
• Low cost set up
interfaces, and how to configure your
• Remote access makes it easy to PC. It also includes programming
control system from remote location
examples and interface trouble-
USB • Quick, easy setup • Does not work with Windows
shooting tips. You can view the
• Low cost NT
guide at http://cp.literature.agilent.
• Good data-throughput performance
com/litweb/pdf/E2094-90009.pdf
GPIB • Ubiquitous interface on test • PC expansion slot required
instruments • Must open PC housing to install
• Maximizes throughput for all block card
sizes • Relatively expensive
• Limited cable lengths permitted
between computer and
instruments
27
conflict. The first products to use to generate a universal approach to One advantage of SCPI is that the
this bus were released in 1972. The instrument control, using TMSL as list of commands typically covers
same year this new bus was dubbed the basis. The outcome was Standard 100 percent of the instrument’s
Hewlett-Packard Interface Bus (HP- Commands for Programmable programmable functions, no matter
IB). In 1975, IEEE adopted it as a Instruments (SCPI) (Figure 3.2). how arcane. For a friendly tutorial on
standard with little modification, and SCPI, go to ftp://ftp.agilent.com/pub/
IEEE-488 was born. A variant of the Today, SCPI is still the most-used mpusup/pc/iop/hpibtut/ib5_scp.html.
original interface is now popularly form of instrument control. In
known as General Purpose Interface SCPI, the instrument programming
Bus (GPIB). syntax became much more robust The I/O software: SICL and
and predictable. SCPI defined a VISA
With GPIB and a desktop computer strict hierarchy, and every command
(actually at the time it was called was associated with a concomitant Instrument commands aren’t the
a “desktop calculator”), the need response. These were defined for whole story. It takes more “layers”
arose for a common communication source, sense and switch devices. of software to communicate with
language. Limited processing power Here’s an example of SCPI code: a computer. Before you send the
in the ‘calculators’ demanded a instrument a command, you need
simple syntax, so ASCII commands CONF:VOLT:DC 0.3,0.003 to define the I/O path, route the
were chosen. A DMM might be sent information through the proper I/O
This command tells the instrument card, find out where the instrument
what was affectionately termed
to configure itself to get ready to read is on the bus and speak to the instru-
“R2D2 code.” Here’s an example:
a 0.3 volt dc signal with 3-millivolt ment in the syntax of the I/O you’re
“F1R2T1” resolution. It should be obvious from using. Assuming the GPIB I/O card in
this statement that SCPI commands the computer is at address 7 and the
The command means “Go to the require some intelligence on the DMM is at address 22 on the bus, the
dc volts Function, the 1 volt Range other end of the wire, as not every simple BASIC command might be:
and Trigger a reading.” Different voltmeter has a 0.3 V range. The
manufacturers had unique ways to commands need to be parsed by ASSIGN @Dvm to 722 !
interpret the command strings, based the voltmeter and this parsing adds
on their instruments’ capabilities. a small layer of delay time to the
If you had to replace a product with system.
one from another manufacturer, or
even a new-generation product from Figure 3.2. Compared to “R2D2” code, SCPI commands standard-
the same manufacturer, it could ize programming and make life easier for the programmer. SCPI
mean completely rewriting the entire commands can access virtually any programming function in the
program. Later versions of IEEE 488 instrument, but the parser does add small delays to the process.
elevated the standard from being a
hardware-only standard to one that PC application software
also specified protocol.
Figure 3.3. SICL I/O software reduces a test engineer’s programming Figure 3.4. VISA is the most popular form of I/O software.
burden by making it easier to change I/O types (USB, LAN, GPIB, USB, Drawing heavily on the work done for SICL, VISA was created to
VXI, RS-232, etc) without recoding the program. SICL adds a software serve multiple T&M suppliers and be a universal standard. VISA-
layer, which has a small effect on system speed. COM is a new variant of VISA.
www.agilent.com/find/open
29
Today’s two main suppliers of PC industry adds language The benefits of these PC software
VISA are Agilent Technologies and independence technologies are compelling, but
National Instruments. (In 1999, As I/O development was proceeding should the test and measurement
the engineers from HP Test & in the T&M industry, the PC industry industry embrace PC operating
Measurement who were involved in was making big strides in I/O-inde- systems?
instrumentation were split from HP pendence and language-indepen-
in the new venture now known as dence. In 1994, Microsoft stated: “The Detractors point out the frequent
Agilent Technologies.) Component Object Model (COM) is operating system upgrades in the PC
a software architecture that allows industry relative to T&M languages.
VISA I/O software uses common components made by different soft- However, as Figure 3.5 indicates,
terminology and syntax to connect ware vendors to be combined into a COM—which is integral to .NET
to and control instruments. A VISA variety of applications. COM defines components—has been around longer
library supports complete control a standard for component interoper- than most T&M standards. It seems
of instrument across the physical ability, is not dependent on any only logical to take advantage of the
interfaces GPIB, RS-232, USB, LAN particular programming language, is investments Microsoft has made to
and VXI. available on multiple platforms, and create this paradigm shift. With 3,000
is extensible.” engineers working for three years on
The VISA library provides the the first version of .NET, Microsoft’s
capability of SICL, in a way that In February, 2001, Microsoft intro- investment is twenty times that of
conforms to industry standards. duced .NET, its 3rd generation of the leading T&M language. Similar
A program written to work with component technology. .NET has correlations apply to software. Visual
Agilent’s VISA library will work with been applied to Microsoft’s inte- Basic has over 6,000,000 users, and
implementations of VISA from other grated development environment, C/Visual C++ has 1,000,000 users
vendors. For those accustomed to Visual Studio®.NET, as well as MS worldwide. This will result in an
using SICL, Agilent’s implementation Office, other applications, operating unprecedented body of software
of VISA is provided along with its systems and web services. the average engineer will be able to
SICL libraries. (Since the introduc- leverage.
tion of VISA, programming based on
the SICL library has gradually been
phased out in favor of the industry- Dr. Dobb’s Journal, Microsoft Corp.
standard VISA library.) December, 1994.
.NET components
to facilitate the use of VISA in Visual industry Microsoft:
Basic.NET and C#. These can be ActiveX
downloaded from http://www.agilent. Microsoft:
COM IVI-COM
com/find/iolib. VISA-COM (components)
VISA
VXIplug&play IVI-C
SICL
1980's 1990's 2000's 2010's
Figure 3.6. The driver is, among other things, a programming aid that works
between the PC application and the I/O software. It can save enormous
amounts of development time and prevent coding mistakes.
PC application software
Your computer
Driver
www.agilent.com/find/open
31
Drivers come in many forms and Figure 3.7. Agilent’s T&M Programmers Toolkit using a VXIplug&play WIN32 power supply driver in
offer various levels of functionality. VB.NET.
A driver can be as simple as a list
that pops up when you hit the next
“dot” in Visual Basic, or it could be
as elaborate as a “panel driver” that
displays a virtual front panel on the
screen of your computer to help you
set up the instrument (Figures 3.7
and 3.8).
Driver coverage
A simple DMM may have only 25
commands, whereas a more complex
instrument may have hundreds.
You can imagine how expensive it
is to write an intelligent driver that
anticipates all the possible permuta-
tions of instrument setup, triggering,
sourcing and measurement. And
that’s why you’ll seldom see a driver
that covers every command in the
instrument.
Figure 3.8. A tiny but interesting program, written in VEE. With its intuitive interface, VEE is the
Instrument manufacturers take their fastest T&M graphical language to learn. Fill in the boxes, and the VEE panel driver generates code
for you. See http://www.agilent.com/find/vee.
best guess at the commands you are
likely to use and craft the driver
accordingly. A typical IVI driver (see
“IVI drivers”) covers about 40-60
percent of the instrument’s command
list. This may sound like a small
number, but consider this: Agilent
surveyed customers who used the
3852A Data Acquisition/Switch Unit,
a complex instrument with over
300 distinct commands. By poring
over customers’ code, we found they
rarely used more than 5 percent of
the available commands. This is an
extreme case, but it tells you that Figure 3.9. If you are using
40-60 percent coverage is a good a driver and need to access PC application software
start. And even if a driver doesn’t instrument functions the
driver doesn’t have, you can
incorporate a particular command,
send direct SCPI or ASCII
there are ways to send commands Direct I/O
commands, or go through Driver
directly yourself (Figure 3.9). the driver with pass-through
(native instrument
commands) Your
commands to control the computer
instrument directly. This 60% ent
um
gives you the convenience of 100% ent instr ability
um cap
drivers, with the 100 percent instr ability
coverage of direct I/O. To cap
avoid command conflicts, this
technique requires in-depth I/O software (VISA)
knowledge on the part of the Physical interface
programmer. Commands sent over GPIB,
RS-232, USB, LAN, VXI or
other physical interface
Physical interface
Figure 3.10. The three generations of drivers represent varying degrees of language independence.
IVI-COM is the newest and the one supporting the widest variety of software environments. For additional information, you can
visit the IVI Foundation website at:
www.ivifoundation.org.
Instrument driver families
Component PC Traditional T&M Proprietary T&M
(based on (based on T&M standards) (specific to one language)
PC standards)
www.agilent.com/find/open
33
Finding drivers and technical Conclusion
advice
If the project you are pursuing is not
Instrument vendors typically provide complex, there are often situations
drivers on a CD with new products where you don’t even know you are
and offer their most up-to-date using a driver. Indeed, the ultimate
instrument drivers on their Web goal of T&M companies is to keep this
pages. For downloads or more process entirely transparent. In the
information on Agilent drivers, I/O meantime, if you do get embroiled
software, connectivity and applica- with issues of driver selection, note
tion software, join us at the Agilent there can be tradeoffs between speed
Developer Network: www.agilent. of development and speed of execu-
com/find/adn. Note that we do not tion. The industry is working through
post drivers written by other parties. these issues by instituting faster I/O
Because you are at the mercy of and software aids, such as tools to
whoever created the driver, it is a keep track of instrument states. The
good idea to use a driver supplied whole idea is to give you both fast
by the same vendor who made the programming and fast throughput.
equipment.
If you choose to use a driver,
Third-party software and systems computer industry-standard IVI COM
integration companies that support drivers and a Visual Studio .NET-
the test-and-measurement industry compliant development program such
can provide driver development tools as the Agilent T&M Programmers
and services. Two such companies Toolkit give you significant leverage.
are Pacific Mindworks (www.pacific- The T&M applications you develop
mindworks.com) and Vektrex (www. will show significant hardware and
vektrex.com). software interchangeability, while
being easily maintainable and exten-
For advice on mixing I/O hardware sible. Plus, the intellectual property
and I/O software from different you create during the development
suppliers, see ftp://ftp.agilent.com/ process will be widely transferable to
pub/mpusup/pc/binfiles/iop/m0101/ other projects.
readme/trouble/niinfo.htm.
Introduction affects the hardware that connects and maintenance costs. Over the
the PC to the instrument as well life of a test system, software
This chapter will help you under- as the software and drivers that support and maintenance costs
stand the tools required to design, control the instrument. alone can exceed hardware costs.
develop and deploy the software
component of your test system • Collecting and storing test data. • Working with open standards. Today,
(see Figure 4.1). The information Data collection is the science of the industry trend is to move away
presented here will help you choose obtaining, moving and formatting from closed, proprietary develop-
the direction for your software based data. The integrity of your test ment environments. More and
on the application you have in mind system depends on obtaining the more people are embracing open,
and the amount of experience you right data at the right time. industry-standard development
have. We will explore the entire environments as their platform of
• Designing the user interface. One of choice for test-system development
software development process, from
the most important (and easily projects. Making the right choice
gathering and documenting software
overlooked) aspects of test systems now will give you the flexibility and
requirements through design reuse
is the graphical user interface capabilities you need in the future.
considerations.
(GUI). This is what the test engi-
• Gathering and documenting software neers, operators and technicians • Developing a test sequence. Test
requirements. Before gathering see when they interact with your executives are applications
and documenting your software software. designed to run a series of tests
requirements, finalize your test quickly and consistently in a pre-
• Choosing the development environ- defined order. Of the 93 percent
system hardware design. Once
ment. The software environment of test-system developers who use
finalized, start working with your
and tools you choose will have a test equipment, approximately
R&D and manufacturing teams to
significant impact on the overall 37 percent use a commercial test
collect the information you need
cost of your test system. When executive for test sequencing, while
to create software requirements
choosing your software environ- the remaining 56 percent use a
specifications (SRS).
ment, consider more than just the “homegrown” test executive.
• Programming and controlling your purchase price of the software.
instruments. The control of instru- Also, consider how easy it is to • Planning for software reuse.
ments is rapidly evolving from learn and use the software, how Designing for code reuse means
proprietary test and measurement hard it is to connect to other you and your co-workers won’t
standards to open, computer-based languages, devices or enterprise have to re-create your software
industry standards. This trend applications, as well as support components every time you start a
new project. Instead, you can build
Figure 4.1. Test-system software development process overview up a company knowledge base
of best ideas, best practices, and
Time
software components. This knowl-
edge base will bring uniformity
and consistency to your company’s
Gather Open standards? product testing functions.
manufacturing Data
requirements collection
Graphical or This chapter will provide you with a
Performance textual? solid overview of the test system soft-
ware architecture as outlined above.
Gather Software Test executive? For more in-depth information, refer
R&D Requirements to the sources listed throughout this
requirements Specifiaction (SRS) document. Now, let’s get started with
Design operator interface the first phase of choosing your test-
Test
system software architecture—gath-
specification
Finalize Prepare data collection strategy ering and documenting your software
test system requirements.
User interface
hardware Design for reuse
35
Gathering and documenting • External interfaces. How does the A good SRS answers the following
software interact with people, the questions:
software requirements system’s hardware, other hardware
and other software? 1. What measurements and tests are
The Software Requirements required to exercise the device
Specifications (SRS) is a prioritized • Performance. What is the speed, under test (DUT)?
list of required test-system software availability, response time and
capabilities and information on recovery time of various software 2. How will the measurements
the software’s external interfaces, functions? and tests be performed given
performance requirements, system the available instruments and
attributes and design constraints. • Attributes. What are the portability, devices?
Typically, some requirements “musts” correctness, maintainability and
are essential and others “wants” can security considerations? 3. What types of data need to be
be traded for time (e.g., to meet a collected?
project deadline). • Design constraints. What industry
standards need to be followed? 4. Where will the data be stored?
The IEEE identifies the following Does a specific language need to be
areas you should address in your 5. What are the external constraints
used? What about internal policies
SRS: (e.g., performance and time
for database integrity, resource
specifications)?
limits and operating environments?
• Functionality. What is the software
supposed to do? 6. How will the operators, test engi-
Ideally, the SRS will describe WHAT
neers and technicians interact
you need the software to do, not
with the software?
HOW the software will do it. In other
words, you can look at the software Within the product development
May be referred to as an ERS or as a “black box” that controls a set of lifecycle, the R&D department should
simply as “the requirements.” external resources such as instru- provide a formal list of testing
For more information, refer to ments, a computer monitor and other requirements to the test-develop-
the IEEE Standard 830-1998 components (see Figure 4.2). ment department. The System
“Recommended Practice for Software Requirements Specifications, also
Requirements Specifications” and the The SRS will include implementation
referred to as a Project Requirements
IEEE Standard 1233-1998 “Guide for details only if those requirements
Specification, refers to the system
Developing of System Requirements are imposed externally. For example,
Specifications” located on the IEEE as a whole and therefore is different
your company may require that
web site (http://standards.ieee.org). from the Software Requirements
a portion of the system be imple-
Specifications. Furthermore, the
mented in a specific programming
manufacturing department will have
language.
its own requirements, such as safety
standards. It is the combination of
Figure 4.2. Scope of the SRS R&D and manufacturing specifica-
tions that determine the hardware
requirements of a test system and
Instruments Results Other Operator
database resources interface provide the basis for the Software
Requirements Specifications.
www.agilent.com/find/open
37
Programming and control- In addition to external data require- debug. For example, in one particular
manufacturing test situation, a DUT
ling your instruments ments, recorded data can be used
to debug a test sequence in ways would fail in the afternoon. The
debugging runs cannot. Debugging test engineer was able to correlate
When designing your test-system
means slowing down and subtly the time of day to the time of the
architecture, you need to think about failure and use that information to
changing the behavior of your test
how your PC will communicate with look more closely at a photoelectric
sequence. This means a defect
different instruments. The two most you see in a normal run may not component of the DUT. It turned
important factors are deciding how show up in a debugging run (and out that sunlight would strike that
to physically connect the PC to other vice-versa). One way to reduce the component directly at certain times
instruments and deciding what burden of diagnosing test software, of the day, causing the component
software will you use to control and and its associated DUT, is to always to charge a capacitor and cause the
collect the data you need to debug a test to fail. A DUT may fail due to the
communicate with other instru-
problem. You will need to balance the temperature variations or relative
ments. Refer to Chapter 2 for advice
benefits of collecting extra data with humidity. Capturing contextual infor-
on choosing an I/O option and to mation and measurement conditions
the costs in performance and time for
Chapter 3 for advice on choosing and your test software. can save days of effort.
using drivers and other instrument
communication software. Just as important as the standard You want to ensure the writing or
types of data (e.g., test limits, formatting of your data does not
Collecting and storing the measured values and pass-fail affect the behavior of your test
system. Today’s PCs use a variety of
judgments) are the contextual
test data data. Contextual data are used to caching techniques that can dramati-
communicate everything relevant to cally affect how long it takes for a
Data collection is the science of the DUT’s testing environment. This given file or network I/O command.
identifying, collecting, formatting includes the test-system configu- If the time it takes to cache your data
and distributing important informa- ration, software version, driver varies between each test run, you will
tion about the behavior of your test versions and other factors. get inconsistent test results. For that
system and the devices it tests (see reason, it’s a good idea to keep your
The more variables you record, the data in RAM until the end of your
Figure 4.4). Quality data collection
more correlation points you and DUT testing and then do your format-
and analysis is the foundation for ting and data transmission.
your colleagues can analyze during
controlling your manufacturing and
test processes—the ultimate goal of a
Figure 4.4. Overview of the data collection process
manufacturing test engineer. Quality
data also can be used to support
many functions throughout your Loaded battery XML
organization and support products response data file
(volts)
throughout their development
lifecycle. Results
database
Response
Communicating results of a test voltage limits
[12,100]
sequence is one use of test data. Test Report ticket
data also may be used to ensure regu- Data printer
latory standards are met, document Model 320a
SN:00010145 Fail!
performance standards, or provide
traceability for the DUT. Given these Operator
interface
applications and others, you may 12:03 pm
Dec. 14, 2004 Fail!
want to collect more data than your
R&D or manufacturing colleagues
request.
www.agilent.com/find/open
39
The manufacturing test industry has Designing the user 1. Customize its behavior based on
already begun adopting XML. Some the user class.
test executive applications support interface
XML data logging. There is a stan- 2. Provide or allow input of detailed
When a user displays generated by information about the DUT.
dard called IPC 2547 that defines an
a test system should vary according
XML format for communication of
to the class of user, such as operator, 3. Provide information about the
manufacturing test data.
test engineer, technician, or service state of the system.
Figure 4.5 is an example of a and calibration engineer. A well-
written SRS defines the commands 4. Provide operations for controlling
standard test run in XML format.
and/or menu selections available the system’s state and potentially
You will still want to know the test
to each user class. You will want to its configuration.
sequence ID, the variant of the test,
if the test limits are modifiable on provide each user class with only the
capabilities and information those 5. Display the DUT testing results.
the “PowerTest” and the hardware
configuration of the test system. people need to do their jobs. The For an operator, the interface you
more choices you provide, the greater design should always show the state
If this were a .csv file, we would have the possibility for confusion and of the test system (e.g., running
to create a field for every record to mistakes. a test, paused or stopped). For
answer those questions. Using XML, example, you could use a large color-
we can insert a record type called To ensure security, you can create
a unique login for each of the users. coded graphic on the PC monitor in
<TestSequence ID=”32”> and fully conjunction with lights mounted on
describe the current test sequence Each user login should be linked to
the appropriate class. the test system. The operator also
in that record. We can then add an will need a way to control the state
XML attribute called “IDREF” to refer of the test system as well as a way to
You can verify that your GUI meets
to that test sequence record in our input DUT information (unless this
the users’ needs with a methodology
<TestRun> records. is done automatically via a bar code
called “User-Centered Design,” or
UCD, which consists of prototyping scanner).
In summary, the data format you
choose will have a large impact on its and storyboarding., In general, a
test system’s GUI should be able to As a general rule, test program
value over time. You need to consider should have the following features:
how easy or difficult it will be for
someone else to read and interpret 1. Commands for starting and stop-
Vredenburg, Karel, et al, “User-
the data once you are no longer Centered Design, an Integrated ping the test sequence.
involved in the project. Approach.” Prentice Hall PTR, 2002.
2. Commands for sending test
Norman, Donald A., “The Design of results to various kinds of
Everyday Things.” Basic Books, 2002. printers (defect report ticket,
IPC 2547: http://webstds.ipc.
org/2547/2547.htm etc.).
www.agilent.com/find/open
41
Choosing the development environment. In addition, it is easier Graphical or textual
environment
to modify small to medium size
programming?
graphical programs versus textual
programming languages. Historically, Before you can decide on which
The next step in choosing your test
textual programming languages ran development environment is best for
system software architecture is
faster in the manufacturing environ- your application, it’s important to
to select a software development
ment and yielded higher throughput. understand the use model of each in
environment. The software environ-
Today, there is less difference greater detail.
ment and tools you choose will have
between the runtime speeds of a
a significant impact on the overall
graphical environment and a textual Graphical programming is accomplished
cost of your test system. When
environment.. by manipulating images, called icons
choosing your software environment,
or objects, and the lines that connect
consider more than just the purchase Even though graphical environ- these images. The images represent
price of the software. You need to ments are easier to use than textual pre-made commands while the lines
consider how easy it is to learn and environments, textual environments represent the program flow, control
use the software, how hard it is to are used more commonly in manu- points, and /or how data are gener-
connect to other languages, devices facturing test systems. Only about 22 ated and consumed. The icons and
or enterprise applications, as well as percent of the half million-plus users interconnecting lines are contained
support and maintenance costs. Over who write code for test and measure- within the integrated development
the life of a test system, just software ment equipment use a graphical environment (e.g., the software
support and maintenance costs can programming language. program).
exceed hardware costs.
You have a number of options when Figure 4.8. Agilent VEE Pro graphical programming environment
it comes to software development
environments, from writing every-
thing yourself in a language such
as C, C++, C#, VB, VB .NET, Agilent
VEE, MATLAB or LabVIEW, to using
an off-the-shelf test executive with
pre-written third party tests. The
software environment you choose
needs to accomplish two goals:
1) meeting your time-to-first test
requirements and 2) meeting your
test-throughput requirements. How
fast can you get your test system up
and running, and how can you get
the greatest throughput?
www.agilent.com/find/open
43
2. Simpler navigation of large programs. 4. Greater choice of development Choosing between proprietary
environments. For example, there
For larger programs, textual
and open standards
programming languages are are few graphical programming
easier to navigate and compre- languages that have develop- Several factors to consider when
hend. A person can observe only ment environments provided by deciding between an industry stan-
about 50 graphical objects at multiple vendors. This means that dard and a proprietary development
a time before the information today’s graphical languages are environment are 1) cost, 2) industry
becomes too complex or too small less likely to have the advantages support, 3) upgradeability, and 4)
to see.12 If a user is forced to move created by competition between extensibility.
around in a program to see all its vendors.
objects, he or she can lose track Development environments for open
of the control and data lines and Graphical programming tends to standard programming languages
find it difficult to understand be easier to learn and comprehend have a greater feature set and are
the overall flow of the program. while textual programming is more less expensive than their proprietary
With that said, you can improve pervasive and open. Table 4.2 counterparts. Simply stated, an
the understandability of large summarizes the differences between open standard environment tends to
graphical programs by breaking the two programming environments. create greater competition, which in
up the program’s large operations turn tends to drive down prices and
into smaller suboperations. This Working with open create innovation.
is called functional decomposi-
tion and is achieved by putting a
standards Open-standard languages generate a
series of commands into a “black lot of interest from both software tool
In addition to choosing between
box”. You then send commands to vendors and open-source developers.
graphical and textual programming,
the functional block and receive Both of these groups spend consider-
you need to consider whether the
its output as appropriate. Make able time understanding the needs of
environment you choose will be
sure the graphical program you the test-system programmer and, as a
based on industry standards or
use supports this functional result, develop both free and for-pay
propriety, vendor specific technology.
decomposition13 if you plan on tools and applications to meet those
C++, Visual Basic, and C# are all
working with larger programs in a needs. A good example is the tremen-
examples of industry standard
graphical environment. dous number of C and C++ libraries
programming environments. Agilent
available on the market, both from
VEE Pro and NI LabVIEW are
3. Higher system throughput. The vendors and from end-users. These
examples of proprietary development
faster runtime speeds of a libraries save development time and
environments although Agilent VEE
textual programming language money given that it is faster and less
Pro 7.0 does allow for access into
can improve overall system expensive for a developer to buy the
industry standard technologies such
throughput. However, be aware domain-specific software (such as
as .NET.
that the time spent during instru- mathematical analysis libraries) than
ment operations will often have create it from scratch.
a greater impact on throughput
than the choice of programming Table 4.2. Graphical versus textual programming
environment. For example, time
lost through inefficient signal Graphical Textual
switching between the test system Free and open Few open standards, less Dominated by open standards,
and the DUT can far outweigh extensible very extensible
any time savings earned through
Rapid prototyping Excellent T&M prototyping Some code wizards, (T&M
choice of programming language.
features Programmers Toolkit, for
example) but slower
12 Begel, Andrew, “LogoBlocks: A T&M support Several graphical environments Several T&M-specific 3rd-party
Graphical Programming Language targeted at T&M, many drivers tools available, many drivers
for Interacting with the World,” 1996,
3rd-party tools Hundreds Tens of thousands
http://www.cs.berkeley.edu/~abegel/
mit/begel-aup.pdf. Learnable and Easy to pick up and use programs Only small or very-well-designed
shareable programs are easy to share
13 Glinert, E. P., “Visual Programming
Environments,” IEEE Computer
Society Press, 1990.
www.agilent.com/find/open
45
One significant advantage of .NET Developing a test sequence and allow the user to create addi-
over older programming technologies tional test modules from scratch
is its extensibility. Microsoft engi- In a survey of more than 2,500 test (as well as customize existing test
neered .NET so that it avoids a lot and measurement equipment users, modules). Test executives control
of the DLL installation frustrations 93 percent of the respondents said the data to and from the test module
Windows programmers experienced they use multiple test instruments and, after collecting and analyzing
in the past. There are already a and /or are connecting their test all of the data, determine if the DUT
large number of third-party tools instruments to a PC. Of that, 37 passed or failed.
for .NET. Many of these third-party percent said they use a commercial
controls (i.e., advanced graphing test executive for test sequencing. One reason for using a test executive
visual controls) are useful to test- The remaining 56 percent of these is it provides a structured framework
system programmers. Additionally, respondents use internal or “home for manufacturing test systems. Test
several test and measurement grown” software for test sequencing. executives work best in medium- to
vendors, including Agilent low-mix, and medium- to high-volume
Technologies, National Instruments, A test executive is a software manufacturing test environments.
and Measurement Computing, have application designed to run a series
of tests quickly and consistently in Test executives are written so that
released .NET-compatible tools. For
a predefined sequence. If any of the sequence design, individual test
a complete list of released .NET-
tests within the test sequence fail, design, and test limits and configu-
compatible tools, refer to Microsoft’s
then the DUT fails. Over the years, ration management are treated as
.NET partner web site at www.
test executives have improved consid- separate tasks. Keeping the three
vsippartners.com.
erably both in terms of flexibility tasks separate results in greater
Agilent Technologies’ first add-in and capabilities. First-generation flexibility, higher quality, and an
for Visual Studio .NET is called the test executives were language- increased opportunity for code reuse.
Test and Measurement Programmers specific and not powerful enough It is the test executive that provides
Toolkit (see the sidebar on page for a mission critical manufacturing the infrastructure and helper
43). The T&M Programmers Toolkit environment. Second-generation services required to connect each of
provides I/O tools, graphing and test executives, such as Agilent’s the separate tasks into a complete
mathematical libraries, T&M specific TxSL and NI’s TestStand are more program.
help and example generators, and powerful but more expensive. They
One of the most important features of
.NET wrappers for instrument also lack the flexibility required for a
a test executive is its test sequencer.
drivers and other software. The T&M low-volume, high-mix manufacturing
As shown on the left side of Figure
Programmers Toolkit is fully inte- environment.
4.10, the test sequencer is a sequence
grated into the Visual Studio envi-
Each of the tests within the test of tests that can be manipulated in
ronment. For more information on
sequence is a separate module. design mode. Various test executives
Agilent’s solutions, go to http://www.
Commercial test executives come provide different levels of flexibility
agilent.com/find/toolkit or http://www.
with a standard set of test modules in this sequence, such as “test
agilent.com/find/iolib. To download
looping.”
.NET-related I/O source files, which
also work with the Agilent I/O
Figure 4.10. The test executive test sequencer
Libraries, go to the Agilent Developer
Network (ADN) web site at http://
www.agilent.com/find/adn.
www.agilent.com/find/open
47
The design reuse process A design reuse example You can achieve similar reuse success
in your own code with good archi-
The first step in the design reuse A good model for design reuse of tecture influenced by the natural
process is to complete a domain individual test modules is the test boundaries of your software’s func-
analysis. This is accomplished by 1) executive—here’s why. tions, types and data. To accomplish
systematically analyzing the func- this, put information that changes
tions and parts of your software 1. Some test executives break test
frequently, such as the limits for a
domain, and 2) using this informa- software up into swappable tests,
test, into a Data File. Put less flexible
tion to develop a software architec- sequencers, limit checkers, test
elements, such as a test class, into
ture with well-defined component sequence and test limit data.18
Types or “classes.” Functions, or
types and algorithms. 2. Test executives rely on the “procedures,” should be reserved for
concept of modules. For example, the least flexible elements.
Next, you will want to look for
natural boundaries in your software. you can have a module that
One software design practice of provides the ability to perform Design reuse and .NET
finding and documenting the natural a single pass or fail judgment,
boundaries is known as Design including the sequencer data While the definitions of the bound-
Patterns.17 To find the natural bound- type, the sequence execution aries of your software domain are
aries, look to those areas where one operation, and the test types. not specifically influenced by the
type of activity or data set links with programming language or software
3. Test executives allow reuse of environment, some environments
another type of activity or data set. tests in different test sequences
These areas can then be grouped into are better than others in helping to
with no change to the test code. keep your software modular and
separate modules and documented The sequencer provides the
accordingly. Once documented, swappable.
necessary data to the tests to
the same type modules can then be customize their operation for the .NET provides software tools that
swapped for one another. current test sequence. make it easier to develop a formal
Once you have identified, collected software reuse program within
4. Test executives keep the tests in your department or company. Since
and documented your modules, separate modules or files from
components and /or individual parts, .NET is object-oriented, it’s good at
the test sequencer or test execu- representing boundaries between
you will need to thoroughly test them tive application. This allows you
before they go into the repository (or different types of objects, such as
to easily swap tests in and out tests or sequencers. Nonobject-based
are passed on to your co-workers). without recompilation.
This will save you and your co- languages, such as C, require you to
workers from problems later in the keep track of which functions apply
5. Some test executives allow you
process. to which objects, without much
to write your own custom limits
context-sensitive help or compile-
checkers or sequencers.
Finally, reusable components are time error checking.
reusable only if your co-workers All of these modules are able to
know they exist. You need a reposi- .NET also includes improved
interoperate because test execu-
tory (such as a relational database) versioning and deployment features.
tives use well-defined application
for your modules where anyone in In addition, .NET has the ability
programming interfaces (APIs) for
your team, division, or company (if to tell Windows that you will only
each module. The modules are placed
appropriate) can browse and search accept a certain version of an
on natural boundaries between
for them based on what the compo- external library. This eliminates one
different types of data and functions
nents are and what they do. of the common frustrations with
within the test executives.
earlier versions of Windows where
you rely on an external library
17 Shalloway, Alan and Trott, James 18 This is a good example of a design (DLL), but then the DLL changes and
R., “Design Patterns Explained: A pattern specific to the test and your software no longer functions
New Perspective on Object-Oriented measurement domain. correctly.
Design,” Addison-Wesley Pub Co,
2001.
www.agilent.com/find/open
49
Conclusion Whether you choose a graphical or
textual environment will depend
Before you begin writing code for on the size and complexity of your
your test system, you need to make a system, your skill set, your company
number of important decisions about standards, and the size of your
the system’s software architecture. programming team. The decision
You will want to start by creating usually comes down to which envi-
a detailed software requirements ronment—graphical or textual— will
specification that defines what you make you more productive. Textual
want the system to do and how it environments are almost always
should operate. The SRS should the best choice for creating code for
include an outline of how you will large, high-throughput manufacturing
gather, store, analyze and present test systems because they offer the
your data as well as how end users most power and flexibility, and they
will interact with your system. allow faster throughput.
Another important decision you need Finally, you need to decide whether
to make upfront is which program- to use an off-the-shelf test executive
ming environment and language or write your own test routines. Test
you will use for writing your code. executives can speed up your test
Using a standards-based environ- system development and lower your
ment such as Visual Studio .NET costs but will require an up-front
maximizes your flexibility and helps training investment. If you are only
you prolong the useful life of your performing a few tests, you may want
software. By combining Microsoft’s to consider writing your own code.
Visual Studio .NET with Agilent’s
T&M Programmers Toolkit, you can
wrap objects written in a variety of
languages such as Agilent VEE Pro.
This allows you to pull them forward
into your new programming environ-
ment, making the most of your legacy
code investment.
51
Instrumentation type: rack- Another cardcage architecture is Another cardcage architecture is
and-stack or cardcage? called PXI (PCIbus eXtensions for compact PCI (CPCI). CPCI technology
Instrumentation). While PXI cards is the basis for PXI, although PXI
There are two major types of instru- are very small, they typically lack the adds triggering options not available
ments for test systems, rack-and- accuracy and performance of VXI or in PCI. CPCI and PXI cards can be
stack and cardcage. Rack-and-stack rack-and-stack instruments. If you interchanged to some extent. CPCI
instruments are standalone test are considering using a PXI system, cards tend to be used in industrial
instruments that can be used inde- be sure to investigate whether you PCs, because they are rack mount-
pendently. For test systems, they are will need to purchase additional able and more rugged than other
frequently stacked in a rack (hence signal-conditioning equipment. Also, card types.
the name) to save floor space, and PXI is based on a PC backplane with
typically, engineers use external PCs no electromagnetic interference
to control them. Newer LXI instru- (EMI) or cooling specs, and therefore
ments often come in both traditional it is not as well suited to be a quiet
rack and stack formats as well as in measurement environment. Note
smaller modular formats. that PXI is also transitioning to PXI
Express, so be sure to look at your
Cardcage instruments needs in the future to determine
Cardcage instruments, as their name if you should purchase a hybrid
implies, are modular test instru- PXI/PXI Express cardcage. See
ments on plug-in cards. You insert the sidebar on page 54 to compare
the cards in a cardcage, or main- attributes of PXI, VXI and rack-and-
frame, and control them either with stack systems.
an embedded controller (a plug-in
card that is a PC) or an external PC. Figure 5.2. VXI mainframe with modular test instruments on plug-in cards
Card-cage systems are often mixed
with rack-and-stack instruments to
provide all the functions needed in a
test system.
www.agilent.com/find/open
53
Comparison of instrumentation types
time update capability. This is especially
Rack and VXI CPCI PXI See notes:
stack true with an oscilloscope, where lack of
immediate feedback can be annoying.
Standalone use Yes No No No 1
Accuracy 2 7. Footprint
**** *** ** **
Price $$ $$$$ $$$ $$$ 3 PXI and CPCI systems have the smallest
Burst speed 4 footprints. However, many instrument
** to **** **** **** **** functions are not fully realizable in PXI,
Single-point ** *** ** ** 5 so engineers typically adopt a hybrid
measurement speed
approach of rack-and-stack plus PXI
GUI response time **** ** ** ** 6 instruments. Once you have a rack for
Footprint 7 part of your system, you use the same
** ** **** ****
amount of floor space as you would for
Ease of use and **** * * * 8
integration a full rack-and-stack system, so you lose
the space-saving advantage offered by
Shielding **** *** * * 9 the small form factor of the PXI cards
1. Standalone use rack-mounted, system cost depends on 8. Ease of use and integration
the configuration of the rack.
With an internal PC, a cardcage can If a racked system has been designed
operate standalone, but you need a 4. Burst speed to accommodate a reasonable amount
monitor if you require an operator GUI. of expansion space (a good idea to plan
Cost of an embedded PC is several times Burst speed is the speed at which the
for unforeseen future needs), adding
that of a standard PC. In any case, card instrument can move a large amount of
instruments to a rack is not a lot more
cages generally require some form of data from a single channel across some
complicated than adding an instrument
computer communication in order to be bus or I/O port to the computer. Burst
to a cardcage. A more important consid-
useful, while rack-and-stack instruments communication is used in data acquisi-
eration is the ease of adding additional
can be used to check out the system tion more than it is used in functional
cables to an existing architecture. For
without a computer present. test. Cardcages typically shine in this
example, whether you use a cardcage or
arena, although recent improvements
several racked instruments, their inputs
2. Accuracy in I/O speed, such as the adoption of
and outputs are usually connected into
fast LAN, have blurred the distinction
Cardcages have power supplies that a switching system or a mass intercon-
between backplane and external I/O.
must be shared among several subsys- nect. If the system has been designed to
tems. Rack-and-stack instruments 5. Single-point measurements handle such new instruments, integra-
are optimized to one use, so they are tion will only take a few minutes. If the
designed to have the right power supply Single-point measurement speed is the system has to be redesigned to handle
for the job at hand, and analog circuitry time it takes to make a single measure- the new instrument, it can take days.
that is not subject to cage-imposed ment, switch channels and then make
restrictions. Rack-and-stack instruments another measurement. This is the 9. Shielding
are designed to minimize magnetic inter- predominant mode used in functional
Dedicated rack-and-stack instruments
ference so they are less likely to induce test. You’ll find more information
are typically well shielded. Since they
currents that would disrupt sensitive about test-execution speeds in the
are designed for a specific purpose, they
instruments. As a result, rack-and-stack “Measurement speed” section on
are frequently more noise-free than their
systems typically outperform cardcage page 60.
card-cage counterparts. VXI has specific
systems in terms of accuracy, crosstalk, 6. GUI response time shielding specifications, and these are
noise, and other factors. lacking in PXI and CPCI. While it is
When a cardcage communicates to possible to shield PXI, the implementa-
3. Price the PC, the PC must often do double tion is left up to the vendor, so placing
Cost of a bench-top system is usually duty as it processes the data and also a new vendor’s product in a slot may
lower when instruments are not updates the GUI. In some rack-and-stack result in unwanted interference with
rack-mounted. When instruments are instruments, these operations happen in nearby instruments.
parallel, giving the operator more real-
www.agilent.com/find/open
55
In manufacturing test and design of high-speed instruments, not and mass interconnect are in the
validation systems you often need simple relays. If you place relays in same plane, and 2) it reduces wire
banks of general-purpose relays a separate box that is tuned for that length from the switching to the DUT.
of varying current capability. You purpose, it will be easier to expand If the box you choose has cards in the
can use such relays to connect DUT the high-performance instrumenta- rear, reverse-mount the switchbox
inputs to ground or to a supply, or tion while allowing room separately using the rails on the rear of the
through resistors to simulate dirty for denser relay cards, more relay rack, as shown in Figure 5.4. There
switches. You also can use them to cards or a bigger switchbox. It are two negatives to this approach:
provide ways to disconnect output also makes a clearer delineation the front panel of the switching
loads in order to allow parametric between the instrumentation and instrument is not accessible from
tests on output transistors, as shown the switching subsystems, which the front of the system, and it can be
in Figure 5.3. makes it easier to keep your system harder to reach the plug-in cards for
organized. service. However, once a system is
You also need to think about where operational, it is seldom necessary
to place and how to arrange your Placing the DUT interface panel to operate a switchbox from its front
switches. While relay cards can be (mass interconnect or feedthrough panel, and cards can be accessed by
placed in a cardcage that is intended panels) in front of a switching pulling the instrument out the back
for high-performance instruments, it subsystem that has the plug-in cards or by removing the side panel of the
is a waste of valuable real estate. The facing the interface panel accom- system.
high-speed backplane in a modular plishes two goals: 1) It minimizes
cage is more suited to the control rack space, because the switchbox
Figure 5.4. Rear-mounting the switching
subsystem reduces rack space and minimizes
Figure 5.3. Switched loads allow parametric measurements cable lengths
Switch in load
for powered test
Typical DUT
outlet driver Load 12v
MOSFET with
zener protection
DVM
Measure protection
zener using current source
Current sense
resistor
R
+
Measure leakage
current using DVM V
voltage source
www.agilent.com/find/open
59
If you want to minimize hardware You typically spend a large it takes extra effort to create and
costs, you can investigate auxiliary percentage of total development time download a decision algorithm into
capabilities. However, if your goal on debugging your system, particu- an instrument, which may increase
is to minimize development time, larly if you are building a new test development time as well as “first-
buy instruments that are specifically system. You can reduce your debug run” time of the test program. Also,
designed to do the jobs you need time significantly by writing a diag- inside an instrument the readings
done. Using instruments with IVI- nostic test routine that loops outputs will be analyzed by a much slower
COM drivers can save you develop- back to inputs through a large part of processor than the one in the PC, so
ment time. If the instrument has an the switching path. This exercise will this must be factored in as well.
IVI-COM driver, you can interchange help you quickly identify the cause of
hardware without rewriting your problems— whether it is a source, a Simply reading the data sheet does
software, as long as you adhere to measurement instrument or a switch not tell the whole story. Maximum
the functionality that is specific to path. reading rate specifications are
the instrument class. See Chapter 3, usually related to burst speed (see
Understanding Drivers and Direct For more ideas on minimizing your Figure 5.6); that is, the speed which
I/O, for to learn how decisions about development time, see Chapter 4, you can sample the signal on a
drivers affect development time. Choosing Your Test-System Software single channel. But that is not the
Architecture. typical mode for functional test. In
Test instruments with download- functional test, the system normally
able personalities also can save you makes a single measurement, then
development time. You download
Measurement speed changes a parameter like range or
the measurement personalities for a If you are building a manufacturing function or channel, and then makes
specific application directly into the test system (and to a lesser extent in another measurement. In this case,
test instrument’s internal memory. design validation applications), the the burst rate is meaningless. Take
Then you can simply choose from a time it takes to execute each test can for example, two multimeters—one
menu of tests, and the personality’s be critical. But figuring out how fast LXI and one PXI. Note that both
“intelligence” automatically performs your system will perform measure- multimeters can perform up to
the tests, from capturing signals ments is harder than it appears. For 10,000 measurements/second
to displaying results. Agilent spec- example, a digitizer may be able to or more in burst mode, but their
trum analyzers, for example, have sample 1000 readings very fast, but single-sample measurement speed
measurement personalities for if those readings are transferred to is much slower due to the transac-
testing cable TV, phase noise, cable the PC over GPIB, it could take a long tion overheads of controlling each
fault, Bluetooth™, cdmaOne, GSM/ time. A digitizer that can have a deci- measurement. Even a high-speed bus
GPRS, as well as a variety of other sion-making algorithm downloaded such as PXI makes little difference
wireless protocols and modulation. into it could allow a simple go/no-go to the readings/second because the
result to be sent back to the PC, total time is dominated by the setup
New LXI instruments from Agilent and measurement time.
which would make GPIB a reasonable
allow instrument monitoring from
option and may save money over a
the instrument web page. This allows
cardcage-based solution. However,
monitoring of the instrument state
from the same computer screen as
your test program. The web page is Figure 5.6. Burst speed can be misleading; since single-sample measurement speeds are usually
also a useful debugging tool. significantly lower.
30,000 350
25,000 300
20,000 250
200
15,000
150
10,000
100
5,000
50
0 0
LXI DMM PXI DMM LXI DMM PXI DMM
www.agilent.com/find/open
61
Make architectural choices • Low current DC voltage source Our system uses many I/O interfaces:
(0-5 VDC) LXI (LAN), RS-232C, FireWire and
Table 5.1 shows the architectural GPIB. Using Visual Studio.NET with
choices we made for this test system. • Waveform analyzer
IVI-COM and VXIplug&play instru-
• CAN interface ment drivers along with VISA I/O
libraries, the control program can
Design the system • Simulated or actual stepper motor
communicate easily with instruments
load
Now, we will apply the architectural on all of these interfaces. In fact,
decisions to a system for testing an The DUT has 14 pins total on 3 should an instrument’s I/O interface
electronic throttle module for an connectors. Looking at various ever change (say from FireWire to
automotive throttle body. According catalogs, and adopting the architec- LAN), all that will have to change
to the test specification, the following ture specified earlier, we chose the in the program is the initialization
equipment is required to run the tests: instruments shown in Figure 5.7. string. It is also possible to specify
use of an aliased name to eliminate
• Programmable volt/ohm/ammeter There are three LXI instruments— the hard-coding of I/O addresses.
the power supply, switchbox, and
• Programmable power supply— Figure 5.8 shows how the instru-
oscilloscope. We will use an 8-port
0-13.5 V/0-10 A ments will be connected to the
LAN hub providing extra ports, thus
• Waveform generator capable of “future-proofing” the system. Table switching subsystem. We are using
pulse-width modulation, 0-10 VDC, 5.2 lists the instrumentation used in a matrix, so any instrument can be
0-3 KHz this sytem. connected to any DUT pin, and we
CAN H/L
pins is small, simply bringing them CAN CTM Pot2 Vref
2 Pot2 Gnd
directly out of the switchbox to DUT
Fgen Pot1 Wiper
connectors may be sufficient. In the Pot1 Vref
future, if the modules you are testing Pot1 Gnd
have more pins, or if you need a Mixed Sig
Scope Brake
place to put other things between Accel
the system and the DUT, you may
need a commercial mass interconnect DAC Mot –
solution. Therefore, we will provide a DMM Mot +
space directly in front of the switch- Digitizer
box for such an interface. VXI Pwr Gnd
1a 1b
VBatt
We chose a 5-wire measurement bus 2a 2b
because it allows all four leads of the Load tray 3a 3b DUT - electronic
DMM to be connected to different – 4a 4b throttle module
pins on the DUT, making 4-wire ohms – Sense 5a 5b
measurements possible. We routed PS + 6a 6b
two matrix points to the same pin + Sense 7a 7b Mass
on the DUT (as shown in Figure 5.8 interconnect
on the Pot1 and Pot2 Gnd pins), to
make the resistance measurement Table 5.2. Instrumentation decisions for sample test system
very accurate, since the remote sense
location is made right at the DUT. If Instrument Reason
you don’t use two wires, you can still Rack-mountable arbitrary waveform/ Need to generate PWM signals inexpensively
make a 4-wire ohms measurement function generator
inside the relay matrix, which in Heavy-duty power supply Module requires 10A of inrush current
some cases may be good enough. The Optional DMM Debug
fifth bus wire is connected perma- Oscilloscope with CAN trigger module Monitors signals including CAN traffic
nently to the star ground, and so it
Dedicated switching cardcage (“switchbox”) Separate cardcage-based switchbox houses
serves as a common reference for low-data-rate instruments more cost effectively
any single-ended devices, such as the
4-slot VXI cage containing: Provides the most channels in a reasonable form
oscilloscope, or for floating devices factor; space for future expansion
that can be connected to ground,
• Digitizer For high-resolution sampling
such as the function generator,
• 16-channel DAC Need a DAC for generation of a brake signal
digitizer, DAC and DMM.
• High-speed DMM Actual measurements are fastest with this one
• An RS- 232C-based CAN interface Module requires CAN interface for putting module
is located on a shelf behind the PC in test mode
www.agilent.com/find/open
63
It is helpful to make a wiring map resources in the columns. Since Conclusion
that shows how the DUT will connect star ground is physically located
to your system. Table 5.3 shows how outside of both the system and the Before you begin choosing test
to make one using a spreadsheet. DUT, it shows up in both a row and instruments for your test system, you
In the future, when it becomes a column. Wires are connected from need to make a series of high-level
necessary to test a different DUT, the DUT pin number to the relevant decisions about your system archi-
all you need to do is to create a new system resource. For example, the tecture. The architecture you choose
spreadsheet and wire the new DUT battery input, Vbatt (J1-1), has for your test system will depend on
accordingly. two wires attached to it—one to whether you plan to use it for R&D,
general-purpose relay 7b and one to design validation, or manufacturing
Since the system has many resources general-purpose relay 6b, which puts test and on your budget and develop-
available and they can be expanded remote sense of the power supply ment-time constraints, your existing
without changing the basic system right at the DUT. In addition to DUT expertise and your measurement
architecture, new DUTs are easily pins, there are other internal system throughput requirements.
accommodated. The spreadsheet is connections that must be made, and
constructed with DUT pin names they are shown in a separate section Important questions to consider
and numbers in the rows and system of the spreadsheet. include the following:
1. Should you use a rack-and-stack,
Table 5.3. DUT wiring spreadsheet cardcage or hybrid (combination)
System Resource Name architecture?
DUT Pin Name Pin Nr Matrix Col GP Relay CAN H CAN L Star Ground 2. If you decide on card-based
Vbatt J1-1 7b (PS+sense), instruments, should you use an
6b (PS+) embedded PC (one that fits inside
Power Gnd J1-2 X an instrumentation cardcage) or an
Brake J1-3 9 external PC?
Accelerator J1-4 10
CAN H J1-5 X 3. Which switch topology—simple
CAN L J1-6 X relay configurations, multiplexers
Pot1 Vref J2-1 6
or matrices—and which switch
Pot1 Wiper J2-2 5
types (reed relays, FETS or arma-
ture relays) should you use?
Pot1 Ground J2-3 7,8
Pot2 Vref J3-1 2 4. Does a mass interconnect make
Pot2 Wiper J3-2 1 sense for your system?
Pot2 Ground J3-3 3,4
5. Which power supplies and loads
Motor + J3-4 12 3b (load 1)
should you choose?
Motor – J3-5 11 2b (load 2)
Other connections 6. Which measurement and stimulus
PS+Sense 7a instruments should you choose?
PS+ 6a
7. What should you do to minimize
PS-Sense 5a
your hardware costs?
PS – 4a
Motor Load + 3a 8. What should you do to minimize
Motor Load – 2a development time?
Earth Ground 1a 9. What should you do to maximize
Switched Earth Ground 1b X system throughput?
DUT Common X
Star Ground 5b (PS-sense), 10. Which hardware vendor should
13,14 4b (PS-) X X you use?
Width Depth
Full Module
1 MW 269.2 mm
(11 in) D
345.4 mm (14 in) D
Half Module Half Module 421.6 mm (17 in) D
497.8 mm (20 in) D
574.0 mm (23 in) D
1/2 MW 1/2 MW
3/4 MW 1/4
MW
65
When you calculate rack size, you To maximize the re-usability of your Instrument layout
need to decide whether the system test system, keep your future needs
controller (typically a computer) in mind when you choose your rack. When you plan the layout of equip-
and monitor also will be installed in In the future, you may want to add ment in your rack, you will attempt
the rack to display test procedures more instruments and more switches to achieve a number of objectives
and results. If you are incorporating and accommodate bigger devices simultaneously:
a computer and monitor, will you under test (DUTs) that require more
also need a keyboard or mouse for power. To maximize your long-term • Ensure rack stability by carefully
operator inputs? If so, be sure to add flexibility, allow at least 20 percent distributing the weight of system
space for these items into your calcu- extra room in your rack to accommo- components in the cabinet
lations, along with space for a work date instrument additions.
• Make it easy for operators to use
surface. If there is a work surface,
Other questions to consider: the system and be productive
consider the fact that it may prevent
the user from easily accessing any • Minimize magnetic interference
• What are the physical constraints
instrument in the space directly
of the location where your rack will
below the surface. • Provide adequate power and heat
be situated? Will the floor support
dissipation
You may also want to consider your system’s weight? Are door-
including space for accessory ways into the facility tall and wide • Route power and measurement and
drawers to provide convenient enough for the rack you are consid- stimulus signals to the right place
storage for manuals, spare connec- ering? Is there adequate power, as efficiently as possible
tors and other small accessories (see and does the room have adequate
Figure 6.2). Slide-out shelves are cooling to support the additional • Minimize conducted and radiated
useful for attaching loads and other heat created by the system? noise
custom equipment, and they make
• Will your system need to be moved • Ensure operator safety
access easy.
to its final destination? If so
consider using multiple smaller Plan your instrument layout on paper
Figure 6.2. Adding an accessory drawer to your
racks and limiting total rack before you start installing instru-
rack provides convenient storage for manuals,
spare connectors and other small accessories. weight. If you need to ship the ments in your rack, since you will
system to another location, also probably change your layout multiple
consider using ruggedized rack times before you determine the
furniture with strain relief fittings optimal layout.
and keep shipping concerns in
mind (shipping company or airline
size and weight requirements, etc.).
www.agilent.com/find/open
67
You also can use ballast, or weights In some situations, reverse-mounting Instrument access and ease of use
that fasten to the bottom of the (or rear-mounting) instruments in a during testing
rack, to improve rack stability. Most rack makes sense. For example, if you If you are designing a manual or
racking systems offer ballast as an place the DUT interface panel (mass semi-automated system, you need to
option. Ballast mounted at the back interconnect or feedthrough panels) ensure that the operator can reach
of the rack cabinet helps keep the in front of a switching subsystem the necessary equipment controls
cabinet from tipping forward if you that has the plug-in cards facing the and connectors/patch panels without
extend heavy, slide-mounted devices interface panel, you minimize rack straining. Decide whether operators
from the rack or if you place a heavy space, because the switchbox and will sit or stand during testing and
object on a work surface that extends mass interconnect are in the same position the work-surface height
from the rack. plane, and you reduce wire length accordingly. If a test instrument has
from the switching to the DUT. If a display the operator needs to see,
Adding ballast, using retractable the switch box you choose has cards place it at eye level or above, and if
stabilization feet and bolting rack in the rear, you can simply reverse- appropriate, provide the ability to
cabinets to the wall or floor provide an mount the switchbox using the rails tilt the display to reduce glare and
extra margin of safety, but you should on the rear of the rack, as illustrated eyestrain.
not rely on these measures to compen- in Figure 6.5. If you choose to mount
sate for poor weight balance in your an instrument in a non-standard If the operator will interact with a
rack. Always make sure the center manner, be sure the cooling airflow is computer, place the monitor where
of gravity of your system is as low as not disturbed. the operator can see it easily. If the
possible and the weight of your system operator needs to use a mouse or
is centered as much as possible. You may be able to rear mount keyboard, avoid placing these items
shallow instruments behind front- on the same work surface as the
mounted instruments to save rack DUT. Provide for left-handed and
Instrument accessibility and space. This space-saving technique right-handed operators by allowing a
operator ease of use can be a practical way to reduce rack mouse to be placed on either side of
height if you have a problem with low the keyboard.
If your system is fully automated, you doors or you need to meet airline size
may be concerned about instrument requirements. However, mounting When you are planning the operator
accessibility only during system instruments in both the front and work surface, make sure operators
development or troubleshooting. If back of a rack can make servicing sit or stand far enough away from the
your system is operated manually the instruments in your rack more rack that they do not inadvertently
or semi-automatically, an operator’s difficult. hit controls with their feet.
ability to access instruments and use
them easily during testing will be If you plan to ship the rack to
Figure 6.5. Rear-mounting the switching
an important consideration as you another country, consider operator
subsystem reduces rack space and minimizes
decide how to rack your equipment. cable lengths. height and local safety rules, and
make sure adequate preparations are
Instrument access during develop- made for power, cooling and so on
ment and/or troubleshooting before the rack is shipped. Obviously,
When they are low on rack space, local-language instructions may be
system designers sometimes “bury” necessary in some cases. Inadequate
instruments inside the rack behind preparation can sometimes cause
other instruments or mount them long delays in system deployment.
backwards or sideways in the
rack. Before you choose this tactic,
determine if you will need to access
the instrument during system
development to verify operation or
for troubleshooting, repair or calibra-
tion. If you perform periodic system
self tests to verify operation, you may
need access to the front panel of an
instrument, making “buried” installa-
tion impractical.
www.agilent.com/find/open
69
If you cannot create enough airflow Many test instruments draw a fixed Wiring your system
to remove the heat with a fan, you amount of current. However, a Power wires radiate electronic noise
may need to consider air condi- power supply draws variable current and both stimulus and measurement
tioning your rack. There are standard depending on how much power it is signal-carrying wires are suscep-
NEMA enclosures that can be used providing to the device it is powering. tible to this noise, so to minimize
for this purpose. When you calculate heat dissipation interference, separate power wires
requirements, plan around a power from signal-carrying cables. Proper
When you install equipment in your supply’s maximum draw. shielding and grounding techniques
rack, do not block instrument fans or can help alleviate noise problems
side air holes and be sure to follow (see “Grounding and shielding”
instrument manufacturers’ recommen- Routing power and signals on page 72). Selecting the proper
dations regarding air flow and clear- type of cable is also important. A
Once you have resolved the weight
ance around instruments. In general, double-shielded or triaxial cable with
and balance issues, calculated your
place your deepest instruments at insulation between the two shields
airflow and power needs and planned
the bottom of your rack. If you place provides the maximum protection
for operator accessibility, you are
a full-depth, full-width instrument against noise coupling.
ready to turn your attention to how
in the middle of the rack, you block
you will get power and signals to
airflow to the instruments below it. In some cases, you may need to
your instruments and your DUT. Your
separate signal measurement cables
Typical top-mounted extractor fans goal is to route power and measure-
(which can be sensitive to noise)
will move about 200 CFM (cubic feet ment and stimulus signals to the
from signal stimulus cables (which
per minute) of air, which is sufficient right place as efficiently as possible
can generate noise). For example,
for dissipating up to 2500 W of power while keeping noise to a minimum.
if your stimulus signal is a high-
inside a rack. If your system uses frequency square wave with rapidly
more than 2500 W, you could install Multiplexing and matrix switching
changing transitions (fast edges)
additional top-mounted fans or use a Switches, or relays that route power
produced by a function generator,
600 CFM fan in the rear rack door to and interconnect system instrumen-
it will radiate more noise than a
increase air flow. tation and loads to your DUT, are an
square wave with slow edges or a
integral part of most automated and
high-frequency sine wave, and it
If your system includes high-power semi-automated test systems and
would be more likely to interfere
instruments like AC sources or elec- some manually operated systems.
with the accuracy of a low-level
tronic loads with their own fans, use Multiplexers and matrix switches
measurement signal. If possible, keep
ductwork to vent them directly out make it possible to minimize the
wires carrying high-frequency square
the back of the rack. You can make number of test instruments in your
waves and other noise-generating
the ductwork out of sheet metal. system instead of using separate
signals away from your measurement
instruments for each test point.
The amount of power an instrument paths to minimize interference.
Switches deliver power and stimulus
dissipates typically is specified by signals to the DUT when they are For a detailed discussion of ways
the instrument manufacturer. If that needed and route the measurement to reduce noise in switch systems,
specification is not available, you can signals back to your test instruments. see the Application Note 1441-2,
estimate power dissipation require-
Reducing Noise in Switching for Test
ments from the maximum current Choosing the proper switch type and
Systems.
specification using the equation topology will impact the cost, speed,
safety and overall functionality of Wiring dress and termination—Good-
Worst case power (VA) = your test system. For a thorough quality cabling is expensive, but you
Voltage x Amperage examination of switching in test will get the best results if you buy
systems, see Application Note 1441-1, the best cabling your budget will
This calculation provides a conserva- Test System Signal Switching.
tive estimate of power dissipation allow. Make sure the cable you select
requirements because power in is designed for the task you have in
watts, the true source of heat, is mind and be careful not to exceed
always less than or equal to power in the manufacturer’s ampacity rating
VA. It is a good idea to use conserva- of the wires you choose.
tive figures to safeguard against
worst-case situations.
www.agilent.com/find/open
71
Radiated magnetic and electric fields Grounding and shielding For a detailed discussion of
occur whenever an electric charge Grounding and shielding are the grounding and shielding issues, see
is moved or a potential difference two primary methods for reducing Application Note 1441-2, Reducing
exists, and can also be a cause of unwanted noise in a test system. Noise in Switching for Test Systems
noise coupling. In a circuit, high- They often work together, such as and the white paper Considerations
frequency interference may be unin- when the shielding of a cable is for Instrument Grounding.
tentionally rectified and appear as a connected to ground. In such cases
DC error. Switch-system circuitry is it is important to understand where In high-frequency systems, radio
also susceptible to electromagnetic to ground the cable shield in order to frequency interference (RFI ) also
radiation from radio, TV, and other maximize the shield’s effectiveness. can cause problems. To minimize
wireless broadcasts, and it is impor- In some cases, the solution to one RFI, make sure your cable diameter
tant to shield sensitive circuitry from noise problem may reduce the effec- is suitable for the signal wavelengths
these fields. If you want to make tiveness of the solution to a different you are transmitting, terminate all
accurate measurements of low-level noise problem, making it imperative cables in their characteristic imped-
signals in a test-system environment, that you thoroughly understand the ances, keep cable lengths as short as
you need to pay careful attention noise source, method of coupling, possible and use only high-quality
to the details of grounding and and noise receiver so you can make cables and connectors. For more
shielding. the appropriate tradeoffs. information, see the white paper
Proper Cable Shielding Avoids RF
It is always a good idea to have a line When you design a grounding system, Interference Problems in Precision
filter and surge protector in the main your goal is to minimize the noise Data Acquisition Systems.
power distribution unit (PDU) of the voltage generated by currents from
rack. Also, each instrument usually two or more circuits flowing through
has its own line filter, to reduce a common ground impedance, and
Safety and interlocks
conducted interference from the to avoid creating ground loops that It is important to protect the safety
instrument and reduce conducted are susceptible to magnetic fields and of test-system operators, as well
susceptibility to the instrument. But differences in ground potential. as safeguarding your DUT and the
remember, there is still some residual equipment in the rack itself. You
noise that each instrument can inject To accomplish these goals, instru-
need to plan for system safety as part
into the power grid. Sometimes it ment, power and safety grounds
of your overall system design, and
becomes necessary to put an addi- should all be connected as close as
you need to comply with company,
tional power filter on an individual possible to the DUT’s power ground
local, national and international
instrument to reduce its conducted via a “star” mechanism as shown in
safety standards and regulations that
noise. Figure 6.7. This eliminates ground
may apply.
loops and contributes to quiet
readings. Install a system cutoff mechanism
that is activated by any action that
Figure 6.7. A star ground minimizes noise and eliminates ground loops. exposes the operator to potential
harm. Make sure you document
Star ground safety procedures and thoroughly
train operators to use them.
Test system DUT
Mechanical safety
Fans are a potential source of danger
Instrument commons Analog ground
in a test system. Make sure that any
fans you use are covered with fan
PS- Digital ground
guards that make them inaccessible
PS-Rem sense RF ground to human fingers. Positioning fans on
top of rack cabinets, instead of in the
Earth ground Power ground cabinet wall, reduces the chance that
someone’s long hair could get sucked
Other commons in unintentionally.
www.agilent.com/find/open
73
Conclusion
It’s one thing to connect a PC to one
instrument, but when a rack might
contain $100,000 or more worth of
equipment, it pays to do some plan-
ning. Arranging your test equipment
in a rack to maximize measurement
accuracy, equipment longevity and
operator ease of use and safety also
takes careful planning. Whether you
are using your test system for R&D,
design verification or manufacturing
test, you need to consider a variety of
issues, including weight distribution,
heat dissipation, instrument acces-
sibility and operator ease of use, and
you need to pay close attention to
minimizing magnetic interference
and conducted and radiated noise.
Introduction validate your designs, the faster you and program it for optimum speed.
can get your new products to market. After you’ve built your system, you
This chapter discusses hardware In R&D, throughput is seldom an can tweak instrument setups and
and software design decisions that issue because you are not likely to operating procedures to further
affect throughput, including instru- repeat tests on large numbers of optimize its speed.
ment and switch selection, as well as devices or to perform the same test
test-plan optimization and I/O and repeatedly on a single device. In general, your system first needs to
data transfer issues. It also discusses set up a test or configure the proper
ways to optimize your system as you Taking the time to optimize system stimulus and send it to your device
prepare to deploy it. throughput may require some under test (DUT). Then your system
additional investment up front, but needs to actually make the measure-
Throughput is a measure of the time the payoff in lower costs and faster ment on the DUT and transfer the
it takes to test a device or product. time to market makes the investment measurement data back to the
Maximizing throughput is most worthwhile. computer. Figure 7.1 shows typical
critical in high-volume manufac- steps a computer-controlled system
turing, where you have thousands of As Chapter 5, Choosing Your Test- would take to make a measure-
products to test and you want to test System Hardware Architecture ment. (The steps do not necessarily
them as fast as possible. In high- and Instrumentation, pointed out, have to be executed in the order
volume manufacturing, you measure a test system is essentially a group presented.) Each of these steps takes
throughput in terms of devices per of subsystems that work together. some amount of time to execute.
unit time. The faster you test your The hardware you choose for these To optimize throughput, you need
devices, the lower your manufac- subsystems and the software you to analyze how long the steps take
turing costs. In design validation write to make these subsystems in your system and decide which
testing, the speed of each individual communicate and interact have steps you can speed up. Depending
test is not as critical, but test setup a huge effect on your system on your application and budget, you
time is important because you need throughput. If throughput is critical may decide to work only on the steps
to be able to adapt to pinouts that in your test application, you need that have the biggest impact on your
change often. In design validation, to choose equipment with the throughput, or you may decide to
you measure throughput in terms of performance and features required invest the time and money to elimi-
tests per unit time. The faster you can for fast testing and then configure it nate every unnecessary millisecond
in the entire process.
Figure 7.1. Steps involved in making measurements with a typical computer-controlled system
3 8
4 1 6 12 10
BUS
Control commands Computer
Measurement data 2, 5, 7, 9, 11
Steps
1. Tell system where to connect the stimulus 7. Wait for switch to settle
2. Wait for switch to settle 8. Tell measurement instrument what parameter to measure and the
3. Tell stimulus instrument what signal to send to DUT range in which that parameter falls
(parameter, range) 9. Wait for instrument to process command and complete configuration
4. Tell stimulus instrument to send signal 10. Tell instrument to make the measurement
5. Wait for stimulus to settle 11. Wait for instrument to process command and make the measurement
6. Tell switch to send DUT signal to measurement instrument 12. Transfer measurement information to computer
75
In a typical test system, the steps • Perform regular maintenance on Upfront design decisions
with the biggest negative impact your system and its components.
on throughput include instrument Clean fan filters regularly to avoid affect throughput
resets, delays (wait statements) heat build up (high temperatures
If you are designing a new system,
programmed into the system soft- contribute to failures). For
rather than optimizing an existing
ware and waveform downloads. more information on this topic,
system, you will have a greater
Power supply settling time, voltmeter see Chapter 8, Operational
opportunity to maximize your system
measurements and switching also Maintenance.
speed. The system hardware and
play a role. Figure 7.2 shows the
This chapter focuses on improving software architectures, instruments,
hierarchy of delays in a typical test
throughput for systems designed switches, and I/O interfaces you
plan.
with rack-and-stack test instruments. select will have a huge impact on
Obviously, if your system stops However, most of the concepts apply system throughput. For a detailed
functioning, your throughput drops to systems built with card-based discussion of system hardware
to zero. In all phases of product test instruments (such as VXI and PXI) and software architectures, see
(R&D, design validation and manu- as well. Card-based systems do have Chapter 4, Choosing Your Test-
facturing test), therefore, minimizing features that lend themselves to System Software Architecture, and
system downtime is critical to optimizing throughput. For example, Chapter 5, Choosing Your Test-
maximizing throughput. To minimize VXI and PXI backplanes have a System Hardware Architecture and
system downtime: built-in triggering bus that makes it Instrumentation.
easy to implement triggering schemes
• Select instruments from vendors
you trust and choose instruments
that can minimize system delays. Making hardware choices
The new LXI standard (see Chapter
with high mean time between 16) also specifies a trigger bus, thus Figuring out how fast your system
failures (MTBF) specifications. bridging some of the differences will perform measurements is harder
between card-based and rack-and- than it appears. For example, you
• Establish a good spares program:
stack instruments. Card-based and may decide to use a digitizer instead
keep backup components for your
rack-and-stack systems are similar of an oscilloscope to take advantage
system so that if an instrument
in most other regards, and you can of the digitizer’s higher resolution.
fails, you can quickly swap in a
use many of the same techniques for The digitizer may be able to sample
replacement and restore system
optimizing measurement speeds in 1000 readings very fast, but if those
functionality.
both types of systems. readings are transferred to the PC
over GPIB, the total process could
take a relatively long time. If you
Figure 7.2. Hierarchy of delays in a typical test plan
can download a decision-making
Time spent during example test algorithm into the digitizer, you can
send a simple go/no-go result back
to the PC, which would make GPIB a
(% of total time shown in parenthesis)
6
5 (36)
5 reasonable option. However, it takes
4
extra effort to create and download
3.5 (25)
a decision algorithm into an instru-
Seconds
3 2.7 (20)
ment, which may increase develop-
2 ment time as well as “first-run” time
1.4 (10)
1 0.7 (5) 0.5 (4) of the test program. You also need to
0 consider the relative analysis time of
Instrument Wait Arbitrary waveform Power supply DMM Switching
resets statements downloads settling readings a routine computed inside the digi-
Source of time spent tizer compared to the time required
to complete it inside the PC.
www.agilent.com/find/open
77
Power supplies delays. For details on using this tech- A good example is testing a DC-to-
Your choice of power supply nique and other techniques, see 10 DC converter with four outputs,
can dramatically impact system Hints for Using Your Power Supply where you need to measure the input
throughput, because waiting for to Decrease Test Time, publication voltage to the converter and each of
power supplies to settle is typically number 5968-6359E. the four outputs in order to fully test
a time-consuming element in a test the device. If you have a single DMM
plan (see Figure 7.2). Check the Another way to reduce test time is to to measure the voltages, you’ll need
settling time specifications of the choose power supplies and electronic a multiplexer to sequence through
power supplies you are considering loads that have built-in measurement the measurements (see Figure
for your system. If you can’t find a features. With power supplies, these 7.3). In addition to the complexity
specific reference to “settling time” capabilities let you measure the of this setup, your test program
on the data sheet, look instead for supply’s output voltage and current. needs to wait for the multiplexer’s
the “programming speed,” “program- With loads, you can measure load switches to move and settle for each
ming response time,” or “rise and input voltage and current. measurement.
fall time” specification. Programming
speed is defined as the amount of
Figure 7.3. Testing a four-output DC-to-DC converter with a single DMM requires a complex multi-
time it takes for the instrument to
plexing scheme and can involve significant delays.
reach a specified percentage of the
voltage setting (typically within 0.1 DC to DC
percent), not including command converter
processing time. Rise and fall times + + Load 1
– –
are typically defined as the time it DC source
+
DC outputs
takes to get from 10 percent of the +
DC inputs
– – Load 2
+ +
final value to 90 percent of the final
– – + + Load 3
value for the rise time, or vice versa – –
for the fall time. Because of the + +
– – Load 4
different terminology and definitions,
you must be careful when comparing
settling times in power supplies from MUX
different vendors. +
–
When you are trying to boost +
– + + DMM
throughput in time-critical produc- + – –
–
tion test systems, look for a multiple- +
–
output supply that can set multiple
+
outputs with a single command, like –
the Agilent N6700 series. Otherwise,
consider using multiple single-
Figure 7.4. By using the built-in measurements in your DC power source and electronic loads, you
output power supplies instead of
can eliminate the DMM and MUX and significantly increase your test speed.you can eliminate the
one multiple-output supply. With DMM and MUX and significantly increase your test speed.
multiple-output power supplies, the
instrument takes extra time to parse DC to DC
commands, because you are sending converter
an additional parameter to indicate +s
+ + Load 1
which of the multiple outputs DC source – –
–s
it should use. Also, with most
multiple-output supplies, commands +s
sent to the various outputs are + + Load 2
DC outputs
+s
DC inputs
– –
processed sequentially, one output + + –s
– –
at a time (this can be avoided with –s
+s
the Agilent N6700 series). With + + Load 3
– –
multiple supplies, one supply can be –s
processing a command while the next
is receiving a command, so you avoid +s
+ + Load 4
– –
–s
Figure 7.5. This diagram shows what happens when you tell a switch to close, take a measurement,
and then reopen. The “switch drive signal” represents the actual voltage that causes the switch to change
states. The resulting “measured signal” is connected from the DUT to the measurement instrument.
Switch timing
Source settling
and input settling
Switch bounce Measurement trigger
safety margin Perform measurement Coil falltime delay
Coil risetime delay Measurement complete
Data to buffer
Time
Mechanical Switch Open Mechanical
delay bounce command delay
www.agilent.com/find/open
79
Electromechanical switches such bigger impact on speed than the close. In some cases it is worth the
as reed and armature relays are switch topology you choose, unless extra money for the test time you
common in low-speed applications. you factor in the time required for save.
They are capable of switching high reconfiguring a switching system
voltage and current levels, but they (which, as we noted earlier, is more Controller issues
are limited to switching rates of critical in design validation applica- Unless your PC is ancient, its
dozens of channels per second for tions.) If you use a switch matrix, processor speed is not likely to
armature relays up to hundreds of you will be able to quickly and easily be a significant factor in your test
channels per second for reed relays. expand and reconfigure your system throughput. Typically, issues associ-
Reed relays are excellent choices to as your test needs change. Expanding ated with stimulus and measure-
connect measurement instruments and reconfiguring systems that use ment instruments, power supplies,
and low-current stimulus to your multiplexers typically is more time switches and test software play a
DUT. They are relatively fast (see consuming. much bigger role in determining
Table 7.1), although they can have a system speed. Your PC is not in
higher thermal offset voltage than A matrix arrangement of reed relays control of data collection speed, and
armature relays. Armature relays provides an excellent way to allow faster PCs don’t necessarily collect
are slower, but you can use them for any instrument to be connected to data any faster. The PC’s interface to
higher current loads. When you use any pin on your DUT, and it permits your test system (GPIB, LAN, USB,
armature relays, group your tests so easy expansion as you add new FireWire, VXI or PXI) will certainly
the relays stay connected to perform instruments to your system or more impact data transfer time, but that
as many readings as possible at one pins appear on your DUT. Matrices is not dependent on PC processor
time. use more relays than multiplexers, so speed. If you are using a LAN or USB
they tend to cost more. If you don’t interface, we recommend using the
Electronic switches, such as field- need to connect multiple instruments highest speed interface and switches/
effect transistor (FET) and solid- to any pin, a multiplexer is a suit- hubs available.
state relays, are frequently used in able solution. If you have a 1 x 20
high-speed applications. (Typically multiplexer for example, you can take Processor speed is a factor only
for voltage or temperature measure- a test instrument and connect it to 20 if you are relying on your PC for
ments). However, some FET elec- pins, but you can’t hook anything else analyzing data and if you are using it
tronic switches cannot handle high to those 20 pins. With those same 20 for your software development. You
voltage or current, and they must be relays in a matrix, you can connect want to use the fastest PC available
carefully protected from input spikes four instruments to five pins in any when you are compiling programs,
and transients. Check the electronic combination. but of course you do not have to
switch ratings carefully. do your development work on the
If you want the ultimate in same computer you use to run your
Switching topologies can be divided throughput and your budget is not system.
into three categories based on their limited, you can use multiple test
complexity: simple relay configura- instruments instead of a switching
tions, multiplexers and matrices. scheme for making measurements on Designing your test plan for
The best one to use depends on the multiple test points. With multiple speed
number of instruments and test instruments, you can set each to
points, whether connections must be the needed range and eliminate the Many test programs spend most
simultaneous or not, cost consider- time spent on configuring the test of the time waiting. Even if you
ations and other factors. Typically, instrument range, as well as the time have selected the fastest-available
the type of relay you choose has a required for switches to open and hardware for your system, software
issues can slow your test-system
Table 7.1. Relay comparison chart throughput significantly. While you
can tweak your test-system program-
Armature relay Reed relay Solid-state relay ming after your system is complete
Switch speed 50/s 1000/s 1000/s (see “Fine-tuning your system for
Contact resistance Low Very low High speed” on page 84), you will achieve
Life 1 million 10 million >10 million better throughput if you design your
Typical failure mode Fails open Fails open Fails shorted test plan up front to optimize test
Typical max input 250 V/2 A 100 V/100 mA 250 V/10 A
sequencing and minimize delays.
Measurement setup
Program step Input conditions(stimulus to DUT) (to measure signal out of DUT) DUT measurements taken
1 Set input condition 1 (e.g., amb. temp. = 0 degrees C)
2 Prepare measurement setup 1 (e.g., voltage)
3 Test point 1 voltage
4 Set input condition 2 (e.g., amb. temp. = 25 degrees C)
5 Test point 1 voltage
6 Set input condition 3 (e.g., amb. temp. = 55 degrees C)
7 Prepare measurement setup 2 (e.g., current)
8 Test point 1 current
9 Set input condition 1 (0 degrees C)
10 Prepare measurement setup 1 (voltage)
11 Test point 2 voltage
12 Set input condition 2 (25 degrees C)
13 Test point 2 voltage
14 Set input condition 3 (55 degrees C)
15 Prepare measurement setup 2 (current)
16 Test point 2 current
17 Set input condition 1 (0 degrees C)
18 Prepare measurement setup 1 (voltage)
19 Test point 3 voltage
20 Set input condition 2 (25 degrees C)
21 Test point 3 voltage
22 Set input condition 3 (55 degrees C)
23 Prepare measurement setup 2 (current)
24 Test point 3 current
www.agilent.com/find/open
81
If you organize the program to Figure 7.6. To minimize overall test time, structure test loops so that the most time-consuming
minimize changes to the stimulus operations are performed the fewest number of times.
conditions and measurement setups,
overall test time is reduced. Note
Test flow diagram — nested programming loops
that the sequence shown in Table 7.3
provides exactly the same number
and type of DUT measurements Loop 1
under exactly the same set of input Contains stimulus conditions that take a long time
conditions as the previous sequence, to change or set up, for example:
but the overall number of program-
ming steps has been reduced from 24 Step through a variety of DUT temperatures
to 14. Also, the number of stimulus
changes has been reduced from 8 to Loop 2
2, while the measurement setup has
Contains test variables that take less time
gone from changing back and forth 5
to change or set up than those in loop 1, for example:
times to changing just once.
Step through a sequence of high and low DUT bias
Organizing nested loops voltage combinations
Structure the basic test flow so that
slow operations like setup, DUT Loop 3
connections and temperature settings
Contains test variables that take the least time
are in the outermost loop. Nest faster
to change or set up, for example:
operations like one-button measure-
ments in lower-level loops. Place your Make a DUT voltage measurement
fastest operations in the lowest-level
loop. You can use a test flow diagram,
as shown in Figure 7.6, to get a better
conceptual understanding of the
test plan and prevent wasted time
in nested loops and poor use of DUT
connects and re-connects.
Measurement setup
Program step Input conditions (stimulus to DUT) (to measure signal out of DUT) DUT measurements taken
1 Set input condition 1 (0 degrees C)
2 Prepare measurement setup 1 (voltage)
3 Test point 1 voltage
4 Test point 2 voltage
5 Test point 3 voltage
6 Set input condition 2 (25 degrees C)
7 Test point 1 voltage
8 Test point 2 voltage
9 Test point 3 voltage
10 Set input condition 3 (55 degrees C)
11 Prepare measurement setup 2 (current)
12 Test point 1 current
13 Test point 2 current
14 Test point 3 current
www.agilent.com/find/open
83
In some test systems, I/O speed is not For more information about I/O Fine-tuning your system
a major determining factor in overall and its effect on system throughput,
throughput. This is especially true see Chapter 2, Computer I/O for speed
in RF systems, where the network Connectivity Considerations, and
Whether you are turning on a new
analyzer or spectrum analyzer Application Note 1475-1, Modern
system or fine-tuning an existing
may take some time to complete a Connectivity—Using USB and LAN
system, there are a number of
measurement. However, in systems Converters.
techniques you can use to improve
that rely on unprocessed data, or
Keep in mind that if your instru- throughput. Relatively small
when real-time control is important,
ment’s throughput is slow, you are adjustments to system software,
your choice of I/O for the connection
not going to get greater throughput instrument setups and operating
between your computer and your
by changing to a faster I/O interface. procedures can help you optimize
test system hardware can have a big
You can improve your throughput your system speed.
impact on the overall test time.
by minimizing the number of
While high-speed LAN and USB have GPIB transactions you send. When Minimize delays
much higher throughput than GPIB, possible, send multiple GPIB
the serial nature of these interfaces commands at one time. This reduces As Figure 7.2 noted, delays (wait
results in performance that may be bus turnaround times and allows statements) programmed into system
similar to GPIB for highly transac- the instrument, in some cases, to software typically cause systems to
tional operations in which you are operate on the commands as quickly run at suboptimal speeds. When you
not waiting for the instruments. The as possible. run a test program there are some
extra cost to use gigabit LAN and Hi- operations — such as measuring
Speed USB is relatively low and will The character format you use to a complex signal or moving data
result in noticeable speed improve- transfer data can also affect the to an array — that take additional
ments. Note that a LAN will run at data transfer rate. You can choose time to complete before the next
its fastest if you make a direct socket from a variety of general formats, command can be executed. If these
connection. including character string, ASCII, or operations do not complete before
binary. Binary code is handled as the next command in your program
Connection to a card-based system bit streams, typically in block-length is executed, errors can occur
such as VXI or PXI should usually message units. These message units and the program may halt. When
be done with a fast interface such are more compact than those made debugging test routines, program-
as FireWire or MXI, as the register- up of string and ASCII characters mers frequently “fix” the problem
based cards generally have minimal and therefore they can be transferred by programming in a delay after
processing capability on-board and more quickly. the operation and before the next
count on a fast interface for good command. This is fine as a temporary
performance. For example, when you are down-
fix for correcting an error, but it is
loading a data file for an arbitrary
important to remove the delays, or
Table 2.1 in Chapter 2 showed the waveform to a function generator,
at least to make them as short as
relative speeds for various operations downloading floating-point values
possible, once you find the real cause
for a stimulus instrument having (a character string) is slower than
of the measurement problem. Leaving
GPIB, USB and LAN interfaces. As downloading binary values, but using
unnecessary delays in a program
you can see from that table, the floating-point values is more conve-
slows down the overall system
instrument’s internal speed clearly nient when creating the arbitrary
throughput.
dominates setup changes, making waveform. Here, you need to decide
I/O choices seem moot, but download which is a higher priority, faster An alternative to using a delay is to
speeds are much better with LAN data transfer (binary), or ease of use use system-level control commands
and USB when large amounts of data (floating-point values in the form of a such as *OPC? (operation complete)
are involved. character string). to inform the control software that
an operation is complete, which is
especially useful for variable-length
operations. Many instruments are
IEEE-command compliant which
means they are able to use the
www.agilent.com/find/open
85
Counters • Shield the measurement setup Power supplies
• Use ASCII format for fastest to reduce noise pick-up from the • If your power supply has list mode,
throughput (note: this is different operating environment. Shielding use it to store complete instrument
from other instruments) may allow you to make measure- setup states and recall them with
ments with shorter measurement a single command, rather than
• Select the trigger level instead of times (aperture) or with less sending a long series of configura-
using auto level filtering and still achieve sufficient tion steps.
noise rejections to obtain the
• Use the auto arming mode • Use the built-in measurement
required accuracy.
capabilities.
• Disable printing operation
• Try to make all readings with the
DMM “LO” terminal connected to • Use power supplies with downpro-
• Define the trigger command so the
circuit low. DMMs have fairly large gramming capability.
fetch command does not need to be
sent for every measurement values of capacitance between “LO”
and earth which must be charged RF/microwave sources and analyzers
(increases settling time) when you • Agilent application notes offer
• For some measurements, a counter
make floating measurements. many tips and tricks can be used to
may produce readings in which the
speed up measurements with RF/
last few digits are not stable. This
can slow a test if a human operator Scopes and digitizers microwave sources and analyzers.
• If you are importing raw data, use See www.agilent.com/find/open.
needs to discern the difference in
readings. Truncating the last digits binary transfer mode. Specifically,
will produce a more understand- use byte or word formats. Word Conclusion
able display, but some tests require format is more accurate but
requires twice as much data to be To maximize system throughput, you
that extra resolution. Have the
sent over the bus. Some scopes need to choose the right equipment
counter calculate the arithmetic
produce more than 8-bit resolu- and program it for optimum speed.
mean if you require high resolution
tion, but many acquisition modes The system hardware and software
and a stable reading or use a limit-
produce only 8-bit data. In these architectures, instruments, switches,
testing mode.
cases, transferring word versus and I/O interfaces you select have a
Digital multimeters byte data will take twice as long huge impact on system throughput.
• When using a scanning meter and not provide any additional If you carefully evaluate the complex
such as the Agilent 34970A, wire resolution. It is important to know interplay of the hardware and soft-
adjacent channels so that the DMM how and when the instrument ware elements of your test system,
doesn’t have to switch modes or produces extra resolution. you will find many opportunities
ranges for improving the speed with which
• Capture only as much data as you your system performs measurements.
• Select the shortest channel delay need to analyze. After you’ve built your system, you
(zero) can tweak instrument setups and
• Turn off special features such as operating procedures to optimize
• Turn off scaling mask test, jitter analysis and FFT speed. The time you spend doing
functions if they aren’t needed. so will help lower your costs and
• Turn off alarms accelerate your time to market.
• Make sure you have an adequate
• Use the fast filter trigger rate, and use the fastest
sweep speed (timebase scale) that
• Turn off T/C (thermocouple) check. is consistent with your applica-
Some scanning meters will check tion. Long acquisition times
for the existence of a thermocouple and/or slow trigger rates can limit
by looking for a short circuit before your throughput if your analysis
attempting to read the thermo- program is very fast.
couple voltage.
2 01
T1 3 6 01
T2 4 7 01
T3 5 8 02
N 6 9 02
7 10 02
10A
8 11 03
10A
9 12 03
LL1 J3 10 13 03
LL2 J4 14 NEU
NEU (A) NEU
16 NEU
GRD Control
board
J1 J2 Grn
Outputs enabled switch 1 3 7
Line/EMO error indicator
2 4
8 Yel
Emergency shutdown 5 J2 Outputs enabled
(EMO) switch 6 9
indicator
1 3
Outlets/
Branch in 2 4 Branch out receptacles
87
Cooling In non-air-conditioned factories, Logistics and ergonomics
temperatures sometimes may exceed
Fans are another problem area when the ability of simple fans to keep The doorways in many older
line voltage varies. A 240 V fan may the instruments operating within European buildings are short, and
work when operated at 120 V, but their specifications. In this case, a 2-meter rack may not fit through
at a much lower speed. Thus, the consider a dedicated air conditioner the doors. Taller racks also require
airflow may no longer be sufficient to for the system. NEMA enclosures are larger aircraft to transport them. If
cool the system. Conversely, a 120 V available for a wide variety of rack you build tall systems, your shipping
fan may burn up when connected to sizes. These completely enclose the costs may be significantly higher over
240 V. It can be a nuisance to replace system, and provide a way to attach the life of the system if it is moved or
the fans every time the system is air conditioner intake and exhaust. shipped frequently.
shipped from country to country. But Appropriate ductwork must also be
fans that can be operated from any added to the factory. See In some Asian countries where real
line voltage are produced in smaller www.nema.org. estate is in scarce supply and space
quantities and are thus much more is at a premium, facility aisles and
expensive than single-voltage AC hallways are extremely narrow. It
fans. Line frequency may be difficult or impossible to
move a deeper- or wider-than-normal
DC fans, though, can be an excel- The frequency of AC line voltage system to its intended location. Once
lent choice for systems that must varies in different parts of the world. positioned, it could be difficult to
be moved often. A small, fixed 12 In the U.S., 60 Hz is standard. In open front or rear doors.
V or 24 V DC power supply with many other countries, it is 50 Hz.
universal AC input (i.e., 100-240 VAC) While this won’t affect most modern The average population height varies
can be installed in the system and power supplies, it can certainly affect country-to-country, too. Use care
connected to the DC fan(s). Other signal measurements. It is common to place keyboards and monitors at
advantages of DC fans are: to take low-noise DMM readings an elevation that is not too high for
with a “1-line-cycle” integration shorter operators. It is also a good
• More control over airflow and time. At 60 Hz, this is 16.667 ms. idea to provide keyboard/mouse
noise. The speed of the fan is At 50 Hz, it is 20 ms. Some DMMs, trays with adjustable heights and
directly related to the input such as Agilent’s 34401A, automati- provisions for left or right-handed
voltage. A 24-volt DC fan can cally adjust their integration time operators. Your safety department
typically be operated between 12 based on internally measured line can provide you with up-to-date
and 28 volts DC. At 12 volts DC, frequency. Others must have this guidelines for ergonomic standards.
the fan will operate at half speed, information programmed into them.
producing less air and less noise. It is important to set your DMM
correctly based on line frequency.
• The life expectancy of a DC fan
is higher than that of a compa- At lower frequencies, the magne-
rable AC fan, since DC fans are tizing current of transformers and
many times more efficient. The motors can go up, even to the point
correspondingly low heat dissipa- of saturating the core. This can
tion reduces the thermal load on cause nonlinear magnetic fields and
the bearings, thereby increasing overheating of the core, especially
lifetime. at 47 Hz, creating a situation where
products designed in a 60 Hz envi-
ronment can cause problems in other
parts of the world.
88 8. Operational Maintenance
Calibration a method for notifying appropriate accuracy I need in both places, and
personnel when the next calibration are the calibration services
Most electronic instruments require due date arrives. You could simply adequate?
periodic calibration that is traceable place a dated sticker in a conspicuous
to a government standards agency place on the instrument whenever it
such as NIST (National Institute of is calibrated (see Figure 8.2), and have
In-house calibration lab
Standards and Technology) in the U.S. someone check dates periodically, or If you do not already have an
This requirement guarantees that you could program the system with in-house calibration department
measurements meet their published “due” software that notifies appro- (see Figure 8.3), you might consider
accuracy specifications. Calibration priate personnel automatically. setting one up, although the cost and
is not the same thing as diagnostics, Figure 8.2. Cal sticker time to do so can be considerable.
which are simple tests to verify that If you intend to offer calibration
the instrument is operating and LOVELAND STDS LAB
17025 services to others outside your
taking measurements that are at DD MM YY BY/NO. company, your customers may
least close to what they should be. CAL require you to have international
Diagnostic tests and fixtures are DUE accreditation. A good place to start
discussed in the next section. is the International Laboratory
Accreditation Cooperation (www.ilac.
It may seem logical to build calibra- In addition to regular calibration, org). Members of ILAC, such as the
tion fixtures that would allow your keeping a log is a good practice. American Association for Laboratory
system to be automatically calibrated It’s helpful to be able to correlate Accreditation (A2LA—www.a2la2.net),
without having to remove equipment. manufacturing anomalies to the will certify your lab after you have
Unfortunately, such fixtures would be particular operator, time of day, met their requirements, a process
prohibitively expensive. Calibration calibration period, run number and that can take from four to nine
requires use of components that meet to many other manufacturing months once the lab is fully opera-
stringent specifications under closely variables. Before you build that test tional. If you desire to have your lab
controlled conditions of temperature system in Germany for shipment to accredited, the international stan-
and humidity. Oil-baths containing Thailand, for example, try to answer dard ISO 17025 will apply. It is not
“standard” resistors at controlled these questions: Do I have the same necessary to become accredited, but
temperatures, frequency-measuring calibration system in both places? If at the least, you may wish to become
equipment that connects to the NIST not, can I guarantee the measure- ISO 9000 certified.
cesium-beam frequency standard and ments made by my test system here
the like are not easily contained in a will be the same after I ship the test
removable fixture. system overseas? Can I get the
There are three ways of assuring that
a test system is calibrated: Figure 8.3. Calibration lab
www.agilent.com/find/open
89
Contract services Diagnostics and Repair • Some instruments can do limited
internal automatic calibration
For a broad range of calibration Perhaps the hardest thing to do once (sometimes called “auto-adjust-
services covering many types of you have a test system finished is ment”). This automatic procedure
instruments, professional instrument to spend some extra time designing should be done periodically, but
calibration services are available a diagnostics test program that can not necessarily every time diag-
from Agilent. See www.agilent. help locate the source of problems nostic test programs are run. Keep
com/find/calibration for details. when they arise. But it is time well a programmatic calendar to remind
Non-Agilent equipment is included. spent. Here’s what to do: the operator to run such programs
Contracted services can be arranged when the due date occurs (usually
in various levels, from single instru- • Execute a self-test on every instru- about every 30 days).
ments on an as-needed basis to ment that has this capability.
scheduled volume on-site calibration • Attach a known good device under
(VOSCAL). • Measure the output of every test (DUT) to the system and run
stimulus device with an appro- a full suite of tests on it. This
priate measurement device to technique is not foolproof, since
Swap and return verify that all instruments are characteristics of such a “golden
working and taking readings that DUT” can change over time as
The third method of calibrating are nominally correct. This is not components age. A useful way to
equipment is to simply replace units sufficient to guarantee that they counter this effect is to periodi-
when they need calibration with are in calibration, but it is good cally run a “Gage Reliability and
others that are still within their enough for a diagnostic tool. Reproducibility” (Gage R&R) test
calibration period. This requires
on the system. There are two
keeping one or two spares on hand, • Feed a small DC voltage from a
sources of variation in any system:
which can be expensive. However, it stimulus device (digital-to-analog
the variation of the product and
is a good idea to keep some spares converter, power supply, etc.)
the variation of the measurement
handy anyway if system uptime is successively through all internally
system. The purpose of conducting
critical, as the next section discusses. available switching paths and
the Gage R&R is to be able to
There is one caveat in swapping back to a DMM. This verifies the
distinguish between the two so as
instruments: A replacement may switching subsystem.
to reduce the measurement system
be completely within its calibration
• Create a special diagnostic fixture variation if it is excessive. This
specifications, but if it is operating
that loops signals that cannot be means running a large quantity of
at the opposite end of its calibration
automatically connected internally known good boards on the system
range from the original instrument
back into the system. Use the same periodically to obtain a statistical
and the production device being
procedure previously described to sampling that can be compared
tested is already near its limit, a
measure continuity of these paths. to reference data to see if there is
statistical variation could result
any long-term drift in the measure-
that is large enough to cause a yield
• Read switch cycle count informa- ments. Such a study can also be
problem. The solution is to run a
tion from any switch box that has used initially to study the measure-
statistical analysis on the results.
this capability. This data can give ment statistical parameters, which
This analysis is called a “Gage R&R”
you early warning of relays that are can be used to set acceptable upper
study, and it is covered in the next
nearing the end of their specified and lower limits on each test. Look
section.
lives. for statistical process control (SPC)
and statistical quality control
(SQC) software tools that can help
you create such data.
90 8. Operational Maintenance
In a production environment, • Maintain an inventory of spare Cleaning
diagnostics can be run daily or at instruments. This can be expen-
the beginning of a shift. In a design sive, but so is a down production Maintaining good airflow through
validation or R&D environment, line. Remember, too, that the cost your system is essential, because it
running the test once a week or less of many plug-in cards for PXI and keeps the temperature under control,
may be adequate. Once a problem is VXI is greater than an equivalent assuring that instruments are
identified, the next step is to fix it. rack-and-stack instrument because operated within their temperature
There are several things you can do rack-and-stack instruments specifications. Many instruments
to ensure fast repair: typically are produced in higher have removable air filters, so be
volumes. Thus, it is less expensive sure to inspect these regularly and
• Make it easy to replace instruments. to inventory spares of box instru- clean or replace them when neces-
Make sure that mounting screws ments, and they can double as sary. Some racks are also available
are not hidden, that cables are debug tools when not in use inside with air filters. These should also
easily removed from the instrument a system. be inspected regularly. Keep cables
(and labeled so they are replaced away from the filters. If cables must
correctly), and that instruments are • Place more than one of a key be moved in order to reach the filters,
not hidden inside a rack, necessi- instrument in your system when the flexing can make the cables
tating removal of other instruments you design it. For example, eventually break, causing reliability
in order to get to them. an inexpensive DMM could be problems unrelated to dirty air.
integrated into the system for use
• Although PCI slots in a rack- during manual debug, but pressed If many operators will be using the
mounted computer are tempting into service should the main, high- system, it is a good idea to peri-
spots to put instruments (since speed DMM require service. With odically clean the keyboard, mouse,
they do not take up additional IVI drivers, such interchangeability barcode reader and touchscreen, as
rack space), remember that should not require a change to the applicable. You generally can use
removing the computer from the software. simple household cleansers. Disease
rack to get to them is tedious and can be spread easily from one person
time-consuming. • Heat and thermal gradients are to the next via these devices. Trained
enemies of any test system. Provide operators may be hard to find, so
• Use a limited set of custom cables adequate airflow to minimize heat keep them healthy!
and keep spares on-hand in case rise, and avoid a situation where
they need to be replaced. Use you are continually changing the
standard, easily available cables thermal environment of the test
whenever possible. equipment.
• Fixture connectors can wear out
over time. Have a good stock of
replacement connectors available.
www.agilent.com/find/open
91
Upgrades and expansion Conclusion
If you’ve designed your system well, Test systems have made the task of
using the concepts highlighted in repetitive testing both faster and
earlier chapters, it will be able to more reliable, but there’s much to
handle new instruments easily. consider to keep them running. You
You’ve left extra space in the rack must factor in worldwide power
for additional or bigger instruments, issues, calibration, diagnostics,
and you’ve allowed expansion room repair, cleaning, upgrades and expan-
in your switching or instrumenta- sion. At Agilent, we appreciate the
tion cardcage if present. You’ve talent and effort required to design,
also designed the switching system build and implement exceptional
in such a way as to allow instru- test systems. If you are creating a
ments to be added to the system by test system or need help with one
simply plugging the new inputs and you already use, you can find lots of
outputs into a place you’ve reserved advice at www.agilent.com/find/open.
for future instruments (such as the
unused rows of a switching matrix,
as described in Chapter 5, Choosing
Your Test-System Hardware
Architecture and Instrumentation).
You’ve got room in your fixturing
system for more pins, and you’ve
developed a small set of reusable
cables to connect those into your
instruments and switches.
92 8. Operational Maintenance
Section 2. Networking Choices
93
94 Section 2. Networking Choices
9. Using LAN in Test Systems: The Basics
Introduction Coping with complexity Figure 9.1. PC and test instruments in a rack
This chapter provides an introduc- The basic purpose of any test system
tion to the essential elements of is to characterize and validate the
local-area networking (LAN), the performance of electronic compo-
basic attributes of test systems, and nents, assemblies or products. The
the benefits of using a LAN interface complexity of this task depends on
for control and data transfer in a variables such as the physical nature
system. of the device under test (DUT), the
number of tests to be performed, the
number of signals to be measured
and the desired time per test.
95
Setting the standard Defining key attributes and The other essential elements of a
elements LAN are the hardware devices that
Today’s most pervasive computer control, manage, direct and amplify
networking standard goes by a few Wired LAN connections are made the data being sent between other
well-known names: IEEE 802.3, with UTP cables called Category devices on the network.
Ethernet, 100Base-T, 1000Base-T, 5e, commonly referred to as
or Gigabit Ethernet. Some variant • Adapter. This refers to the LAN card
Cat 5e, which is the name of a
of this standard is almost always and connector in a PC (and some
wiring standard defined by the
used when PCs share files, exchange new-generation test equipment)
Telecommunications Industry
e-mail, access the Internet and so on. that provides an electrical inter-
Association and Electronics
With steady improvements in cost, face to the network.
Industries Association (TIA/EIA).
speed and functionality, Ethernet has (Cat 5e replaces the Cat 5 standard • Hub. A small, standalone unit that
achieved virtually universal adoption that has been in use with LANs for a connects multiple devices. Hubs
for local-area networking (to the number of years.) A CAT5 LAN cable use a broadcast model to transmit
extent that LAN and Ethernet are contains four pairs of copper wire data, a method that reduces the
sometimes used as synonyms). and uses locking RJ-45 connectors at effective bandwidth (or data rate)
both ends (Figure 9.2). It is largely when network traffic is heavy.
Devices are often described as immune to interference and crosstalk
being 100Base-T, 1000Base-T and and can support data rates of up to
100/1000Base-T. The number 1000 Mbps.
indicates the data rate in megabits
per second: 100Base-T is 100 Mbps
and 1000Base-T is 1000 Mbps;
100/1000Base-T devices are compat- Assessing wireless LAN • IEEE 802.11a: Uses 5.0 GHz transmis-
sions to send data at up to 54 Mbps
ible with both standards. The T alternatives and with an indoor range of 25-75
indicates unshielded twisted pair
In many companies, an increase in feet. 802.11a is not compatible with
(UTP) wiring to differentiate it from
workforce mobility has led to greater 802.11b and g because it uses a
older standards that used coaxial
demand for flexible networking different modulation method.
cable.
solutions, most notably wireless LAN
• IEEE 802.11n: An emerging standard,
Today, 100 Mbps technology is the (WLAN). As with wired LAN, there is an
backwards compatible with 802.11b
most widely deployed standard and evolutionary series of standards that go
and g, which should offer data rates
provides ample performance for most by various names, but all are generally
up to 10 times faster than 802.11a
uses. Consistent with its history, the known as Wi-Fi (short for “wireless
or 802.11g and up to 50 times faster
standard continues to evolve: Gigabit fidelity”). The four main standards are
than 802.11b.
Ethernet was standardized in 1998 described below in order of commercial
and is now widely deployed and introduction: For test systems, WLAN can enable
10 Gigabit Ethernet has been more measurements in remote or hazardous
recently standardized. • IEEE 802.11b: Uses radio transmis-
settings and provide an alternative to
sions at 2.4 GHz to send data at up to
costly cable runs. However, none of
Tremendous competition among 11 Mbps and with an indoor range of
the current standards can match the
vendors of Ethernet-based LAN 100-150 feet.
combination of speed, reach and noise
devices and cables has driven down
• IEEE 802.11g: Uses 2.4 GHz transmis- immunity possible with a 100Base-T
prices and driven up the volume
sions to send data at up to 54 Mbps wired LAN. What’s more, WLAN signals
of products sold. The net result is
and with an indoor range of 100-150 are susceptible to interference from
a wide selection of high-quality,
feet. It interoperates with 802.11b. other devices that operate in the same
low-cost solutions for local-area
(Some vendors also offer proprietary frequency range, including cordless
networking.
extensions that can provide higher phones and microwave ovens. Wi-Fi
performance.) signals may also interfere with the
testing of wireless DUTs.
Figure 9.2. A CAT5 LAN cable with RJ-45 Figure 9.3. A simple network with one switch and multiple PCs in a
connector. star topology.
Switch
www.agilent.com/find/open
97
Using LAN in test systems This situation is driving the addition 33220A function generator (Figure
of LAN connectors and adapters to 9.5), which has LAN, USB and GPIB
The computer in a test system plays current- and next-generation test interfaces built into its rear panel.
three important roles, two of which equipment. The LXI Consortium has Agilent recognizes the pervasiveness
rely on the I/O connection to the test chosen LAN as the interface for test of GPIB in existing test systems and
equipment: equipment (for more information on will continue to support it. At the
LXI, refer to Chapter 16). The inclu- same time, we are also committed to
• It provides fast, reliable control sion of both LAN and GPIB may be providing LXI compatibility in new
by sending commands, config- quite common for the next few years. instruments, making it the most prev-
uring instruments, reading One recent example is the Agilent alent interface in the near future.
status messages and initiating
measurements.
• It gathers test data—raw or prepro- Figure 9.4. An example test system that utilizes a PC, several instruments, a LAN router, a LAN/
cessed—from the instruments and, GPIB gateway and three types of I/O
if necessary, stores it for postpro-
cessing and archival purposes. USB
Communicating with Figure 9.6. The virtual front panel of the Infiniium oscilloscope enables
instruments browser-based interaction with the instrument.
www.agilent.com/find/open
99
Conclusion
Fast and inexpensive LAN tech-
nology has achieved widespread
adoption in the computer world and
is now shaping the future of test
system development and operation.
LAN-based systems provide several
advantages for test-and-measurement
applications: lower-cost hardware
and cabling; pervasive availability
throughout most enterprises; remote
or shared system control; fast data
transfers; file, drive and printer
sharing; and browser-based interac-
tion with individual instruments.
The LXI standard standardizes the
LAN protocols and connectors as
well as a host of other items for the
test and measurement industry, thus
ensuring easy interoperability of
instrumentation.
The decision to use LAN in a test system from internal and external For a system on an island, the biggest
system delivers important benefits to risks. Our recommended starting risks come from human interfer-
your company and your team. From a point is to create a protected, private ence. System errors might arise if
business perspective, intense compe- LAN for the test system. The stan- a configuration change is made via
tition among equipment vendors has dard capabilities of most Microsoft the front panel or if two instruments
produced a wide selection of high Windows PCs and many low-cost are set to the same GPIB address.
quality, low-cost solutions for local networking products enable two These problems are easy to fix and
area networking. From an organiza- viable approaches, one router-based the integrity of the system remains
tional view, widespread use of LAN and the other PC-based. Several intact.
technology simplifies connectivity factors will influence your choice,
and your decision has implications
and enables new levels of commu-
for the selection and configuration Recognizing potential threats
nication and collaboration between
team members, wherever they may of the PC, the network and the test Connecting the system’s host PC to
be in the world. instrumentation. the company network builds a bridge
to the island. It also opens the door
Figure 10.1. The Blaster worm infected more than 300,000 computers in less than 18 hours
to a wider range of threats—including
some that may compromise system
security and integrity.
Blaster: The first 18 hours
400,000 Inadvertent threats may reach the
system via the company intranet.
350,000
Some are programmatic, as when
300,000 another PC on the network causes a
configuration change in one or more
Number of infections
101
Malicious threats from the Internet partial support) probably can’t create The router-based approach
may breach the company’s firewall, a locked LAN I/O session between
spread via the intranet and infect the the instrument and a PC. A router is a standalone box with
system’s host PC. These threats also multiple LAN connectors, one for
pose a potential risk to any instru- Locking ensures a stable PC-to- the external or “public” network and
ments that contain a Windows PC. instrument connection and also four (or more) for the internal or
One answer is to include a hardware blocks other attempts to access “private” network. The router links
or software firewall in each instru- the instrument for the duration these networks through its ability
ment—a solution Agilent is enabling of a session. Instruments that do to handle high-level communication
in next-generation instruments. support VXI-11 have an important protocols such as TCP/IP. Routers
shortcoming when not locked into a allow one- and two-way communica-
session: they have no authentication tion between devices and also enable
Examining other issues capabilities (e.g., password protec- “awareness” among devices on a
tion) to block unauthorized access. network.
A LAN-based system is also subject
In this case, any PC on the network
to the quirks and limitations of the Routers also utilize a feature called
that supports VXI-11 can access the
deployed hardware. As an example, network address translation (NAT)
instrument and easily disrupt its
the simplest way to connect a system that allows devices to hide their
behavior. The solution is a private
to the corporate network is through presence from public networks. It
LAN that limits access to only those
a hub. However, hubs let all network does this by using a private set of
devices you trust.
traffic flow in both directions; all IP addresses that are not revealed
intranet traffic would be present
within the test system and all test Designing the private, to devices on the public side. This
is the key attribute that enables the
system traffic would appear on the protected LAN creation of a private LAN for a test
intranet. Excess network traffic could system.
degrade system throughput and the Our basic prescription for any
broadcasting of test results on the LAN-based test system is to create As shown in Figure 10.2, the router
intranet could be a security risk. a private, protected network that is the focal point of the network.
Using a switch or router is a better includes the host PC and the test In the simplest router-based
choice because both are specific equipment. Fortunately, there are system, its “external” port, usually
and selective about filtering and two practical, effective ways to set up labeled Internet or WAN (wide area
forwarding network traffic. this type of network. One approach network), is connected to the corpo-
is built around a LAN router, which rate intranet. Its other ports, usually
Some older LAN-enabled instruments provides a buffer between the test labeled LAN, are connected to the
also have two weaknesses that must system and the corporate intranet. host PC and a few LAN-enabled
be addressed or acknowledged: the The other approach uses the host instruments. Additional instruments
inability to lock connections and PC as the buffer by configuring it can be added by connecting a switch
authenticate devices. For example, with two LAN cards and the Internet or hub to one or more LAN ports on
those that don’t support the VXI-11 Connection Sharing (ICS) feature of the router (Figure 10.3).
communication protocol (or provide Windows XP and Vista.
Figure 10.2. A test system that uses a router-based private, protected LAN Figure 10.3. An expanded network that uses a switch to connect
additional instruments to the test system
To corporate
intranet To corporate
intranet
Router The router-based approach has
Router
several advantages. First and
foremost, it protects the test system
Instrument 1 Instrument 2 from the potential hazards carried
Instrument
on 1the intranet or Internet.
Instrument
It also2
prevents any type of outside access
by limiting Switch
communication to only
those devices that reside within the
private LAN— and, unlike a hub, it
Instrument 3 Instrument 4
102 10. Using LAN in Test Systems: Network Configuration and Basic Security
shields the system from intranet
congestion by isolating all but local Defining router and PC • Adequate port speed. The router
should support at least 100 Mbps
traffic. What’s more, the router features (100Base-T) on each LAN port (the
safeguards system operation from
the effects of administrative activity Successful implementation of the router- private side) as well as 100 Mbps on
or hardware problems on the local based private network requires a few the WAN port (the public side). 1000
intranet because it provides all of essential capabilities. Must-have router Mbps (Gigabit or 1000Base-T) routers
features and PCs are widely available and
the network services needed by the
instruments and the host PC. At the recommended for best performance.
• Network address translation (NAT).
same time, the router gives the PC NAT allows the router to act as • Built-in DHCP server. The router
unhindered access to the system an agent between the public and assigns IP addresses to the LAN
network as well as the corporate private networks, mapping private IP devices attached to its private LAN
intranet and the Internet. It also addresses to public IP addresses and ports. Some routers keep a table in
gives all LAN-enabled instruments enabling communication between non-volatile memory that provides
access to TCP/IP resources on the networkss. a mapping between the assigned
intranet and the Internet.
IP addresses and the associated
• De-militarized zone (DMZ) . The Ethernet devices on the private
The configuration process is rela- DMZ feature makes it possible to
tively simple for the router and the network. Vendors call this capability
give a PC (or instrument) complete by many names, including “static
PC—to the extent that you usually access to the Internet, effectively
won’t have to burden corporate DHCP,” “DHCP client reservation,”
putting it “outside the firewall.” “fixed mapping,” and “MAC address
IT personnel with the task. For With DMZ, other computers outside
example, a host PC that’s already to IP mapping.” (MAC stands for
the private network can connect media access control.)
equipped with a LAN card doesn’t to the host PC and use its public
require any hardware modifications. services (e.g., shared file folders The router-based approach also requires
The only PC configuration change, or a Web server). Because DMZ is a host PC that uses TCP/IP rather than
made after the router is installed implemented differently on various NetBEUI, IPX or SPX1 as its network
and enabled, is to activate dynamic router models, you should verify that communication protocol. The PC may use
host configuration protocol (DHCP), you are able to achieve essential DHCP to ensure assignment of a unique
which is a method of automatically IP address, or it can be configured with a
tasks such as communication with a
assigning an IP address to any device static, internal IP address that is compat-
manufacturing database Should-have ible with the router’s configuration.
connected to a LAN. (DHCP may be router features:
turned on by default, but it’s best to
verify this setting.) • Sufficient ports. Each device should
have its own LAN po. Each device 1 NetworkBIOS Extensions User Interface,
Instrument configuration is also should have its own LAN port, either Internetwork Packet eXchange and Sequenced
quite simple. The only changes are Packet eXchange are alternatives to TCP/IP for
in the router or via one or more LAN network communications. If installed in the host
deactivating DHCP then setting the switches or hubs connected to router PC, their presence can create problems within
IP address, subnet mask and default ports. the network.
gateway. These tasks are easy to
complete via the front panel or web
interface of most LAN-enabled instru-
ments. A more detailed description
of the configuration process is
presented in Appendix 10A.
www.agilent.com/find/open
103
The PC-based approach The most important step is the Through the Network Connections
configuration of ICS within the host control panel, both LAN cards can be
By adding a second LAN card and PC, which must be running Windows enabled and the one connected to the
activating ICS in Windows XP or Vista, Windows XP with Service public network can be shared. You
Vista, the host PC can serve as the Pack 1 (SP1), Service Pack 1a (SP1a) then use the Local Area Connection
router in the network (Figure 10.4). or Service Pack 2 (SP2) (Microsoft Properties window to enable ICS
ICS routes traffic from one LAN service packs are cumulative). (Figure 10.5).
card to the other and provides NAT
capabilities for the private addresses.
Other operating system configurations
This method has several advantages may work but this note focuses on the
in common with the router-based most recent versions of Windows.
solution: it provides access control,
blocks Trojans and worms, and gives
Figure 10.4. The PC-based solution, with two LAN cards in the PC and a
the host PC unhindered access to the
switch to connect the instruments
system network, the intranet and the
Internet. LAN-enabled instruments
can also access the intranet and the To corporate
Internet. However, if the host PC is intranet
configured to use DHCP rather than
a static address then it will have to
rely on the corporate intranet being Switch
functional and able to provide an IP
address.
Instrument 1 Instrument 2
Although it probably isn’t a major
obstacle, this approach requires that
you are comfortable with the pros-
pect of opening up the PC, installing Figure 10.5. Use the Local Area Connection properties window in
the second LAN card, and config- Windows XP to activate Internet Connection Sharing.
uring the PC to ensure the peaceful
coexistence of two LAN cards.
104 10. Using LAN in Test Systems: Network Configuration and Basic Security
Instrument configuration It is also necessary to configure the Appendix 10A: Configuring
instruments with the proper subnet
Keep one important caveat in mind mask (usually 255.255.255.0) and the router-based system
when using either approach: the default gateway, which is the IP
Of the two solutions described in
default mode of both ICE and a address of the router itself (typically
this chapter, the router-based system
router is to dynamically assign an 192.168.0.1, 192.168.1.1 or similar,
is more flexible and therefore more
IP address to every device that joins depending on the router maker).
likely to be widely used. The specifics
the network. This is done via DHCP,
Once you’ve saved these settings, of the configuration process depend
which prevents addressing conflicts
you may have to cycle power on each on the actual products used to
but also creates the possibility of
instrument for the changes to take assemble the system. However, three
assigning a different address to
effect. After each instrument has essential steps provide a framework
each test instrument every time it
completed its boot-up operations you for the implementation of any router-
is powered up or reconnected to
can then connect it to a LAN port on based solution: capturing network
the network. As described earlier,
the router. information, configuring the router
unwanted address changes can result
and setting up the test instruments.
in improper operation and damaged
DUTs. Conclusion
Capturing network information
The easiest way to prevent address The decision to use LAN for system
changes is to disable DHCP in each I/O delivers valuable benefits to your You’ll need to record some informa-
instrument and then enter a static company and your team. However, tion about the network and use it
(fixed) IP address. Though this will it also opens the door to malicious to set up the router, which will be
be easy to accomplish via the front threats and inadvertent risks that inserted between the PC and the
panel or web interface of most newer can affect system performance and intranet. That way, the PC is already
LAN-enabled instruments, it may not integrity. The creation of a private programmed with everything you
be possible with some older equip- LAN can protect the test system from need to know about its network
ment. In those cases, the easiest solu- those risks and ensure maximum configuration. You’ll need to record
tion is to use the GPIB interface on throughput. Using the standard that information and use it to set up
the instrument and add a LAN/GPIB networking capabilities of today’s the router.
gateway such as the Agilent E5810A PCs and the low-cost networking
to the network. products now available, you can What you need
chose either a router-based or PC- Configuration requires the host PC,
The IP addresses you assign to the based approach. powered up and connected to the
instruments should only differ from intranet; the router; one LAN cable
the IP address of the router by the Both approaches protect the test for the PC and one LAN cable for
last of the four numbers in the IP system from the potential hazards each instrument.
address (e.g., 192.168.0.x). You may carried on the intranet or Internet,
want to use numbers higher than prevent any type of unauthorized The process
200, reserving the first few digits outside access, and shield the system 1. Power up the router.
for any DHCP-enabled devices for from intranet congestion by isolating
which the router will typically all but local traffic. The router-based 2. Disconnect the intranet LAN
assign an address in that low range. approach has the additional benefits cable from the PC. Use another
Applying these ideas to a system of safeguarding system operation LAN cable to connect the PC’s
that includes a router at the IP from the effects of administrative LAN port to any LAN port on
address 192.168.0.1, the instruments activity or hardware problems on the router. Wait a minute or so
could use numbers in the range of the local intranet because the router to ensure the PC and router are
192.168.0.200 to 192.168.0.255. provides all of the network services synched.
needed by the instruments and the
host PC.
www.agilent.com/find/open
105
3. From the PC’s Start menu, open 1. Enter the PC’s host name. Setting up the instruments
a DOS or Command window and
type in ipconfig/all. This will 2. Enable cloning of the PC’s MAC The final step is to configure the test
display several items including address. instruments with static IP addresses.
“Host Name” and “Physical Use the front panel keys of each
3. Modify the security settings to LAN-enabled instrument to access
Address.” The PC’s host name
disable blocking of anonymous the I/O, Utility or IP Setup configura-
is registered with the corporate
ping requests. (Allowing other tion menu and disable DHCP. Next,
DNS services. The physical
computers to ping the host PC give each instrument a unique IP
address is the unique MAC or
may be a requirement of some address in the range of 192.168.x.200
Ethernet address of the LAN card
corporate intranets.) to 192.168.x.255 (Figure 10.6). These
in the PC. Write down the host
name, the physical address, and 4. Enable the DMZ capability and values are outside the range of IP
the IP address of your computer: set the DMZ host IP address to addresses routers typically assign to
you’ll use that information later 192.168.x.100 (the x must match network devices (192.168.x.100 to
when configuring the router. the value used by your router). 192.168.x.149).
This is the default first address You’ll also need to navigate the
To create a new, private network that
assigned by the router’s DHCP configuration menu and set the
consists of just the PC and the router,
server and must be used as the subnet mask to 255.255.255.0 and
return to the DOS or Command
IP address for the host PC. Some the default gateway to the router’s IP
window and type in ipconfig/renew.
routers may use different initial address (192.168.0.1, 192.168.1.1 or
addresses: type in the ipconfig/ similar, depending on which brand of
Configuring the router all command to find out what router you’re using).
address the router assigned the
The router must be configured to PC after they were connected. Once you’ve saved these settings,
mimic the test system PC on the you’ll have to cycle power on each
corporate intranet. Most routers 5. Save all of these settings. Locate instrument for the changes to take
provide a browser-based interface the intranet cable that was effect. After each instrument has
that lets you use any Web browser to originally connected to the PC completed its boot-up operations, use
log in and modify the configuration. and plug it into the router’s a LAN cable to connect each one to a
Consult the router’s manual for its WAN or Internet port. To verify LAN port on the router.
URL and the default login values for proper operation, open a DOS
user name and password. Launch or Command window and type To verify proper configuration, open
your Web browser, type in the proper ipconfig/release. Next, type a DOS or Command window and
URL and log in to the router’s ipconfig/renew: the host PC type ping 192.168.1.200 or any other
configuration page. At this point, the should now be able to access the valid IP address you assigned to an
details vary by vendor and product. corporate intranet via the router. instrument. To verify access to the
There might be a built-in wizard intranet, launch a Web browser and
function, or you may have to navi- try a few internal URLs. If these
gate through various configuration load as expected, this verifies proper
screens and enter values manually. communication with the intranet.
Either way, you need to accomplish
five tasks: Figure 10.6. The IP Setup menu of the Agilent 33220A function/arbitrary
waveform generator makes it easy to set the IP address, subnet mask and
default gateway.
106 10. Using LAN in Test Systems: Network Configuration and Basic Security
11. Using LAN in Test Systems: PC Configuration
Introduction On the surface, the use of LAN to manufacturer and model. Right
communicate with instruments clicking on a Local Area Connection
This chapter builds on the informa- seems like it should be as simple entry will bring up a menu that
tion presented in Chapters 9 and 10, as connecting a printer to a PC: includes a Properties selection that
describing the additional capabilities just grab a network cable and make opens the Local Area Connection
required to enable communication the connection. Unfortunately, it Properties window (Figure 11.1). This
between a PC and LAN-enabled requires a bit more effort to create window includes a list of check-
instrumentation. the right environment within a PC box items, and “Internet Protocol
for transparent communication with (TCP/IP)” is the last item in the
Creating the right LAN-equipped instruments. Making list. Clicking on that item (not the
it work depends on the LAN services check box) and then clicking the
environment of Windows XP and Vista, and the “Properties” button will bring up the
additional capabilities provided by TCP/IP configuration window. This
The evolution of LAN technology
a suite of I/O libraries from Agilent is where you can change IP address
continues to drive improvements in
make it nearly “printer easy.” settings (static or dynamic) and other
cost, speed, functionality and ease
networking parameters.
of use. This has created at least
two noteworthy trends. One is the Exploring network settings
widespread use of LANs within most in Windows XP and Vista Figure 11.1. Use the Local Area Connection
businesses. The second is the inclu- Properties window to set networking parameters
sion of LAN as a standard feature A standard Windows XP or Vista such as the IP address.
of most new PCs. A third trend is installation includes three software
emerging in test-system development: components that enable networking:
the advantages of LAN technology Client for Microsoft Networks, File
are making it an attractive alterna- and Printer Sharing for Microsoft
tive to GPIB for system input/output Networks, and Internet Protocol
(I/O). (TCP/IP). TCP/IP is the network
protocol that enables data communi-
Chapter 16 offers a closer look at cation with the Internet, your corpo-
LXI, the new LAN-based standard for rate intranet and other computers.
computer-instrument communica-
tion. Backed by the LXI Consortium, You access the PC’s TCP/IP settings
which includes every major test and through the Windows Control Panel
measurement company, the LXI stan- and its Network Connections icon (or
dard ensures interoperability among Network and Internet Connections
vendors’ LAN-based instruments and category). The Network Connections
simplifies configuration and program- window will present every installed
ming of LAN-based systems. Given LAN card under the default name
the advantages of LAN and LXI, “Local Area Connection” (you can
LAN interfaces are becoming more change this name). Each entry will
common in test equipment—though indicate the device’s connection
LAN ports will likely coexist with status (enabled or disabled) and
GPIB for years to come.
107
Using multiple network Of course, the “dynamic” part of
DHCP means that a device may Running a network without
connections receive a different IP address if it is DHCP
disconnected and later reconnected
It is possible to install multiple
to the network. This can cause prob- If a device running TCP/IP is set to
network cards in a Windows XP or
lems in a LAN-based test system. For obtain its IP address automatically
Vista machine. One reason to use
example, if the system contains two but there is no DHCP server available,
two LAN interfaces in one PC is
power supplies and DHCP reverses the device will use a capability called
described in Chapter 10: the PC can
their IP addresses, the device under Automatic Private IP Addressing
do double duty as the host computer
test (DUT) could receive the wrong (APIPA or “auto IP”) to assign itself
for a LAN-based test system and as
voltages at the wrong points and an address about two minutes after
a router that links a private network
suffer severe damage. There are two boot-up. This feature is built into
(for the test system) to the corpo-
alternatives. One is to assign perma- Windows PCs and most Agilent
rate intranet. The Interconnection
nent (static) IP addresses to every instruments. Auto IP creates
Connection Sharing (ICS) feature
device on the network. This lets you addresses that are designed to be
of Windows XP and Vista, accessed
fine tune the network and its settings; compatible with each other, enabling
through the Local Area Connection
it also isolates the network from any the establishment of a network
Properties window, makes this
failure of the corporate DHCP server. without DHCP or the configuration
configuration possible.
On the downside, this approach can of static IP addresses. Also, most of
become overwhelming in a large Agilent’s LAN-enabled instruments
Managing IP addresses network, increasing the chances of use the NetBIOS (Network Basic
configuration errors or bad settings Input/ Output System) protocol. In
In the dual LAN card configura-
that could cause problems across the a network configuration that lacks
tion, the PC acts as the network
network. a DHCP server, the PC can connect
controller for the private network,
to the instrument by using the host
enabling access to the public network The other alternative is to use name that is configured from the
and providing network address a dynamic domain name server instrument front panel. This makes
translation (NAT) and dynamic (dynamic DNS or DDNS) on the local it easy to create a direct connection
host configuration protocol (DHCP) network. DDNS lets an instrument, from a PC to a single instrument, and
services to connected devices. NAT PC or other network device establish all it takes is a LAN crossover cable.
is the key capability that enables the a specific host name when it connects
private, protected LAN by shielding to the network. (The host name can
private IP addresses from the public typically be entered via the front
network. panel of a LAN-enabled instru-
ment.) Large corporate intranets
DHCP automatically assigns an IP
usually have such a server, which
address to any device connected to
allows other devices to use the host
the network and also ensures that
name with DNS to find the device’s
no two devices receive the same IP
IP address and connect to it. If an
address. This is the default setting in
instrument’s IP address changes,
Windows XP and most LAN routers.
DDNS ensures a quick update of
It’s very simple and it works well 99
the DNS table of addresses. This
percent of the time—as long as the
approach has one major caveat:
default DHCP server on the network
DDNS will not typically be avail-
is up and running.
able on the small, secure networks
we recommend for test-system
applications.
www.agilent.com/find/open
109
Enabling communication with Conclusion
instruments
With LAN ports in most current
With connectivity completed and generation PCs and many new-gener-
verified, the next step is to ensure ation test instruments, connecting
and enable communication between the two is almost as simple as plug-
the PC and the instruments. As ging in a network cable—if you apply
one example, the VISA I/O API the capabilities and tools that are
can use two different methods to part of the Agilent IO Libraries Suite.
communicate with LAN devices: the
VXI-11 communication protocol and To discover more ways to simplify
raw TCP/IP socket communication. system integration, accelerate system
VXI-11 is the preferred choice when development and apply the advan-
moving existing (GPIB-based) code or tages of open connectivity, please
when you want to keep a consistent visit www.agilent.com/find/open.
programming style with GPIB instru-
ments. Sockets should be used when
the host PC does not support VXI-11.
Sockets also provide the highest level
of performance, but at the cost of
some programming complexity.
111
USB speeds Agilent support for USB The 82357B is a Hi-Speed, fully
compliant, hot pluggable USB device,
The USB 2.0 Specification encom- instrument connectivity so you can connect it whenever
passes all USB data transfer speeds: you need it, without rebooting your
To provide users with the utmost in
Hi-Speed (480 Mb/s), full-speed (12 PC. Instruments connected via the
convenience, Agilent has committed
Mb/s) and low-speed (a 1.5 Mb/s 82357B have GPIB-style VISA and
to providing USB connectivity as a
alternative designed for keyboards, SICL addresses, just like Agilent’s
standard feature in nearly every new
mice, and other low data-rate PCI- or older ISA-based GPIB cards,
test and measurement instrument. In
devices). Just because a device is so legacy programs in systems that
most cases, new instruments support
USB 2.0 compatible doesn’t auto- use these cards require no reconfigu-
480 Mb/s Hi-Speed USB, which
matically mean it can operate at ration or code changes.
delivers greater bandwidth and lower
480 Mb/s. The best way to verify the
latency (the time required to respond You’re not limited to locally avail-
speed of USB devices is to look for
to programming commands) than able instruments, either. With the
the official USB logo. Devices that
GPIB. Those few instruments that Agilent E5813A networked 5-port
are certified to run at one of the two
support full-speed USB (12 Mb/s) USB hub, you can access remote USB
original USB rates, 1.5 Mb/s or 12
offer bandwidth similar to GPIB with devices or instruments through your
Mb/s, should carry a white and blue
somewhat higher latency. standard LAN. With the E5813A
Certified USB logo (Figure 12.1).
Devices certified to run at 480 Mb/s connected to your PC and properly
You can also take advantage of USB
rates carry the red, white, and blue configured, those remote instruments
with your existing GPIB instruments.
Certified Hi-Speed USB logo. and devices will function as though
The Agilent 82357B USB/GPIB
they were locally attached.
Figure 12.1 USB logos identify the device’s
Interface (Figure 12.2) connects
speed rating GPIB instruments to a USB port on Agilent also provides a USB solution
your computer, giving you a way to for RS-232 instruments. The Agilent
control up to 14 GPIB instruments E5805A USB/4-port RS232 interface
from a laptop or other PC for each provides a direct connection from
82357B. the USB port on your notebook or
desktop PC to up to four RS232
instruments or devices.
Figure 12.2. Agilent 82357B USB/GPIB Interface
To simplify programming of instru-
ments over USB connections, Agilent
and other test equipment vendors
co-developed the industry standard
USB Test and Measurement Class
(USBTMC) and USB488 I/O proto-
The speed rating of hubs in a USB cols. These protocols, along with
system determines the operating Agilent’s IO Libraries Suite, allow
speed of the system. For instance, if you to easily switch from GPIB to
you connect Hi-Speed USB devices USB connections without making big
through a full-speed hub, the investments in new PC software or
maximum speed you can expect from rewriting existing programs. Aside
any of the devices is 12 Mb/s, not 480 from address conventions, your USB
Mb/s. To take advantage of Hi-Speed instruments will look and act just as
data rates, you must connect these they would under GPIB control.
devices through a Hi-Speed hub.
www.agilent.com/find/open
113
Communicating with tion, which discourages lower-level, Similarly, the 82357B USB/GPIB
USB-connected devices error-prone interface manipulation interface looks and acts like a PCI/
operations. Consult the documen- GPIB adapter for typical instrument
You don’t need to worry about the tation for the instrument or I/O communication, so instruments
details of a USB connection, so most adapter for any limitations. connected to it have GPIB-style
programs written to talk to GPIB address names and act like any other
As mentioned earlier, an instrument GPIB-connected instruments. The
devices work with USB-connected
connected via a USB cable acts like source code is exactly the same as
devices without modification.
an instrument connected over a above, with the exception that the
However, if your program uses
GPIB bus, aside from a different I/O instrument address would be some-
low-level commands that affect the
address. Here’s some example C code thing like “GPIB0::23::INSTR” instead
entire GPIB bus (such as through a
that communicates with a USB- of “FuncGen.”
VISA session such as GPIB::INTFC),
connected instrument, either natively
you may encounter some unexpected
or with the E5813A networked 5-port
results. USB devices are optimized
USB hub:
for modern instrument communica-
#include <iostream>
#include <tchar.h>
#include <stdio.h>
#include “visa.h”
/* Error-checking routine */
void CHECKERROR(ViSession vi, ViStatus status)
{
char desc[256];
ViStatus err = 0;
if (status < 0)
{
err = viStatusDesc(vi, status, desc);
fprintf(stderr, desc);
viClose(vi);
_exit(status);
}
}
www.agilent.com/find/open
115
Conclusion
The low cost and simplicity of
USB—and the dramatic speed
improvements of Hi-Speed USB
make—make USB an ideal connec-
tivity option for simple, ad hoc test
systems. Achieving expected data
rates does require some attention to
the hubs used in a USB system; the
system will operate only as fast as
the fastest hub, so make you’re your
hubs are also Hi-Speed rated. Agilent
offers comprehensive support for
USB-based test systems, including
built-in USB ports in many instru-
ments and support for USB in the
Agilent IO Libraries.
117
Enabling connectivity Standardizing the API API; it exposes the VISA API to the
application layer through use of the
In the early days of automated To make it easier to create mixed- Component Object Model.
testing, system controllers—called vendor test systems, a group of
desktop calculators or instrument instrument vendors created the The result: VISA COM gives you
controllers— had limited processing Virtual Instrument Software the freedom to pick from the most
and sparse memory. To keep the Architecture (VISA). This provided popular I/O configurations and
syntax as simple as possible, equip- a standardized API that allowed also choose from a wealth of “COM
ment vendors used short commands, control of instruments through a friendly” languages such as C#,
initially in binary and later in ASCII. common interface—directly or with Visual Basic 6 and Visual Basic .NET.
drivers. From the application’s point As we’ll discuss later, the application
Different manufacturers defined of view, every vendor’s VISA inter- development environment (ADE)
their own command strings and these face looks the same. you choose will influence the best
were typically unique to the specific choice of library and API for your
capabilities of each instrument. In One important caveat goes with application.
a system, replacing an instrument VISA: Although the VISA API is
with one from another manufacturer, standard, each vendor employs Achieving communication
or even a new-generation product different layers beneath the VISA
from the original maker, could mean layer to control the hardware. In Once you’ve enabled connectivity,
completely rewriting the system addition, each vendor may have it’s time to decide how to achieve
software. made enhancements to enable unique communication between the host
features in its application layers. To computer and the system instru-
Instrument commands aren’t the make it all work, the version of VISA mentation. The two alternatives are
whole story, however. Enabling installed on the host computer must direct I/O and instrument drivers.
connectivity between a controller be compatible with the I/O hardware. Direct I/O creates an explicit connec-
and the system instruments requires (In contrast, this points to another tion to each instrument, which
additional layers of software. advantage of PC-standard I/O such makes it faster but limits instrument
Historically, the I/O software layer as LAN and USB: any version of VISA interchange and software reuse. Most
contained libraries such as the that supports those interfaces will instrument drivers utilize direct I/O
Standard Instrument Control Library work because the low-level drivers and SCPI but sometimes hide that
(SICL) or NI-488. The application are standardized.) connection. Generally speaking,
used these libraries to achieve direct drivers trade decreased flexibility
communication with an instrument. (and possibly speed) for improved
Each vendor had a proprietary appli- Expanding freedom of choice interchange and reusability. However,
cation programming interface (API) in most situations you can use
As I/O development was proceeding
that communicated exclusively with both instrument drivers and direct
in the test and measurement
its own I/O interfaces. This made I/O to achieve the best balance of
industry, the PC industry was
it difficult for system developers speed, flexibility and measurement
pursuing independence in both
who were building mixed-vendor functionality.
I/O and programming languages.
test systems—and, of course, many
Microsoft created the Component
systems used (and continue to use)
equipment from multiple vendors.
Object Model (COM), which is a Standardizing direct I/O
software architecture that allows
components made by different An early attempt at improving
For more about the evolution of software vendors to be combined consistency and ease of use came
instrument control, see Chapter 3, into a variety of applications. COM in 1989 when Hewlett-Packard
Understanding Drivers and Direct I/O. is not dependent on any particular introduced an instrument commu-
programming language. nication language called the Test
& Measurement Systems Language
To incorporate the advantages of (TMSL). HP and eight other manu-
language independence, Agilent
initiated the creation of VISA
COM as a companion to the VISA
standard. VISA COM is an object- HP spun off its test & measure-
oriented representation of the VISA ment businesses as part of Agilent
Technologies in 1999.
Figure 13.2. This block of Visual Basic 6 code uses SCPI and VISA COM I/O to communicate with a function generator.
Next I
End With
www.agilent.com/find/open
119
• IVI-COM. This standard does the Exploring the application mated) that are typically performed
most to enable interchangeability during the early phases of product
and reuse by leveraging the COM alternatives development or during design
computer standard. IVI-COM verification.
Shrink-wrapped software often
drivers integrate with standard PC
provides convenience in measure- Applications such as the free Agilent
component architecture software
ment and analysis at the expense of IntuiLink connectivity software and
and enable control of instruments
performance and flexibility. Such low-cost Agilent BenchLink make
from familiar, conventional ADEs
products are often a good fit with the it easy to perform semi-automated
that provide major productivity
small or one-off systems used during measurements, collect data and
improvements. IVI-COM drivers
product development. In contrast, analyze results from a wide variety
that control VXI or GPIB instru-
custom-built software is often the of instruments. Both applications
ments use VISA (either VISA COM
best answer for applications such as utilize either drivers or direct
or VISA-C). Because many new
design verification or manufacturing I/O—transparently—to enable instru-
instruments include computer-
test that require high performance ment communication, control and
standard I/O such as LAN and
and maximum flexibility. data transfer.
USB, IVI-COM drivers for non-GPIB
instruments are not required to
• IntuiLink. This connectivity applica-
use VISA, although many do. Simplifying basic analysis tasks tion simplifies data transfers by
If you are unsure of which I/O There are alternatives to general- adding a toolbar to popular PC
technology an application or driver purpose development environ- applications such as Microsoft
is using, take a look at the connec- ments. One example is “targeted Word and Excel. IntuiLink enables
tion string or “instrument address” applications,” which address specific direct retrieval of data and images
used for instrument communication. measurement or technology domains, from a measurement instrument,
VISA-type strings look like “TCPIP: or specific phases or tasks in the letting you remain in the PC
34980A.tm.agilent.com::inst0::INSTR” product development lifecycle. application and use its familiar
while SICL-based strings are similar to These applications include software interface. IntuiLink also eliminates
“lan[34980A.tm.agilent.com]:inst0.” designed to make the infrequent barriers between instruments and
measurements (manual or semi-auto- PCs by supporting GPIB, USB, LAN
and FireWire interfaces.
Figure 13.3. Agilent BenchLink Data Logger provides spreadsheet-like test set up and real-time • BenchLink. This application is
display and analysis of measurements. available in versions that support
numerous instruments. BenchLink
is a Windows-based application
(Figure 13.3) that uses a familiar
spreadsheet format to streamline
data collection, presentation and
analysis. It can communicate
with measurement instruments
via LAN, USB or GPIB using the
included I/O software.
www.agilent.com/find/open
121
Visual Studio with Agilent T&M • Instrument Explorer. This tool makes Agilent VEE Pro
Toolkit it easy to see and edit the instru-
For those who want an alternative to
ment I/O configuration and initiate
The Agilent T&M Toolkit 2.0 with test communication with instruments. textual programming, Agilent VEE Pro
automation extends the .NET-enabled is a powerful, easy-to-use graphical
• IO Monitor. This utility makes programming environment that accel-
versions of Visual Studio with a suite
it much easier to use instru- erates the process of building and
of integrated, easy-to-use software
ment-control software and programming test systems (Figure
tools and components—project
instrument drivers—IVI-COM and 13.5). To create a program, you
wizards, APIs, class libraries,
VXIplug&play—and diagnose prob- choose high-level graphical objects
widgets, graphs, drivers and more.
lems by letting you watch both the from a huge library and connect
This creates an environment that
underlying direct I/O commands them with lines or “wires.” The wire
simplifies the process of incorpo-
that are sent to the instrument and connections specify functionality
rating tests and measurements
the resulting data that is returned and sequences within intuitive block
into custom applications. Using
(Figure 13.4). diagrams. Because VEE Pro is an open,
T&M Toolkit 2.0 within the Visual
Studio environment lets you use standards-friendly environment, it
In all, the combination of Visual also offers several advantages in test
your preferred textual programming Studio and T&M Toolkit eliminates
language and integrate your new system development:
many of the difficulties often associ-
code with existing code written in ated with connecting to and control-
other languages. ling test equipment from a custom
application.
T&M Toolkit 2.0 offers several other
capabilities that speed and simplify
system development: Figure 13.4. T&M Toolkit’s IO Monitor traces I/O layers for Agilent’s VISA, VISA COM, SICL and SICL
Detail, helping you find bottlenecks in your source code.
• DirectIO class. This is the easiest
way to send commands directly to
an instrument.
www.agilent.com/find/open
123
ADE vs. instrument driver standard. These represent the Conclusion
past, present and future of driver
As mentioned earlier, three types of technology. Open standards such as COM and
standardized instrument drivers are LAN have achieved widespread adop-
available: VXIplug&play, IVI-C and To accelerate test-system develop- tion in the computer world and are
IVI-COM. These work with multiple ment, we recommend using the latest now shaping the future of test-system
ADEs and enable communication IVI-COM drivers and VXIplug&play development. Standards accelerate
with an instrument through any WIN32 drivers for instrument system development by maximizing
vendor’s I/O hardware. control. The IVI-COM driver tech- software reuse and enhance system
nology is the only one built on a flexibility by making it easier to swap
Reading from left to right, Table 13.2 PC-standard architecture. A compo- out instruments—different models
shows a continuum that ranges from nent driver built on COM works in all and even different brands. Standards
least to most standardized across popular PC languages and most T&M also enhance system functionality
three generations of drivers—propri- languages. What’s more, it utilizes the and performance by letting you
etary, T&M standard and PC-industry most popular types of I/O and can be utilize direct I/O, SCPI and drivers
used in the latest .NET technologies. within a single application.
125
Figure 14.1 shows the physical layout 4. Configure your PC. For Windows By default, the VPN client PC can
of such a setup, and here are the XP and Vista, use the “Create access only the virtual network
basic configuration steps: New Connection” task in the task behind the VPN router when the
pane of the Network Connections VPN connection is active. However,
1. Physically connect the router utility that is accessible from the it is possible to configure Windows
and instruments as illustrated in Windows Control Panel. When XP/Vista so that both networks can
Figure 14.1. prompted by the wizard, use the be accessed at the same time, with
public IP address or host name of Windows deciding which connection
2. Via the web interface or other
the router. to use based on the IP address of the
means, configure the VPN router’s
remote device. VPN routers typically
basic DHCP (Dynamic Host 5. After creating the connection, use non-routable, private network
Configuration Protocol) and right-click the new VPN connec- addresses in the range 192.168.x.x
other network settings to create tion and configure the VPN type, or 10.0.x.x for the private networks
a simple, private network on tunnel name and password/key they create. If the local intranet also
which the instruments (and PC you configured on the router. uses private addresses, care must
clients connected via VPN) can
be taken to configure the subnet of
communicate. See Chapter 10 for A VPN router configuration for
the VPN router’s private network
an example configuration. instrument sharing protects the
so that it doesn’t conflict with the
instruments on the private side of
3. Configure the router to have a intranet, otherwise the PC client
the router from public intranet/
VPN end-point to give roaming won’t be able to route traffic to the
Internet access but gives PCs config-
users a connection point. proper network interface (either the
ured with the VPN’s parameters and
Windows XP and Vista provide real network interface or the virtual
passwords unlimited access to those
a basic VPN client that supports network interface created by the VPN
instruments. (Note that for the dura-
the L2TP/IPsec and the PPTP connection).
tion of the VPN session, the external
VPN protocols, so pick a VPN client PC has two IP addresses—one
configuration using one of those for the VPN connection and one
protocols unless you have a for the standard intranet/Internet
different preferred VPN client. connection.)
(See “Configuring a VPN” for
more on choosing a protocol.)
Router
with VPN
endpoint
Sharable LAN
instruments
PCs using
the protected,
sharable VPN tunnel
LAN instruments
Public Private
Ethernet Ethernet
port port(s)
Figure 14.2. Recommended network design for remote monitoring and data acquisition applications
Remote
router
with VPN
endpoint
Public
Ethernet
port
Private Remote data DUT or data
Data acquistion VPN tunnel Ethernet acquisition acquistion
system controller port(s) instruments point
Intranet
Test (wired or wireless)
engineer and/or Internet
www.agilent.com/find/open
127
Remote monitoring and acqui- gain antennae are available for the the system controller computer
sition over wired connections common 802.11b/g wireless Ethernet local to the instruments and point
standards that can, when properly of measurement so that information
If the point of measurement is paired and aligned, create a long and control is not lost for periods
physically located near or in the distance wireless Ethernet bridge of time, rather than keeping the
local corporate intranet, a wired between two points, connecting two controller local to the test engineer
LAN connection is the best choice. In Ethernet networks as though they for easier maintenance.
most situations, using a VPN router were local (see Figure 14.3).
to provide security is desirable to Choosing a wireless technology
It’s important to note that long Unfortunately, choosing a WLAN
prevent unauthorized access to the
distance WLAN installations require technology is not a simple matter,
measurement equipment and to
specialized knowledge and equip- since there is a confusing variety of
prevent infection by viruses. See
ment, and they are highly dependent WLAN standards, both implemented
Figure 14.1 for the recommended
on terrain and other environmental and under development. Although all
VPN configuration for such a data
factors. Due to FCC restrictions on of these stem from the IEEE 802.11
acquisition system.
unlicensed equipment in the 2.4 base standard, you can see from
GHz radio band in the United States, Table 14.1 that the various single-
Remote monitoring and the signal cannot be amplified to letter suffixes represent a variety of
acquisition over wireless achieve greater transmission range, technologies and protocols within the
but effective range can be increased
connections by using a pair of highly directional
802.11 framework.
Figure 14.3. Establishing a remote data acquisition connection via wireless LAN
www.agilent.com/find/open
129
However, WPA may not be avail- Scenario 3: Functional by letting you assign a friendly
able in all possible configuration name to each instrument or
modes for wireless access points. test systems resource in the form of an alias.
For instance, in tests at Agilent, For instance, you can assign your
Functional test systems represent a
we were unable to use WPA with a function generator any name you
third potential for LAN connectivity.
D-Link DWL-2100AP access point in like even though it resides at a
In addition to the benefits discussed
wireless bridging mode, where two specific IP address.
in earlier chapters, several points
access points seamlessly bridge the
need to be considered when applying
two networks they connect into one,
LAN technologies to functional test
larger network. Only WEP security
systems: security and independence
was available in this mode, and we
from network infrastructure, timing
have found no commercial products
and deployment.
that claim to implement WPA for
bridging mode. If this remains true • Chapter 10 discusses security
for 802.11i, we recommend either and independence from network
not using bridging mode for bridging infrastructure through the use
from your intranet or putting VPN of static network configurations
routers on either side of the wireless and inexpensive never systems as
access points to use secure IPSec required, since the GPIB primary
communications (see “Peer-to-peer address is relative to the computer
IPSec tunneling/bridging”) to guar- and the GPIB adapter (as opposed
antee wireless security will not be to being a globally unique number).
broken (see Figure 14.4 for a sketch The Agilent IO Libraries Suite
of such a configuration). brings back some of this simplicity
Figure 14.4. Wirelessly bridged Ethernet network with VPN tunnel for additional security
VPN VPN
router router
Wireless Ethernet
Ports for radio signals Ports for
network Wired Wired network
#1 Ethernet Ethernet #2
connection connection
802.11b/g (high gain)
antennae
IPSec tunnel
www.agilent.com/find/open
131
Peer-to-peer IPSec There are a few situations where configuration tools are too complex
tunneling/bridging IPSec tunneling is the preferred to use without instructions or
choice. Because of the always-on experienced help. However, the
IPSec provides its own tunneling configuration of this tunnel, IPSec Windows implementation is very
mode where two networks are tunneling is a good choice for virtu- flexible and powerful, allowing traffic
virtually joined by establishing a ally connecting a set of measurement destined for the private network
secure tunnel between the network hardware and a controller/monitor, behind the remote IPSec endpoint to
endpoints on a larger network, such as in a test system with a be automatically encrypted and sent
such as a corporate intranet or the remote controller or a remote data over the tunnel and all other TCP/IP
Internet. The networks that the acquisition application (see Figure traffic to continue to its destination
tunnel endpoints connect can be 14.5). Because the endpoints of the unimpeded.
anything from a single computer to a tunnel can be networks of devices,
IPSec tunneling is a good choice for Many VPN routers also have IPSec
large corporate LAN. This tunneling
connecting two separate test systems tunneling support, with varying
mode is designed for permanent
or permanently connecting a test degrees of configuration help. VPN
network configurations between two
system controller. routers without a hardware encryp-
points with known IP addresses or
tion co-processor can be an order
host names (making it peer-to-peer,
IPSec tunnels may not work across of magnitude slower than the most
rather than a client/server archi-
corporate firewalls, so the tunnel powerful, more expensive routers
tecture). An example application of
end-point hardware may have to be with an encryption co-processor
IPSec tunneling would be to virtually
exposed to the Internet to allow such built-in. Configuration informa-
connect two campuses of an organi-
tunnels to be connected over the tion and contributed utilities for
zation via an IPSec tunnel over the
Internet. configuring an IPSec tunnel between
Internet so that the two ends of the
a Windows 2000/XP client and a
tunnel are combined into one large, As part of its IPSec feature set, D-Link DI-804HV router are also
secure virtual LAN. Windows provides the capability available at www.agilent.com/find/
for IPSec tunneling, although their adn_vpn_examples.
IPSec endpoints
(routers, computers, etc.)
Private Private
network #1 network #2
VPN tunnel
Public Public
Ethernet Ethernet
port port
0 or more Corporate intranet
and/or Internet 0 or more
private private
Ethernet Ethernet
port(s) port(s)
1400
1200
1000
800
KB/s
600
400
200
0
0 10,000 20,000 30,000 40,000 50,000 60,000 70,000
Waveform size in bytes
www.agilent.com/find/open
133
Conclusion
LAN is a powerful, compelling choice
for many test and measurement
tasks and systems, but engineers
need to be aware of its limitations
and complexities to create reliable,
long-lasting configurations. The
huge variety of LAN media, devices,
protocols and technologies mean
that a large body of complete tools
and solutions is available to choose
from when designing test systems.
Picking the right technologies to use
and deploy is essential to developing
the best system in the least amount
of time.
instrument connections that perform test and measurement Agilent VISA COM
(T&M) operations. The utilities This is a Microsoft Component Object
The advantages of LAN technology enable quick, easy connections of Model (COM) implementation of the
are making it an attractive alterna- instruments to a PC by helping you VISA standard; it is based on the
tive to GPIB for system input/output debug test programs and diagnose Agilent VISA architecture. Agilent
(I/O). As a result, LAN interfaces problems in a test system. VISA COM also conforms to IVI
are becoming more common in test Foundation standards.
equipment—though LAN ports will The IO Libraries Suite is included
likely coexist with GPIB for years to with more than 150 Agilent instru- Agilent SICL
come. ments as well as the Agilent VEE Many test systems still rely on the
and Agilent T&M Toolkit software Standard Instrument Control Library
On the surface, the presence of LAN products. The IO Libraries Suite also (SICL), which Hewlett-Packard (now
ports in most current-generation PCs works with instruments and software Agilent) developed to make software
and many new-generation test instru- from other vendors. as I/O-independent as possible.
ments may make connections seem This modular library for instrument
as simple as finding a network cable As an introduction to the IO
communication works with a variety
and plugging it into both devices. Libraries Suite, let’s take a closer
of computer architectures, I/O
Today, making the connection work look at its three major components:
interfaces and operating systems. We
depends on the LAN services of Agilent IO Libraries, Agilent
include it in the IO Libraries Suite to
Microsoft Windows and the addi- Connection Expert and the I/O
enable compatibility with customer’s
tional capabilities provided by the utilities.
legacy programs.
Agilent IO Libraries Suite. A quick,
one-time configuration process will Agilent IO Libraries In most cases, we now recommend
make LAN-based instrument connec- VISA over SICL. The exceptions are
tions as easy as using GPIB. The suite includes three separate applications that require capabili-
I/O libraries. Each provides similar ties such as parallel polling that are
Once the IO Libraries Suite is functionality that lets you program- unique to SICL.
installed and configured, it acceler- matically control instruments, send
ates the connection process with commands to them and receive Suggested approach
software libraries and utilities that responses and data. If you are developing new test and
let you quickly connect instruments measurement applications, we
to a PC, configure and verify the Agilent VISA generally recommend VISA and VISA
connections, and get on with your Agilent’s implementation of the COM as the most effective solutions
job—whether it entails the creation Virtual Instrument Software for instrument I/O. The best choice
of instrument-control software or Architecture (VISA) is an industry- of I/O library depends on your
the use of pre-existing application standard I/O application program- preferred programming language,
software. ming interface (API). You can use refer to Table 13.1 on page 124 for
it to develop I/O applications and recommendations.
instrument drivers that will be
interoperable with applications from
135
Agilent Connection Expert • Remote IO Server. Enables connec- • VISA Assistant. Included with older
tions to instruments that are versions of the IO Libraries and
The Agilent Connection Expert physically connected to another PC provided in the latest version
(Figure 15.1) is a software utility on the network. When the Remote as a convenience. Most of its
that helps you connect instruments IO Server is running on one PC capabilities are replaced by either
to a PC—via GPIB, LAN, USB, RS-232 (the server) you can use instru- Connection Expert or Interactive IO.
or the VXI interface—in just a few ments connected to that server
minutes. from separate client PCs by using • VXI Resource Manager. Used to
Connection Expert to configure configure the Agilent E8491 IEEE-
You can use Connection Expert to remote interfaces on the clients. 1394 PC link to VXI interface in
speed and simplify several essen- systems that include modular VXI
tial configuration tasks: configure • ViFind32 Debug Utility. When called hardware.
instrument I/O interfaces; connect from a script, viFind32 uses
to instruments over the LAN; VISA functions to find available • IO Control. Provides easy access to
automatically discover instruments resources and then list them in all of the IO Libraries Suite utilities
connected directly to the PC; and a console window. This utility is and the associated documentation.
browse the test system’s structure useful for verifying that Connection The IO Control icon appears in the
and connections (PC, interfaces and Expert has configured all expected Windows notification area (Figure
instruments). Connection Expert also interfaces, and that the expected 15.2), enabling a quick launch
helps you detect and troubleshoot devices are all attached. whenever you want to use the
connectivity problems, either during utilities.
the configuration process or later
when the system is in use. Figure 15.1. Agilent Connection Expert’s easy to follow tree view helps you see connected devices
instantly
To enhance program portability and
readability, Connection Expert also
lets you create programming aliases
for each instrument in the system.
The ability to update an alias with
a new IP address or hostname can
make it easier to handle system
migration and changing network
settings.
I/O Utilities
A set of six software utilities
enhances your ability to quickly
configure and debug instrument-to-
PC connections:
www.agilent.com/find/open
137
Step 3: Gather network private LAN will support DHCP but Step 4: Connect your
information not Dynamic Domain Name Server
instruments
(Dynamic DNS), a service that maps
With a private LAN configuration specific names to IP addresses and This step assumes that the PC’s
you, as the network designer, are enables use of names in place of IP hardware and software are properly
effectively the system administrator addresses in test-system programs. configured for LAN operation and the
and can define network parameters PC is connected to the private LAN.
The key information to be gathered
that best suit your test system.
or defined is summarized in Table As mentioned above, this configura-
One of the key details is support for 15.1. You’ll want to make a copy of tion uses either a router or a PC
Dynamic Host Configuration Protocol the card for each instrument that configured with two LAN cards and
(DHCP), which automatically assigns will be connected to the test system’s a LAN device such as a hub, switch
IP addresses to devices on the private LAN. or router. Before connecting any
network. In general, a router-based instruments, turn off the power to
all instruments. Next, connect each
instrument to the LAN device with a
Most PCs and most Agilent instru- Some of Agilent’s LAN-enabled
standard CAT5e Ethernet cable then
ments will use a function called instruments may be addressable turn on power to each instrument.
auto-IP to automatically assign an by hostname via NetBIOS, an older Verify the completion of the power-on
IP address if they are configured for networked PC protocol that is still sequence for each instrument and, if
DHCP but cannot connect to a DHCP supported by Windows and works anything seems unusual, refer to the
server. with most LANs. instrument’s user’s guide for detailed
information.
Table 15.1. Private LAN information card
Network DHCP enabled Yes ____ No ____ Hostname versus
Dynamic DNS enabled Yes ____ No ____ IP address
UPnP enabled OK Yes ____ No ____
Subnet mask ______.______.______.______ We recommend that you configure
DNS server IP address ______.______.______.______ LAN instruments with a hostname
and use it to connect to them
PC Ethernet (MAC) hardware address ____ ____ ____ ____ ____ ____
whenever possible. IP addresses
IP address ______.______.______.______ assigned by DHCP can change
Subnet mask ______.______.______.______ without warning, breaking estab-
DNS server ______.______.______.______ lished connections to your instru-
Hostname __________________________ ments. If your network doesn’t
Domain name __________________________ support connecting by hostname, we
Instrument Instrument serial number __________________________
recommend the use of a statically
configured IP address for each instru-
Ethernet (MAC) hardware address ____ ____ ____ ____ ____ ____
ment. If any of your instruments are
IP address ______.______.______.______ connected to the site LAN, you will
Subnet mask ______.______.______.______ need to ask your network adminis-
DNS server ______.______.______.______ trator to provide static IP addresses
Hostname __________________________ for those instruments.
Domain name __________________________
DHCP On ____ Off ____
Auto IP On ____ Off ____ N/A ____
UPnP Enabled ____ Disabled ____ N/A ____
Figure 15.3. Browser-accessible welcome page for the Agilent 33220A function/arbitrary waveform
generator
3 Individual instruments will respond
differently to these changes. In most
cases the new settings will not take
effect until you cycle the instrument’s
power off then on. Some instruments
may change the IP address imme-
diately and you will have to enter
the new address into the browser’s
address box before making additional
changes.
www.agilent.com/find/open
139
Communicate via Interactive IO. You
Coexisting with other versions of VISA can use this software utility from
within Connection Expert (or from
Although the VISA API is a standard, each vendor employs different layers beneath IO Control) to verify communication
the VISA layer to control their hardware. For proper operation, the version of VISA with instruments via LAN:
installed on the system’s host PC must be compatible with the I/O hardware.
• Interactively send commands and
Fortunately, it’s possible to perform a side-by-side installation of Agilent VISA with view responses without writing
other VISA libraries and achieve proper I/O operation. For example, you can use program code
Agilent VISA and National Instruments’ NI-VISA in the same PC.
• Quickly verify connectivity to an
The VISA standard requires that visa32.dll, the dynamic-link library that implements instrument
the VISA interface, be installed in specific locations. Thus, only one installed
implementation can fully comply with the standard at one time. • Troubleshoot communication
problems
To avoid conflicts, the Agilent IO Libraries Suite gives you the option to install
Agilent VISA in side-by-side mode. In this case, Agilent VISA is installed in a • Learn the instrument’s command
different location and does not overwrite any other VISA already present on the PC. set
After installing Agilent VISA in side-by-side mode you can use either the Agilent or • Rapidly prototype commands and
NI VISA library in your programs. The side-by-side installation allows you to choose check the instrument’s responses
which VISA library you link your program against so you can take advantage of the before you begin writing code
support and features of each vendor’s VISA implementation.
You can start Interactive IO from
If you choose to use Agilent interface hardware in a program linked against within Connection Expert in one
NI-VISA, IO Libraries Suite online Help can guide you through the configuration of three ways: by clicking the Send
of NI-VISA. This is the easiest way to use your VISA code without rebuilding your commands to this instrument task in
program. (NI-VISA does not allow side-by-side mode by default and must be manu- the task guide; by clicking Tools then
ally enabled.) Interactive IO on the Connection
Expert menu bar; or by right-clicking
If you use NI GPIB cards and devices, you don’t need to install NI-VISA. Instead, the instrument in the explorer panel
you can simply install NI-488.2 as a driver for these devices and avoid the complica- and selecting Send Commands To
tions of side-by-side operation. You could then use Connection Expert to configure This Instrument.
Agilent VISA to use the NI GPIB card.
Whenever you mix interface hardware from different vendors, it’s best to configure Conclusion
each device using the configuration tools from its manufacturer. NI interfaces
Connecting instruments to LAN is
should be configured with the NI MAX utility prior to use. Agilent interfaces are
almost as simple as plugging in a
automatically configured when you install the IO Libraries Suite, but if you add a
network cable if you take advantage
new interface or decide to change interface properties, use Agilent Connection
of the libraries and utilities that
Expert.
are part of the Agilent IO Libraries
Suite and follow the quick, six-step
configuration process described in
this chapter. Once that process is
completed, GPIB, LAN, USB and
other standard interfaces are equally
easy to use within a test system.
What’s more, included tools such as
Agilent Connection Expert and a set
of six I/O utilities make it easy to
configure, debug and troubleshoot
those connections.
141
142 Section 3. LXI: The Future of Test
16. Value, Performance and Flexibility: The Promise of LXI
143
• System size. Whenever systems must There are also a few issues specific Addressing the challenges
be shipped elsewhere or deployed to each of today’s three major legacy
where floor space is at a premium, test-system architectures. with LXI
system size matters. Unfortunately,
• GPIB. Although this remains the To help test system engineers over-
with existing approaches, this may
current instrumentation standard, come all these drawbacks, Agilent is
also mean sacrificing functionality,
it has slower data transfer rates leading the way to a new vision for
performance and accuracy as the
than other architectures, forces test systems.
system shrinks.
you to install an interface card in
• Consistency. In systems that require your PC, requires expensive cables, Building on the widespread
source, measure, power and RF/ and allows only 14 devices on the
microwave capabilities, developers bus.
LAN foundation
often need to mix two or more Many current-generation instruments
of the current instrumentation • VXI. This architecture requires
include LAN ports (see Figure 16.1),
standards. This type of inefficiency an expensive cardcage, a slot-0
and LXI is the next logical step in the
also affects cost effectiveness. controller and an expensive,
evolution of LAN-based instrumenta-
proprietary interface (MXI).
tion. LAN is gaining momentum in
• Future-proofing. With limited versa-
• PXI. In addition to the VXI over- T&M because it has several inherent
tility, existing test architectures
head costs mentioned above, PXI advantages over most parallel and
make it difficult to meet future
has issues with size, power and serial interfaces. For example, LAN
needs—higher frequencies, greater
EMI that limit the range of solu- can handle an unlimited number of
accuracy, faster throughput and so
tions to those normally covered nodes and provides long distance
on. As more systems are deployed
by PC plug-in cards. PXI is also inter-device connectivity. It also
to remote locations, they become
transitioning to PXI Express, includes TCP/IP error checking and
increasingly difficult to manage
limiting the longevity of PXI fault detection—and these functions
and troubleshoot without onsite
modules or requiring expensive minimally impact throughput rates.
expertise. In addition, test systems
hybrid mainframes. Better still, LAN enables automatic
often remain in service longer than
device discovery, addressing, asset
the lifetime of most backplanes and
management and network manage-
interfaces. Computer backplanes—
Figure 16.1. Current-generation instruments ment. LAN also has a cost advantage:
ISA, EISA, VME, PCI, and Compact
such as the Agilent 34980A switch/measure the prices of cables, interface cards,
PCI—change every few years but unit include LAN, USB and GPIB interfaces. hubs, routers, switches, wireless
usually offer little or no backward
access points and so on are low and
compatibility. The instrumentation
continue to fall.
versions (VXI and PXI) have the
same drawback. To compound the
problem, standardized test and
measurement interfaces such as This is especially true if the test
GPIB and MXI have fallen short of system uses a dedicated network. To
both the increased speed and wide- learn more about creating private
networks for test systems, please see
spread adoption of LAN and USB.
Chapter 10.
Instead, new GPIB or MXI cards
must be developed and purchased
whenever computer architectures
change.
www.agilent.com/find/open
145
The advantages of LXI (typically used in small or ad-hoc Simplifying physical integration
networks). LXI also recommends The LXI standard also includes
LXI’s visionary approach delivers support for DNS, which instru- an optional mechanical specifica-
numerous benefits in virtually every ments can use to publish their tion that simplifies the integration
aspect of system design, implementa- host name through a DNS server of modules within system racks.
tion, operation, and maintenance. By (another feature usually available Compliant modules are full- or
addressing all of the shortcomings in corporate networks). half-rack wide and standard EIA
described earlier with other architec- heights (e.g. 1U, 2U). The standard
tures, LXI does more to help you • Application layer. LXI-compliant also specifies that signal input and
reduce the expense and effort required instruments support the VXI-11 output connectors (and status lights)
to create cost-effective test systems. protocol (based on remote proce- are placed on the front of the module
dure calls) for automatic discovery while power, Ethernet, triggering and
of new instruments and identifica- other control connectors are placed
Ease of use tion through the *IDN? query. on the rear.
The LXI standard address ease of use Enabling easy interaction Reducing set-up time
for system developers in a variety A few noteworthy instrument Through proven standards such as
of ways, including harnessing the control features help LXI surpass Ethernet and IVI drivers, LXI ensures
Ethernet standard, enabling easy GPIB in ease-of-use. For example, that everything is compatible—and
interaction via Web browsers, making LXI devices include a built-in Web setup will take less time. Because
programming more efficient with server to enable configuration and Web pages built into every LXI
standard drivers and simplifying troubleshooting. In many Agilent LXI instrument, a standard Web browser
physical integration. products, the instrument page also is all you need to view device
allows interactive instrument control information, change its configuration
Harnessing the Ethernet standard and monitoring, a capability that can
LXI includes some key elements and even monitor results and control
be very useful during system configu- measurements (in many Agilent LXI
that simplify the use of LAN in test ration and deployment. It also
systems: devices). You can also use proven
simplifies remote troubleshooting. tools such as LAN hardware, LAN
• Physical layer. To help ensure cables and ping servers to communi-
Making programming more efficient cate via LAN and troubleshoot local
successful instrument interac-
For programmatic control, you can or remote systems.
tion, the LXI standard specifies
use IVI instrument drivers (see
automatic negotiation of LAN
Chapter 3). Recommended by the LXI
transmission speed and duplex
standard, IVI-COM drivers are based Flexibility
communication. The standard also
on the widely used Microsoft COM
recommends Auto MDIX, a feature LXI’s modular approach makes it
architecture and work with today’s
that enables the use of either easy to mix and match modules
most popular test software environ-
straight-through or crossover LAN that provide the exact functionality
ments. These object-oriented drivers
cables in direct controller-to-instru- required for each system or applica-
use a hierarchical API, making it
ment or peer-to-peer connections. tion. LXI-compliant instruments
easy to utilize the advanced features
The instrument automatically provide new levels of flexibility in
of modern, object-oriented environ-
adjusts to the existing cable and its hardware selection, product testing,
ments. One key example is easy navi-
communication counterpart. software reuse, instrument commu-
gation through a driver’s hierarchy
of functions and simplified coding nication and even organizational
• Network (IP) layer. LXI instruments
via autosuggestion and autocomple- responsiveness.
support automatic IP configura-
tion through a DHCP server (often tion. Agilent LXI instruments also
available in managed corporate support ASCII-based SCPI commands
networks and in cable/DSL for programming flexibility and to
routers) or through dynamic support non-Windows environments.
configuration of local addresses
Working independently
With their embedded processors, Performance Assuming IPv4 and maximum frame
size, the bandwidth remaining for
today’s test instruments have
Test-and-measurement interfaces application data is about 95 percent
enough computing power to carry of the transmission rate.
out measurement tasks on their such as GPIB and MXI are chal-
own, freeing the system controller lenged by the need for increasing
for other tasks. LXI uses this power bandwidths and faster data-transfer
to provide greater flexibility in rates. One key advantage of LXI is its
communication, too: instruments ability to leverage ongoing innova-
can communicate without arbitra- tions in LAN that satisfy the need for
speed.
www.agilent.com/find/open
147
Accelerating system throughput Of course, security is a concern In comparison, Ethernet is an
Other aspects of LXI enhance for any application that requires extremely stable standard. Like
performance by enabling faster a connection outside your secure, GPIB, it’s more than 30 years
system throughput. For example, LXI well-controlled corporate network. old—and Ethernet is clearly here
makes it possible to build high-speed Rest assured that solutions designed to stay. With its stability and other
distributed systems comprised of for the IT world also work with virtues, Ethernet has been adopted
intelligent instruments that can LXI. You can utilize routers that in many industries, including corpo-
communicate with each other and include security features such as rate communications, consumer
operate in parallel. Devices stay access filtering based on MAC or IP electronics, industrial automation
synchronized through the IEEE addresses, WLAN encryption and and now test equipment.
1588 timing standard, LAN-based so on. If a distributed application
triggers, peer-to-peer and multicast needs to access the public Internet, Ethernet is also a living, evolving
messaging, and a hardware trigger you can use a virtual private network standard. It has accommodated the
bus. IEEE 1588 also accelerates (VPN) to send IP packets securely, addition of higher-layer protocols as
throughput via time-based triggering, encrypted through IPsec or other well as enhancements such as Gigabit
which initiates instrument opera- encryption protocols. Ethernet at the physical layer and
tions at a specific time rather than IPv6 at the network layer. Amazingly,
after a trigger or command. these enhancements have retained
Leverage and longevity backward compatibility, protecting
investments in previous versions of
Distributed applications In general, test systems address two
the standard.
large classes of devices: long-lived
Unlike cardcage-based systems, and short-lived products. Many Extending the life of LXI systems
LXI modules can be easily distrib- devices developed for aerospace and In addition to the continued evolu-
uted in a test rack, across a lab or defense applications require test tion and assured compatibility of
throughout a building. This allows systems that are easy to maintain LAN technology, two additional ideas
you to place instruments where and update far into the future. In extend the life of LXI systems in
they can best meet the needs of contrast, rapidly evolving commer- particular: the ability to download
each measurement or application. cial wireless products require test new capabilities or personalities
Examples include the monitoring systems that can be developed into intelligent instruments and the
systems used in environmental rapidly and within budget—and be possibility of injecting updated or
applications, power generation easily reused as the products evolve. upgraded technology into SI-based
and the process-control industry. The ability to meet the needs of systems. These capabilities simplify
Another intriguing example is testing either long- or short-lived devices the task—and reduce the cost—of
of wireless base stations: protocol improves with LXI, which is designed keeping pace with evolving measure-
test equipment can be placed near to fulfill the promise of long-lived ment standards, wider frequency
or inside base stations located many measurement hardware, I/O and ranges and tighter accuracy require-
miles apart. software. ments. Taking a wider view, LXI
enables new levels of versatility by
With LXI, these solutions can be This need for stability is in sharp
making it possible to configure or
designed using the same instruments contrast to the rapid innovation
reconfigure a system through soft-
you would use for local applications cycles in today’s computer buses. For
ware changes to IEEE 1588 clocking
and rack-based systems. There is no example, in just a few years instru-
and LAN triggering.
need to create custom gateways— ments based on computer buses have
remote access comes without extra had to change from ISA to EISA to LXI also helps you address future
effort. Using your corporate intranet PCI and now to PCI Express (a serial organizational needs. Test-system
or the public Internet, large distances bus not compatible with previous experts are becoming scarce in
can be bridged easily and the connec- parallel implementations). many organizations and can’t be
tion is transparent to the end user. everywhere at once—onsite, offshore
or anywhere in between. With LXI,
you can place test systems virtually
anywhere on your intranet, enabling
your team to perform centralized
troubleshooting, remote monitoring
and more.
www.agilent.com/find/open
149
1. The master clock sends a sync What’s especially appealing about A closer look at LXI
pulse and the current time IEEE 1588 is that it works across
to every other device on the Ethernet—the same Ethernet being The LXI standard defines instrument-
network. All slaves set their used for instrument control. No specific requirements needed to
clocks to the master time. additional cables are required. ensure reliable, accurate operation
Depending on the size of the network in a system environment:
2. Each slave sends a time-stamped and its variation in latency times,
reply to the master. The master it is possible to achieve precise • Cooling
calculates the offset between the synchronization of LXI devices • Triggering
original transmission time of the located anywhere on a network—local
sync pulse and the various or remote. • Interrupt handling
reception times.
• Mechanical interfaces
To learn more about IEEE 1588, visit
3. The master sends an offset value the National Institute of Science • Software interfaces
to each slave, which adjusts and Technology (NIST) Web site at
its clock to compensate for the http://ieee1588.nist.gov. • Electromagnetic and radio
difference between the master’s frequency interference
sync pulse and its reception time
at the slave. After this initial Security • Network routing and switching
Figure 16.3. In an IEEE 1588-enabled network, a simple process ensures precise synchronization
between all devices
Switch/router
2: Reply 2: Reply
2: Reply 2: Reply
A B D E
(slave) (slave) (slave) (slave)
1: Sync 3: Offset
C
Master clock
Indicator
lights
Width
(IEC full- or half-rack)
Signals
www.agilent.com/find/open
151
• Addresses. LXI devices must Instrument pages Trigging and synchronization
support IP addresses (assigned
by the server), MAC addresses Every LXI-compliant device must One especially intriguing aspect of
(assigned by the manufacturer) be able to serve its own Web page. LXI is its triggering and synchroni-
and hostnames (assigned by the This page provides key informa- zation capabilities. LXI provides a
user). tion about the device, including its variety of optional triggering modes
manufacturer, model number, serial that are not available in GPIB, PXI
• Configuration. Compliant devices number, description, hostname, or VXI.
must support ICMP (ping server), MAC address and IP address. The
DHCP-based assignment of IP standard also requires a browser- The three classes of LXI devices
addresses, manual Domain Name accessible configuration page that implement these capabilities to an
Server (DNS) and Dynamic DNS. allows the user to change parameters increasing degree. As an example
Because DNS can translate domain such as hostname, description, IP of what is possible with LXI, all
names into IP addresses, it can address, subnet mask and TCP/IP Class B and A LXI instruments
contribute to the longevity of configuration mode. Accessing these (optional in Class C) can utilize
system software: IP addresses may Web pages is as simple as typing triggers embedded in LAN packets
change but domain names will not. the instrument IP address into the that can originate from any device
address line of any web browser. on the network—a PC or another
• Default conditions. As a safeguard, instrument. One device can send a
LXI defines a set of default LAN Agilent’s LXI-compliant instruments multicast message that triggers all
conditions and requires a “LAN go beyond the LXI requirements, instruments on the network without
configuration initialize” (LCI) providing monitor and control the need for a real-time computer.
switch that will reset a device to capabilities (see Figure 16.5). For Peer-to-peer messages can enable
this set of known conditions. example, you can set up a DMM, measurement scripts or cause data
command it to start making measure- to be passed from one LXI device
Programmatic interface ments and then read the results. to another without involving the
Some of our LXI devices even allow system’s host computer (a potential
Because the LXI standard you to download complete measure- communication bottleneck).
requires that all devices have an ment personalities (for CDMA, GSM,
Interchangeable Virtual Instrument or Wi-Fi, for example) into the instru-
(IVI) driver, it allows you to use ment and perform specific measure-
whichever programming language or ments with one command. The ability
development environment you prefer. to control an instrument through its
IVI-COM and IVI-C are well-estab- browser interface opens up a realm
lished industry standard drivers that of new possibilities for test engineers
instrument makers supply with their who need a simple way to access test
products. systems from virtually anywhere in
the world.
The LXI standard also mandates that
compliant devices implement LAN Figure 16.5. LXI specifies an informative instrument page that can be accessed with a standard
discovery, which enables the host PC Web browser
to identify connected instruments.
Currently, the LXI standard requires
use of the VXI-11 protocol, which
defines LAN-based connectivity for
all types of test equipment, not just
VXI. Going forward, future revisions
to the LXI standard may include
other proven discovery mechanisms
such as Universal Plug&Play (UPnP).
www.agilent.com/find/open
153
Expert troubleshooting traditional instruments such as spec- Peer-to-peer triggering
trum analyzers, signal generators and
Whether a system is located in the oscilloscopes. These fundamental By making it possible for one instru-
next room, the building next door or building blocks depend on PC ment or device to send triggers and
a site halfway around the world, your software that dynamically aggregates information to another, LXI frees up
system developer (or product expert) and “synthesizes” different measure- the PC to perform other, higher-level
can check its operation and trouble- ment tasks. tasks. Peer-to-peer triggering also
shoot problems. No travel is required: eliminates the need for an expensive
simply type an instrument’s URL As an example, an RF downcon- real-time controller to issue precise
or IP address into a standard Web verter LXI module could be used triggers to the instruments in a
browser and the instrument page will for spectral measurements in one system. Ultimately, overall test time
appear. test sequence and then be reconfig- can be reduced because techniques
ured for network measurements in such as wait states and status
another. To create the stimulus signal queries will be used less often in
Intelligent instruments for network analysis, simply adding system software.
a different LXI upconverter makes it
Without the size restrictions of
easy to change the output frequency
VXI and PXI, LXI enables use of
range without having to purchase Time-based triggering
intelligent instruments within a
an entirely new signal generator.
system. You can download measure- With IEEE 1588, time-based
Reducing the redundancy—and
ment personalities into a spectrum triggering may prove to be a
increasing the utilization—of these
analyzer, sophisticated signals into revolutionary way to synchronize
fundamental hardware elements
an arbitrary waveform generator measurements within systems and
helps trim the size and cost of test
or complex power sequences into a between instruments. For example,
systems (Figure 16.6).
programmable dc supply—and let the this method eliminates the need for
instrument handle the details. The trigger-specific external cabling so is
capabilities built into these instru- Synthetic instruments not limited by the distance between
ments help you save time, too. instruments. All measured data can
In addition to the attributes be time stamped, making post-test
You can reduce programming time mentioned earlier, SIs create two analysis easier, more efficient and
by taking advantage of the soft- additional possibilities. SI hardware more meaningful. System throughput
ware (and firmware) developed by and software modules can be used also increases because each instru-
the vendor rather than writing it to emulate obsolete instruments, ment can start at a specific time
yourself. Instrument set-up time can removing the burden (and cost) of rather than waiting for a trigger or
be reduced by creating configurations supporting outdated equipment in command.
in advance and recalling them as long-lived systems. SIs also make it
needed. Data transfers take less time possible to create and perform totally
because the instrument can make unique measurements that are not
measurements, perform the required currently possible with traditional
analysis and then send results—not instruments.
large data blocks—to the host PC.
Figure 16.6. LXI devices reduce the size and footprint of test systems
Rapid reconfiguration
LXI-based synthetic instruments
reduce system size and cost by
utilizing multi-purpose modules—
digitizers, waveform generators,
upconverters, downconverters and
more—that can be combined to create
RF/
microwave Analog Digital
signal signal data
Control To/from
embedded CPU
www.agilent.com/find/open
155
Appendix 16B: Creating Create versatile solutions with Achieve efficient development
cost-effective measurement system-ready instrumentation with open software tools
solutions with Agilent Agilent Open instruments are Configure a typical system in less
Open to test your way designed for faster throughput and
easier integration—in test software
than 15 minutes with the Agilent
IO Libraries Suite, which supports
LXI solves the key problems faced and system racks. Choose classic literally thousands of instruments
by system developers: it cuts costs, benchtop instruments for R&D then from hundreds of vendors. Get your
reduces system size, simplifies use their modular, system-optimized systems up and running sooner with
integration, accelerates throughput equivalents in manufacturing—and industry-standard IVI drivers that
and provides more opportunities for run the same software with few or no put instrument functionality at your
reuse of both hardware and software. changes. To reduce software develop- fingertips—and work in the application
These benefits make LXI a test archi- ment time, many instruments let you development environment you prefer.
tecture for today and into the future. install measurement personalities
that address specific applications,
To help you fully realize these including Mobile WiMAX, jitter,
Develop hybrid systems that
benefits, we’ve adopted LXI as part phase noise and more. combine new and existing
of the Agilent Open concept (Figure assets
16.8). Your test system architecture
should give you choices. Its range
Simplify system communication Protect your existing assets by easily
of possibilities should fit your and connectivity integrating GPIB instruments into
requirements, your preferences and LAN- and USB-based systems with
your existing test assets—hardware, Choose the best connection for your Agilent interface gateways and
software and I/O. This is the power requirements through instruments converters. You can even add VXI
of Agilent Open, a combination of equipped with GPIB, LAN and USB and PXI equipment to LAN-based
proven standards and time-saving ports. You can connect those instru- systems via LAN slot-0 cards. Utilize
tools for test automation. By giving ments quickly and easily with the the multiple I/O ports of Agilent Open
you greater flexibility, Agilent Open Agilent IO Libraries Suite software, instruments to connect via GPIB now
accelerates the creation of cost- which supports the major test-system and LAN or USB in the future. Using
effective measurement solutions— interfaces—GPIB, LAN, USB, VXI and VISA, making the transition from
and enables testing, your way. RS-232C. With support for LXI, you GPIB to LAN or USB requires nothing
can control instruments and monitor more than simple address changes in
measurements remotely via the your system software.
Figure 16.8. Through Agilent Open and LXI, LAN Web servers built into Agilent Open
becomes the backbone of test systems that easily instruments. Agilent is leading the way in migrating
incorporate present and future test assets.
test systems to the advanced capa-
bilities of LAN. We’re continually
introducing new additions to what is
currently the industry’s largest port-
folio of LAN-enabled instruments. At
the same time, we’re also protecting
your investment in GPIB instruments
by offering devices such as the Agilent
E5810A LAN/GPIB gateway and the
82357A USB/GPIB interface.
157
Typical LAN-based systems you can transition your system with an IVI driver and meets LXI’s
without changing its software. The physical specifications.
Making the transition to LXI doesn’t E5810A gateway can be mounted
require sweeping changes to your Scenario 3—All LAN. LXI-based prod-
in the system rack, which, with a
system architecture. Instead, a ucts are becoming widely available
LAN connection, is freed from the
variety of evolutionary system so it is now possible to evolve to an
distance constraints of GPIB.
structures are possible when using all-LAN structure. These systems will
LAN communication along with Scenario 2—GPIB plus LAN. A typical include one or more LAN routers as
LAN-enabled and LXI-compliant next step in the transition to LXI is needed to accommodate all of the
instruments. the addition of a LAN router between LXI instruments (see Figure 17.3).
the PC and the LAN/GPIB gateway Every instrument will be able to
In all cases, the starting point is (see Figure 17.2). This makes it take advantage of LAN’s speed while
a PC with a built-in LAN port: possible to incorporate GPIB, LAN utilizing low-cost I/O cabling. The
unlike GPIB-based systems, the PC and LXI equipment into a single browser-based interface within every
needs no physical modifications. system by connecting GPIB instru- LXI instrument will help speed and
However, the system structure does ments to a LAN/GPIB gateway and simplify instrument—and system—
require the addition of a switch then connecting the gateway and configuration and troubleshooting.
or gateway (external to the PC) to any LAN or LXI instruments to the The long reach of LAN and the
enable communication with multiple router. synchronization made possible by the
instruments. IEEE 1588 precision time protocol
.One important point is worth (PTP) will enable a variety of new
Scenario 1—GPIB-to-LAN. The easiest remembering: although many test capabilities and applications. (Please
initial transition is to use LAN to instruments are equipped with LAN refer to Chapter 10, Using LAN in
communicate with an existing GPIB ports, not all can be controlled via Test Systems: Network Configuration
system. A device such as the Agilent LAN. Some use the LAN port only to and Basic Security, for advice on
E5810A LAN/GPIB gateway (see communicate with external peripher- setting up a private, protected
Figure 17.1) enables remote access als—so it’s best to check the product measurement network using either a
to GPIB instruments via LAN—and manual or built-in help function to router-based or PC-based approach.)
eliminates the need to install a GPIB verify LAN-specific capabilities. Of
card in the PC. Addressing remains course, if an instrument carries the Please refer to Chapter 16, Value,
the same: your system software will LXI logo, it is has passed compliance Performance and Flexibility: the
see the gateway device as a GPIB testing and, at a minimum, can be Promise of LXI, for a closer look
interface even though it communi- controlled via LAN, has a browser- at the fundamental concepts and
cates via LAN. Because your instru- accessible Web interface, is provided advantages of LXI.
ments still look like GPIB devices,
Figure 17.1. A LAN/GPIB gateway can connect Figure 17.2. A router plus a LAN/GPIB gateway Figure 17.3. Using multiple routers enables
GPIB instruments to a PC’s LAN port enables connections of GPIB, LAN and LXI connection of local and remote LXI instruments
instruments to a PC’s LAN port to a PC’s LAN port.
LAN LAN
Router
LAN/GPIB LAN LAN
gateway
LAN/GPIB
GPIB gateway Router
GPIB
Router
www.agilent.com/find/open
159
Many Agilent instruments provide Simplifying software via search and replace. For larger
additional built-in pages that let programs or test suites, you can
you interact with the instrument changes modify the device declarations within
and perform various tasks: make the program. You can also use Agilent
Making the transition to LXI doesn’t
measurements, generate signals, IO Libraries Suite, which allows a
require sweeping changes to your
close channels, read values and table of aliases. This approach may
system architecture or your system
display results (see Figure 17.5). cause slightly slower communication
software. Four important items can
Some will also let you try program but provides a convenient way to get
simplify the process of modifying
commands or command sequences your system up and running.
your system software to communi-
and verify the instrument’s response.
cate with LXI-compliant devices.
Agilent IO Libraries Suite, Agilent VXI-11
Connection Expert and the LXI Dual-interface instruments From the perspective of a PC applica-
browser interface are a powerful
tion, many instruments implement
combination that can reduce set Many of Agilent’s GPIB instruments
the VXI-11 communication protocol
up time from days to minutes. Best that are updated to LXI compliance
that makes LAN I/O look just like a
of all, the IO Libraries Suite and will have both LAN and GPIB ports.
GPIB connection. In practice, this
Connection Expert are designed to These instruments can be used via
means software written for GPIB is
work with instruments from virtu- the GPIB port without modifying
likely to work with identical LAN-
ally every manufacturer. Agilent your existing software or through
enabled instruments that implement
customers can download the IO the LAN port with a simple address
VXI-11.
Libraries Suite at no charge from change in your program. For smaller
www.agilent.com/find/open. programs, you can change addresses
from GPIB to IP either manually or Command-compatible
instruments
Figure 17.5. Many of Agilent’s LXI instruments include built-in web pages that let you configure the
Instruments that are command-
device and make measurements. For example, this page on the 34980A multifunction switch/mea-
sure unit makes it easy to configure each module in the system.. compatible with older instruments
can be used with your existing
software. As an example, the Agilent
34410A and 34411A digital multime-
ters have a compatibility mode that
mimics the widely used 34401A or
E1412A DMMs. The LXI-compliant
34410A and 34411A, equipped with
both GPIB and LAN interfaces, can
replace a 34401A in a system with
either no changes (GPIB) or minor
changes (IP addressing for LAN) to
the system software.
www.agilent.com/find/open
161
162 17. Transitioning from GPIB to LXI
18. Creating Hybrid Test Systems with PXI, VXI and LXI
163
Advantages of PXI and VXI Disadvantages of PXI and VXI
Programming register-
Both PXI and VXI have useful Both PXI and VXI have shortcom- based devices
advantages in hardware, connectivity ings that can affect your ability to
and programming when compared to create a system that fully satisfies Because PXI and VXI are leveraged
rack-and-stack systems. the budgetary, technical and lifecycle from computer buses (PCI and VME),
requirements of a test specification. their plug-in cards usually depend on
Hardware. One key advantage is the register-based operations to read or
density of switching, sourcing and System host. Unlike a typical rack-and- set attributes, initiate measurements,
measuring capabilities that can be stack system, a PXI- or VXI-based load or unload data, and so on. While
packed into a single mainframe. PXI system can be heavily dependent on this type of low-level programming
and VXI will usually be smaller than the performance of the host PC—and enables detailed computer control
a rack-and-stack system with similar higher performance commands a of each module, it can be quite
functionality. PXI and VXI also have higher price. What’s more, the PC- complicated and time consuming.
an edge over rack-and-stack in trig- dependent approach does not scale
gering and synchronization, thanks well for large, complicated systems: One popular solution is device
to their high-speed backplanes and as more modules move more data drivers, which handle the low-level
included triggering capabilities. more often, the PC can become a details and enable programming
processing bottleneck that slows at a higher level. The best choice
Connectivity. PXI and VXI offer a overall system performance. of driver depends on the type of
variety of I/O alternatives: MXI, hardware or software being used.
GPIB, LAN, USB, FireWire and serial. Embedded controllers come with For example, National Instruments
This allows you to make case-by-case their own set of shortcomings. uses IVI-G drivers with LabVIEW and
tradeoffs between performance and Because these are a specialty item IVI-C drivers with LabWindows. While
convenience. produced in limited quantities, they IVI-C and IVI-G drivers are available
typically cost three to eight times for many Agilent instruments, Agilent
Programming. System creators can as much as an equivalent desktop and others have provided IVI-COM
use graphical or text-based develop- PC. They also tend to lag behind the drivers. These are language- and
ment environments to create the latest advances in performance and platform-neutral and one version will
required measurement and analysis capabilities. work in all Microsoft® COM (and
functionality. While it can be difficult
compatible) environments, and with
to work with register-based PXI Hardware. In addition to the high
Microsoft Excel. Not only does this
and VXI plug-in cards, the use of entry cost of PXI and VXI, you may
provide additional flexibility because
device drivers can greatly simplify need to buy a mainframe that has
you can work in your preferred
communication and programming more slots than needed if you want
development environment, but it
(see “Programming register-based to allow for future expansion. Once
also can enhance your productivity
devices”). The resulting measurement a mainframe is filled, there is also
through features such as IntelliSense
solution may be smaller and faster the potential cost of adding another
pop-up menus that provide onscreen
than an equivalent rack-and-stack mainframe if the system needs just
command-completion help.
system built with benchtop GPIB one more plug-in card.
instruments.
When the required functionality
isn’t available in a modular format it
will be necessary to add bench-type
instruments to the system. Examples
include many RF measurements as
well as high-wattage power supplies.
The inclusion of standalone instru-
ments can increase the complexity
of both system integration and
programming. It may also negate the
size advantages of VXI or PXI.
164 18. Creating Hybrid Test Systems with PXI, VXI and LXI
Connectivity. Using either MXI or GPIB Exploring LAN-based Scenario 1: VXI and GPIB
as the interface adds hardware cost
and configuration complexity to an hybrid systems If a VXI mainframe contains a GPIB
external host PC. slot-0 card, it can be connected to the
As earlier chapters have noted,
PC via LAN by adding an interface
Programming. Because most PXI and LAN is rapidly gaining favor as the
converter such as the Agilent
VXI devices lack any sort of built- interface of choice for automated
E5810A LAN/GPIB gateway (Figure
in user interface—front panel or test systems. While the earliest
18.2). With the gateway connected
browser-based—you typically have to LAN-enabled instruments offered
between the VXI mainframe (GPIB)
purchase, install and configure some inconsistent implementations of
and the router (LAN), any applica-
type of software to control even the the interface, the LXI standard now
tion running on the PC will be able
simplest device. Additional program- ensures a consistent approach that
communicate transparently with the
ming may be required to perform a makes it possible to use compliant
VXI hardware as GPIB devices.
measurement, manipulate the data instruments from multiple vendors.
and analyze the results. What’s more, Chapter 16 offers a closer look at • Advantages. This hybrid structure
T&M-specific software that provides LXI. eliminates the need to install a
these capabilities tends to be much GPIB card in the PC. With the
In most cases, it is relatively straight-
more expensive than commercial gateway, addressing can be kept
forward to create hybrid systems
programming environments available the same so no software changes
that utilize LXI devices alongside
from Microsoft and other vendors. will be required.
GPIB, PXI and VXI hardware. A
Cost. Both PXI and VXI incur a hybrid structure lets you harness • Disadvantages. System performance
large overhead cost in the main- the advantages of each architecture may decrease if the gateway cannot
frame, controller, connectors and within a single system. In addition keep pace with the demands of
I/O subsystem. PXI and VXI both to saving money by protecting your any high performance measure-
require a considerable investment existing investments in test assets, ment cards installed in the VXI
before the first module can be used. this approach also helps you save mainframe.
Additionally, the per-slot overhead time because you can continue using
costs can be prohibitive when the familiar hardware, interfaces and
Figure 18.2. With a router and a LAN/GPIB
PXI or VXI mainframe contains software. gateway, test software on the PC can commu-
low-cost modules such as digital IO, nicate transparently with VXI instruments, as if
A typical LXI-based system starts they were GPIB devices.
DACs, or simple switching.
with a host PC and its built-in LAN
port, which provides a connection to
local and remote LXI-based devices
through commercially available LAN
LAN switches or routers. This is Router
also the starting point for hybrid LAN
configurations that include LXI LAN/GPIB
devices working alongside a VXI or gateway
PXI mainframe. Today, four possible GPIB
scenarios are likely and feasible.
www.agilent.com/find/open
165
Scenario 2: VXI and LAN Scenario 3: Embedded Figure 18.3. Adding a LAN slot-0 card to a VXI
mainframe lets you create a LAN-based hybrid
When a VXI mainframe is equipped
controller VXI/LXI system..
with a LAN slot-0 card, adding it If a PXI- or VXI-based system is
to the system network is as simple using an embedded controller within
as connecting it to the LAN router the mainframe, the controller can LAN
(Figure 18.3). Even if the LAN- be connected to the test-system Router
equipped VXI system is not LXI network through its built-in LAN
compliant, it can coexist on the port. The PXI or VXI portion of the LAN
network with any LXI devices. system would still be controlled by
the existing software running on
• Advantages. Every instrument in
the embedded PC. To simplify the
the system—LXI or VXI—can utilize
overall system structure, the existing
LAN’s I/O speed. If the system
software could be modified to control
software is already programmed
the LXI devices, eliminating the need
to communicate with the VXI
for an external PC that controls only
hardware via LAN, addressing
LXI devices (Figure 18.4).
should remain the same so few VXI mainframe LAN and LXI
or no software changes will be with LAN slot 0 instruments
• Advantages. This is a straightfor-
required. Any required program- ward way to add the advantages of
ming changes should be relatively LXI to a PXI- or VXI-based system.
modest, even when you replace an If suitable modular LXI devices are
MXI or FireWire slot-0 card with a available to provide functionality
LAN slot-0 card. that isn’t available in PXI or VXI Figure 18.4. An embedded PC with a LAN port
formats, the resulting system can be used as the system controller in a hybrid
• Disadvantages. Depending on the
may also be simpler and more VXI/LXI or PXI/LXI system.
devices installed, this configuration
compact than one that uses GPIB
may provide less performance than
instruments.
a purely backplane-based system
(e.g., one that uses an MXI inter- • Disadvantages. Modifying the Router
face) but should be faster than the existing software to control the
LAN/GPIB configuration described LXI devices could hinder system LAN
in Scenario 1. performance by putting an addi-
tional burden on the embedded PC;
however, this may have a modest
One example is the VXI Technology
EX2500 LXI-VXI Gigabit Ethernet
impact given the built-in intel-
Slot-0 Interface. ligence of most LXI devices. This VXI or PXI
system structure also requires the mainframe with
embedded PC
addition of a LAN router, which
will cause a slight increase in LAN and LXI
system cost and complexity. instruments
166 18. Creating Hybrid Test Systems with PXI, VXI and LXI
Scenario 4: LXI-compliant Going beyond hybrid Figure 18.5. An LXI-compliant mainframe brings
the benefits of PXI switching to a hybrid PXI/LXI
mainframe to all-LXI system.
The Pickering Interfaces 60-100 and The use of LXI-based synthetic instru-
60-101 are seven-slot chassis that ments is more similar to PXI and VXI
support a variety of 3U PXI modules. in philosophy and approach. This
topic is covered in detail in Chapter 19.
www.agilent.com/find/open
167
Conclusion
The PXI and VXI architectures offer
a number of advantages over rack-
and-stack approaches, but they also
present some disadvantages than can
limit a test engineer’s ability to meet
technical and economic constraints
in a test system. LAN-based hybrid
systems that incorporate PXI or
VXI instruments can be a cost-effec-
tive way to leverage equipment
and programming efforts across
the product life cycle. This chapter
explored four hybrid options that
deliver the benefits of LAN connec-
tivity while protecting investments
in existing instrumentation. As
LXI solutions are becoming more
pervasive across the T&M spectrum,
the transition to an all-LAN LXI
approach is becoming increasingly
feasible and attractive from both
functional and financial perspectives.
168 18. Creating Hybrid Test Systems with PXI, VXI and LXI
19. Assessing Synthetic Instruments
Introduction This chapter will help you assess These are ambitious goals, but the
the potential value of SI relative to DoD, defense contractors and equip-
For decades, automated test systems your present or future requirements. ment manufacturers believe they can
(ATS) built around benchtop instru- It presents a brief history of SI, be achieved over time by applying
ments have been the dominant compares a rack-and-stack system advances in commercial technologies
test-system architecture. In the late to an SI-based system, describes the (LXI is one important recent example).
1980s, modular VXI-based systems initial applications of SIs and illus-
addressed several shortcomings trates the emulation of conventional The greatest progress toward these
of the rack-and-stack approach. In instruments with SIs. goals is coming from the use of
particular, card-based instruments SIs. According to the Synthetic
mounted in a multi-slot mainframe Reviewing the roots of SI Instruments Working Group (SIWG),
a synthetic instrument is a recon-
reduced the size and weight of
systems. The speed and capabilities In the mid 1990s, the U.S. figurable system that links a series
of the VXI backplane also enabled Department of Defense (DoD) of elemental hardware and software
enhanced triggering and faster assigned the U.S. Navy the task of components via standardized
data transfers. However, all such developing new types of ATS for interfaces to generate signals or
commercial technologies tend to the testing of avionics and weapons make measurements using numeric
have lifecycles that are much shorter systems. This ongoing project has six processing techniques. The key word
than a typical aerospace or defense driving goals: is reconfigurable: the elemental
system, possibly affecting long-term blocks can be arranged and rear-
maintenance and support of an ATS. • Reduce the total cost of ownership ranged via software commands—and
of ATS the signals rerouted via switching—
These issues are the driving force to emulate one or more types of
behind an approach called synthetic • Reduce the time to develop and traditional test equipment.
instrumentation (SI). The concept deploy new or upgraded ATS
is simple: SIs let you configure and To achieve this flexibility, an SI
• Reduce the physical footprint of will contain as many as four major
reconfigure modular hardware and
each system components: signal conditioners,
software elements to create the
functionality of multiple measure- frequency converters, data
• Reduce the logistics footprint via
ment devices. This building-block converters and numeric processors.
decreased spares, support systems
approach makes it possible to update For example, the basic block diagram
and training
or upgrade an ATS or a Test Program shown in Figure 19.1 describes most
Set (TPS) by simply replacing a • Provide greater flexibility through microwave instruments, including
single module such as a digitizer or systems that are interoperable spectrum analyzers, frequency coun-
downconverter. It can also reduce the among U.S. and allied services ters, network analyzers and signal
burden of software updates over the generators.
lifetime of an ATS. • Improve the overall quality of
testing
Includes joint participation of the
DoD, prime contractors and suppliers.
Figure 19.1. Basic architecture of an RF/microwave synthetic instrument
Control To/from
embedded
Control CPU
169
Unlike general purpose instruments, Software is another matter. Currently, Comparing present and
which are optimized to perform one substantial effort is required to
task (e.g., spectrum analysis or signal create the software modules that future approaches
generation), the synthetic instru- provide essential functionality such
Whether you view SIs as a near-term
ment architecture is optimized to as the measurements and calibration
requirement, a long-term curiosity or
provide greater efficiency in an ATS routines needed to replace a stand-
something in between, a comparison
by reducing redundant elements such alone instrument. There is also the
with traditional approaches reveals
as the digitizers and downconverters time and effort required to support
some interesting highlights. Within
found in multiple instruments used software written in-house. If you add
the context of the DoD’s driving
within current systems. to that the typical effort required to
goals, it is easy to illustrate the
create the mainline test program or
The DoD expects these SI modules advantages of synthetic instruments
suite of TPS, then the total up-front
to come from a variety of vendors, over GPIB, VXI or PXI solutions.
development cost is acceptable only if
enabling easy mixing and matching SIs are required.
as requirements change or modules Reviewing purpose and usage
become obsolete. What’s more, any Looking ahead, SI vendors recognize
substitution of modules—replacement the need for software tools that will The main purpose for a military-
or “technology insertion”—should reduce effort, accelerate development related system is to test devices or
require only minimal changes to the and ensure accurate, repeatable assemblies in locations such as the
core system software. results. As these tools become readily flight line, an aircraft hangar or a
available and reuse of software repair depot. The same system may
Putting SIs in perspective modules becomes more commonplace, also be used in the original manufac-
the development costs for SI-based turer’s facility.
SIs are clearly intended to address systems should decrease. However,
a specific set of needs that are vendors need to address one key When the test system is fielded for
especially important to the military, issue: the interchangeability of soft- military use, the top priority is
but may also be relevant to some ware components. If vendor substitu- to identify and replace defective
commercial organizations. For tion is equally viable with both the electronic systems or assemblies
example, if your company is bidding hardware and software elements of as quickly as possible to return an
on a contract that requires or gives SIs, then the major benefits of the aircraft or vehicle to operational
preference to the NxTest concept, NxTest vision will be within reach. service. The second priority is
then SI will be required. Longer to repair the defective system or
term, commercial organizations that assembly and put it into the inven-
utilize outsourcing and offshore
Weighing commercial tory of spares.
manufacturing may benefit from the applications
use of SIs in test systems they define The usage model for such test
or provide. If you develop systems within a systems involves rapid deployment,
commercial organization, the busi- perhaps into areas of conflict. Putting
ness model for most automated test the systems closer to the aircraft or
Assessing the situation applications probably can’t support vehicles they support translates into
the higher initial costs of developing higher levels of operational readi-
Two factors will affect the rate SI-based solutions. Of course, this ness—and reduced downtime—for
of SI adoption in the near term: requires a case-by-case assessment, those aircraft or vehicles. In this
hardware availability and software and only you can decide if the poten- scenario, flexibility and easy main-
effort. Gradually, a wider variety of tial benefits outweigh the current tenance are more important than
hardware is becoming available, and tradeoffs. Some early adopters may absolute measurement throughput.
Agilent is in the vanguard of both SI find SI to be very useful in a specific
and LXI. The LXI standard, which application.
addresses the needs of synthetic
instrumentation, is perhaps the most Over the longer term, the promise of
promising platform for SI due to SI is well worth watching for many
the potential longevity of the LAN commercial firms. As more hardware
interface (see Chapter 16 for more and software modules become
on LXI). available, the economic benefits will
increase for commercial applications.
www.agilent.com/find/open
171
Understanding the SI approach type of system is smaller, lighter and Deciding if SIs are right for you
easier to transport. It also simpli-
With SI, the fundamental elements fies support by making it easier to In the near term, SIs offer useful
of multiple instruments are real- replace or upgrade individual instru- benefits that must be weighed against
ized through functional modules ment modules as needed. the tradeoffs. The main overall
such as digitizers, upconverters, benefit of the modular, building-block
downconverters and arbitrary A rear-panel view of the SI system approach is greater flexibility in
waveform generators. By arranging would reveal LAN ports on each less space. This approach will also
and rearranging the interconnec- module. By creating LXI-compliant make it easier to replace individual
tion of these building blocks and SIs, Agilent is providing a PC-to- modules or implement technology
the associated software modules, it instrument interface that delivers the insertion when new, updated
is possible to emulate the function- stability, longevity and performance capabilities are available. Longer
ality of an oscilloscope, a spectrum of LAN. This simplifies PC connec- term, this should also make it easier
analyzer, a power meter and other tivity and also helps lower the total to replace any modules that become
instruments in much less physical cost of ownership for the ATS. obsolete.
space. Operationally, this is a
software-intensive process in which The rear view would also show a Currently, the major tradeoff is the
the system could perform a series of hardware trigger bus cable that intensity of the software effort, but
tests by configuring the hardware, complements a variety of LAN-based as mentioned earlier, this is likely to
any needed switching and the associ- triggering capabilities. In combina- change as hardware vendors begin to
ated software module for one type of tion, these triggering capabilities provide the necessary software tools.
measurement and then reconfiguring equal or surpass the capabilities VXI Looking at the whole system, another
the hardware, switching and soft- and PXI. possible tradeoff is in the ability
ware modules for the next type of of SI-based systems to scale grace-
System software fully: larger systems will tend to put
measurement.
The points mentioned earlier greater demands on the host PC. A
System hardware regarding software development and higher-performance PC may be able
As a comparison, a rack-and-stack maintenance still apply. Currently, to handle the demands of a complex
system containing a spectrum creating the necessary measurement system, but a faster processor (or
analyzer, three microwave sources and calibration functionality requires multiple processors) and more
and a power meter might occupy a significant development effort. memory also means a higher price
18U of rack space. Using a variety However, any software modules for the PC and therefore a higher
of half-rack SI modules that don’t designed for transportability can be total cost for the system. The use of
have displays or keypads, the same reused with other SI-based systems intelligent, LXI-based instruments,
functionality occupies 11U of rack and potentially with other hardware which can offload many computing
space, as shown in Figure 19.3. This modules. tasks from the PC, is another way to
head off this potential issue.
Figure 19.3. An SI-based system can provide equivalent (or greater) functionality in less rack space.
Upconverters
Spectrum
Downconverter analyzer
11U Microwave
sources
IF digitizer 18U
Signal processing for downconverter
Power meter
www.agilent.com/find/open
173
Utilizing current SI devices N8212A. This high performance 20 Figure 19.4. Agilent N8211A 20/40 GHz perfor-
mance analog upconverter
GHz vector upconverter functions
In May 2006, Agilent’s initial offering as a microwave source with greater
of six synthetic instruments became than 2 GHz I/Q modulation band-
the first Class A LXI products to width. It features AM, FM and pulse
achieve certification from the LXI modulation (via external or internal
Consortium. These SIs demonstrate modulation) and multisource
Agilent’s ability to leverage proven coherent carrier capability. The
RF technologies into innovative LXI- N8212A is based on the Agilent PSG
based solutions that serve the needs vector signal generator and includes
of the DoD, its prime contractors and options for greater spectral purity
others who can benefit from the flex- and enhanced phase noise.
ibility of modular instrumentation.
N8221A. This 30 MSa/s IF digitizer has
a 7.5 MHz IF input and provides 80
Reviewing the original six dB dynamic range, 14-bit resolution,
and 8 MHz modulation bandwidth.
N8201A. This high performance Figure 19.5. Agilent N8241A arbitrary waveform
26.5 GHz downconverter provides generator
IF output frequencies of 7.5, 21.4
and 321.4 MHz, enabling three
different signal bandwidth capabili-
ties. External mixing can be utilized
to downconvert microwave signals
as high as 110 GHz. The N8201A
is leveraged from the Agilent PSA
Series spectrum analyzers.
A caveat
Emulating RF instruments You can use SIs to emulate a legacy
instrument up to the point where
This versatile set of modules can be
quickly and easily reconfigured to Figure 19.6. For signal generation, the SI chain includes numeric processing, data conversion and
make a host of measurements that frequency conversion.
would ordinarily require a vector
signal analyzer, spectrum analyzer
and oscilloscope. They can also be
used to emulate the capabilities of an
obsolete instrument such as the HP
uW signal out
8902A measuring receiver. Two brief
examples will illustrate some of the Arbitrary
possibilities. Signal generator waveform Upconverter
software generator
SI stimulus unit
This requires signal generation hard-
ware and software modules to create Figure 19.7. Agilent N8241A arbitrary waveform generator
the required signals and perform
scalar or vector signal analysis. Vector signal analyzer
Signal generation might utilize the software
N8241A AWG module (for maximum
Input
signal bandwidth and accuracy), the
signal
associated signal-creation software,
and the N8211A or N8212A upcon- 7.5 MHz
verter, depending on requirements
High dynamic range
for modulation, output power and Upconverter IF digitizer Radar signals
signal purity (Figure 19.6).
www.agilent.com/find/open
175
the SI hardware is too different.
For example, software can emulate
a legacy instrument that has poor
noise floor; however, it can’t emulate
a legacy instrument that has a better
noise floor than the SI hardware.
Also, most GPIB instruments have
a unique set of timing, network and
bus issues that are very difficult to
reproduce. In other words, SIs can
emulate legacy instruments, but no
emulation will be a perfect duplica-
tion of the original.
Conclusion
The basic premise of synthetic
instruments is very appealing: they
let you configure and reconfigure
building-block modules to create the
functionality of multiple measure-
ment devices. With benefits such
as smaller test systems, easier
transport, single-module updates or
upgrades, long-lived I/O and simpler
software updates, LXI-based SIs
readily support the DoD’s vision for
NxTest.
Overview
The three chapters in this section
explore some of the unique chal-
lenges of automating RF/microwave
tests, particularly as devices and test
requirements become increasingly
complex.
177
178 Section 4. RF/Microwave Test Systems
20. Optimizing the Elements of an RF/Microwave
Test System
Introduction Letting the DUT define The ability to meet the needs of
either long- or short-lived DUTs
Whether you need to test the latest “future” improves when the test system
cell phone, a next-generation military includes long-lived hardware, input/
When discussing the future-proofing
radio or an advanced radar system, output (I/O) and software. Careful
of a test system, it’s important to
proving the device’s ability to meet selection of these three elements
clarify what “future” means within
customer requirements depends will enhance a system’s flexibility
the context of the DUT and its
on a test system that can provide and its ability to perform accurate,
expected lifetime. For RF/microwave
accurate, repeatable results. For both repeatable measurements of multiple
test systems, there are two large
parametric and functional testing, DUTs and applications—today and
classes of DUTs that have specific
the ability to achieve accuracy and tomorrow.
future requirements.
repeatability becomes more difficult
as devices become more complex. • Long-lived DUTs. Many devices and
Greater complexity often translates systems developed for aerospace
into more tests, which may mean and defense applications require
longer development time and a test systems that are easy to
more complicated test system. The maintain and update far into the
challenge grows when you try to future. An example of this is the
create a system that meets budget NxTest program from the U.S.
and schedule constraints but is also Department of Defense (DoD).
flexible enough to meet both current Its guiding vision combines a
and future testing needs. common hardware architecture
with software-driven functionality
This chapter offers ideas and
to enable rapid deployment across
suggestions that can help you create
different programs and facilitate
flexible, long-lived RF/microwave test
easy updates in the future.
systems that will provide accurate,
repeatable assessments of the device • Short-lived DUTs. Fast-cycle
under test (DUT). Our focus is on aerospace/defense programs
making it easier for you to configure, and rapidly evolving commercial
update and modify your systems now wireless products require test
and in the future. systems that can be developed
rapidly and within budget. For
example, creating a new test
system from scratch for every new
phone model—or new wireless stan-
dard—becomes less desirable as
introduction cycles become shorter
and budgets get tighter. The ability
to leverage existing investments in
test equipment and software will
accelerate system development and
deployment while also reducing
system cost.
179
Reviewing some essential Specific attributes Constraints on the test system
The electrical attributes of the DUT
considerations often drive the shortlist of viable A combination of business and
instrumentation candidates. Most technical factors will also influence
When it’s time to define and assemble
DUTs contain a mix of circuitry system decisions. On the business
an RF/microwave test system, two
that is becoming less analog and side, budget and timeline are often
major factors will affect your deci-
more digital while going higher the primary drivers of tradeoffs
sions about test equipment: the key
in frequency with every genera- when selecting test equipment.
attributes of the DUT and the various
tion. On the analog side, operating At one extreme, for example, you
constraints on the test system. A
parameters such as frequency range, might need to get the system up and
quick review of important attributes
bandwidth and resolution—along running as quickly as possible—and
and constraints will lay a foundation
with headroom for today’s harmonics the ideal solution may be a one-box
for the discussions that follow.
or tomorrow’s enhancements—define tester, which trades rapid develop-
the essential specifications for signal ment time and optimized measure-
Key attributes of the DUT analyzers, signal generators, oscil- ments for decreased flexibility. At
loscopes and so on. The availability the other extreme, your contract
The attributes of your DUT obviously of test equipment with the necessary may require compliance with NxTest,
affect test-system design, and it’s performance or capabilities will have which specifies the use of modular
helpful to look at them from both a strong influence on the design of synthetic instruments—an approach
general and specific perspectives. your system. that yields tremendous flexibility but
at the expense of development time.
General attributes Greater digital content makes it
At a high level, it’s helpful to possible for new devices to support Within those constraints and
consider the DUT’s complexity, its multiple communication standards. tradeoffs, numerous expectations
stage in the product lifecycle, and the This might be CDMA, TDMA and GSM are placed on the test system.
nature of the manufacturing process. in a cell phone or various protocols These include its capabilities and
For example, multi-function devices in the military’s Joint Tactical performance: inputs, outputs
are often the most difficult to test: Radio System. The need to support and switching; measurement and
cell phones with built-in cameras, all relevant standards will demand analysis; speed, accuracy and
military radios that carry voice much greater flexibility from the repeatability; and data handling and
and data, and LAN devices with test system—and perhaps lead to the reporting. There are also expecta-
both wired and Wi-Fi capabilities use of instrumentation that also has tions about cost effectiveness, which
may require a much wider range of greater digital content in the form of may suggest the use of hardware
measurements and a more costly and advanced digital signal processing elements that are easy to reconfigure
complex test system. (DSP) capabilities. or replace and software that is easy
to modify or reuse.
Whether a product is simple or The physical configuration of the
complex, the early stages of its DUT will also affect choices about Expectations about system longevity
lifecycle generally require thorough handling, fixturing, switching, power, follow from both the length of time
testing of numerous characteris- loads and test accessories. As an the DUT will be manufactured and its
tics—parametric and functional—to example, the number and kind of estimated service life. Those require-
ensure expected performance and ports available for external connec- ments define how long the test
operation. As a product matures, tions may change as the device moves system itself must also be supported
fewer characteristics are tested, and through the manufacturing process. and maintained.
often in less detail. Once the circuitry is loaded into its
enclosure, any built-in test points
Within the manufacturing process, may become inaccessible and the
the product volume and mix also test interface may have to shift from
affect equipment choices. The most hard-wired to antenna-based.
difficult case is high volume and high
mix, which might require several
identical test systems that are able
to measure multiple products or
product variations.
Figure 20.1. A complex multi-tone test system implemented with conventional analog instruments
Amp
Signal Spectrum
generator LPF
analyzer
Combiner
Signal
PC generator LPF DUT
Signal
generator LPF
www.agilent.com/find/open
181
Example 2: Next-generation and its hardware and support costs. Example 3: Modern vector
modular instruments The DoD and others believe the
instruments
building-block approach offers the
Compared to the conventional greatest potential for enabling longer- As shown in Figure 20.3, the use of
approach, this type of system lived test systems. modern vector instruments produces
requires a somewhat less complex the simplest system, requiring just
arrangement of hardware (Figure Disadvantages one vector signal generator and
20.2) that includes four building- Initially, this architecture will require one vector signal analyzer. The PC
block modules: an arbitrary a significant investment in software does more than serve as host and
waveform generator (AWG), an development. The main reason is controller: it also adds functionality
upconverter, a downconverter and a the need to understand, define and via the Agilent Signal Studio soft-
high-speed digitizer. The PC provides create the individual measurement ware, which makes it easy to create
system-control functions that algorithms and analysis func- the required multi-tone signal and
arrange and rearrange the building tions that will utilize data from download it into the vector signal
blocks as needed to send or measure the hardware modules. (This is in generator.
a variety of signals. The PC also runs sharp contrast to a fully integrated
user-written software that provides instrument that has a vendor’s Advantages
system functionality, ranging from measurement expertise built into The tight integration of analog and
calibration to measurement algo- its firmware.) As a result, software DSP technologies delivers excep-
rithms to data analysis. development costs will tend to be tional versatility and functionality.
higher for this type of system. Comparing this system to the conven-
Advantages tional approach, one vector signal
Another key issue is measurement
The modular approach provides the generator replaces three analog
accuracy. Because manufacturers
ultimate in flexibility, enabling a signal generators and seven external
cannot anticipate every possible
high level of hardware reusability accessories. On the measurement
combination of modules, developers
and making it easy to rearrange the side, some vector signal analyzers
will have to create routines that, for
building blocks to create function- also provide waveform analysis capa-
example, calibrate every on-the-fly
ality that is equivalent to multiple bilities, possibly replacing a separate
rearrangement of the modules.
instruments. For example, because digitizer or oscilloscope. These
Consequently, traceability may be an
the AWG can generate virtually any flexible instruments can also be
issue for the earliest systems built on
type of signal, this configuration used for a variety of measurements,
this foundation.
can handle much more than just the not just the multi-tone example. In
multi-tone test. a system, fewer instruments mean
fewer connections, less complexity
Modular hardware also offers Over time, Agilent expects to provide and fewer opportunities to introduce
the possibility of obtaining better a broad and deep set of software
measurement errors.
performance by simply replacing an tools to accompany its building-block
hardware modules. Possible software
outdated module with a new, higher- Vector instruments can also provide
tools include individual measurement
performance building block. What’s better longevity: because they are
routines (e.g., group delay, VSWR),
more, this approach can also elimi- complete measurement modules (e.g., firmware-based, it is easy to enhance
nate redundant hardware elements, spectrum analysis) and even legacy their functionality and add new capa-
which may reduce a system’s size instrument emulation modules. bilities. Because so much of their
Figure 20.2. The multi-tone test system implemented with LAN-based building-block instruments
LAN
PC
Upconverter Downconverter
AWG Digitizer
DUT
www.agilent.com/find/open
183
Pulling it all together test system, we suggest you consider lived or short-lived, (2) the character-
the use of vector instruments, LAN- istics of your DUTs—including both
Table 20.1 compares analog, modular based I/O and instrument drivers. general factors such as stage in the
and vector instruments based on five This combination will provide a product lifecycle and manufacturing
essential aspects that affect system highly future-proof system that volume and specific factors such as
performance: measurement capabili- should be easy—and cost-effective—to degree of digital content and physical
ties, measurement performance, I/O modify in the near-term, maintain configuration—and (3) any financial
connectivity, system software (and and update in the future. If you are and technical constraints on your
instrument communication) and required to comply with NxTest, then test system.
potential longevity. Those elements substitute modular instruments for
capture the value of each approach, vector instruments in the preceding In addition, be sure to consider your
and that overall value provides recommendation. options for connectivity, software
a broader context for the sixth development tools and instrument
element, which is hardware cost. Conclusion communication.
Ultimately, the best answer will For new test systems, our baseline
Conventional analog instruments,
depend on the attributes of your DUT recommendation is a combination of
next-generation modular instruments
and the constraints on your system. vector instruments, LAN-based I/O,
and modern vector instruments
However, if you are creating a new graphical programming and instru-
each offer compelling benefits for
ment drivers. Modify this approach
RF/microwave test systems. Choosing
as needed, of course, but in general it
In many cases the lack of software the approach for your next system
will provide a high degree of future-
transportability will drive the cost of depends on a number of factors: (1)
proofing and the ability to modify,
developing new software far beyond whether your DUTs tend to be long-
maintain and update quickly and
the hardware cost.
cost effectively.
Measurement instruments
ment instruments (Figure 21.1). This & power supplies Performance
chapter suggests a framework for In RF and microwave test equipment,
Analog Digital Power
those tradeoffs and offers six sets of Agilent’s definition of “performance”
hints that address common prob- focuses on instrument accuracy,
Switching system
lems that may exist along RF signal measurement range and bandwidth.
pathways. Instrument accuracy includes the
specified absolute accuracy of
Interface & fixturing amplitude and frequency measure-
ments. Measurement range refers
to dynamic range, distortion, noise
level and phase noise, which are the
Device under test (DUT)
attributes that enable precise charac-
terization of signal levels. Bandwidth
refers to the frequency width or
data rate that can be processed and
analyzed.
185
Speed Repeatability measurement. It can be optimized by
Test system speed or throughput For any test system, the ability to minimizing changes to measurement
depends on hardware, input/output produce consistent results—test-to-test settings such as center frequency,
(I/O) and software. Our focus is on and day-to-day—is crucial. However, span and attenuation level.
the hardware and the four factors repeatability does not infer a high
that influence speed: measurement level of precision, which depends Summarizing the interrelationships
set-up time, measurement execution on the performance of individual The test requirements and business
time, data processing time and data instruments. Instead, repeatability drivers for a DUT will help you assess
transfer time. At RF and microwave means a consistent result, whatever the relative importance of perfor-
frequencies, a key aspect of set-up the specified accuracy. For any given mance, speed and repeatability. Once
time is the settling time of the DUT instrument, repeatability may be you’ve identified the dominant factor
or the test system whenever a change different for certain measurements and the intensity of its requirements,
is made (such as switch closures and or modes so it’s important to check sorting through the interactions and
power level). the product specifications or ask their impact on the system becomes
the manufacturer. To some extent, easier. Tables 21.1 through 21.3
repeatability can be improved with summarize the implications of these
more averaging or through modified interactions in two cases: when the
algorithms that produce an accurate intensity of the dominant factor is
approximation of a standardized either high or low.
Table 21.1. When performance dominates, the most important interaction is between performance and speed.
Table 21.2. When speed dominates, the key relationship is between speed and repeatability.
Table 21.3. When repeatability dominates, the key relationship is once again between repeatability and speed.
186 21. Six Hints for Enhancing Measurement Integrity in RF/Microwave Test Systems
Repeatability and performance Hint 2: Review the nature a wide frequency range. Hint: Use
In Tables 21.1 and 21.3 there is an high precision cables and acces-
important secondary relationship and behavior of the DUT sories, and fully characterize their
between repeatability and perfor- actual impedance using a vector
A typical automated test system
mance. This is an indirect relationship network analyzer (VNA), especially
performs three basic tasks: sourcing,
linked by measurement uncertainty. if the DUT is an active device.
measuring and switching. Decisions
When dealing with uncertainty, some
about which signal generators, power • Minimize VSWR. The combination of
system developers create an “error
meters, spectrum analyzers, network a switch matrix, its connectors, its
budget,” the size of which depends on
analyzers, switch matrices and cables internal and external cables, and
the margin between test requirements
to use depend on the electrical and even the bending radius of any RF
and system uncertainty. The two
mechanical attributes of the DUT. cables can induce errors caused
major contributors to uncertainty are
At RF and microwave frequencies, a by voltage standing waves in the
absolute accuracy (instrument perfor-
few essential characteristics require DUT. Hint: To minimize this error,
mance) and measurement consistency
special attention. use a switch matrix with a voltage
(repeatability). If the instruments in
a system have high absolute accuracy, standing wave ratio (VSWR)
then there is a wider margin in the Electrical parameters specification of 1.2:1 or better.
error budget for lower repeatability.
The basic nature of the DUT is a key • Enhance isolation. If your test
If the instruments provide consistent
consideration: Is it passive and linear requirements call for simulta-
results, that leaves more room in the
or active and nonlinear? Passive, neous measurements of high- and
budget for somewhat lower absolute
linear devices are easier to deal with low-level signals then the isolation
accuracy.
because they typically have fixed gain specifications of the switch matrix
Multiple “high” requirements and phase shift at any allowed input will affect measurement integrity.
Satisfying requirements such as “high power level across their operating Hint: If there are multiple pathways
speed with high repeatability” or bandwidth. In contrast, active through the DUT, use a signal
“high performance with high speed” devices demand greater care because generator and spectrum analyzer
will probably require sophisticated they usually have a nonlinear oper- to characterize the isolation
instrumentation that is somewhat ating region that is highly sensitive properties to the extent possible.
more expensive than less-capable to input power, producing different If this can’t be done then the
equipment. However, many high results at different levels. Within system should be configured and
performance instruments may a test system, this may suggest the programmed to route high- and
include hardware accelerators that addition of amplifiers or attenuators low-level signals on non-adjacent
speed up time-consuming operations to precisely control power levels, and pathways or through separate
such as averaging and calibration. perhaps the addition of couplers to switch units.
Some models may also include split off and verify the power level
multiple algorithms for calculating being delivered to the DUT. These Mechanical attributes
parameters such as adjacent channel additions should not be taken lightly:
power (ACP). At high frequencies, every system One set of details to consider is the
element has a complex-valued imped- number and type of connectors for
If all three requirements rate “high” ance (with associated S-parameters), signals and power (AC or DC). This
then every element of the system—test and every additional connection will influence factors such as the
equipment, switching, cabling, creates the possibility of undesirable required size of the switch matrix
connectors, and so on—must be scru- interactions with the DUT. and the complexity of system cabling.
tinized. The best solutions will likely Hint: Use a switch matrix with enough
demand a high price, but may provide • Avoid mismatches. An impedance ports to let you make all system-
additional capabilities and benefits. mismatch at any connection to-DUT connections just once. This
can cause insertion loss, which will minimize delays while waiting
robs power from any sourced or for signals to settle, and minimize
As an example, some Agilent PSA measured signal. As a truism, the chances of damaging the switch
series spectrum analyzers include power is expensive at high frequen- matrix or DUT with sudden changes
a standard “ACP mode” and a “fast cies—and it becomes more expen-
ACP mode.” The fast mode provides in power level.
sive if it has to be delivered across
an accurate approximation of the
standard-compliant measurement.
www.agilent.com/find/open
187
Hint 3: Understand, Correcting active pathways Dealing with DUT distance—
characterize and correct The performance of active devices
near or far
RF signal paths may vary with changing input power. Accurate correction can be difficult
The process required to improve whether the DUT is mounted in a
Without additional correction, measurement accuracy depends on fixture at the test system or located
product specifications extend only whether the device is operating in several meters away in a test
as far as the “calibration plane” that the linear or nonlinear portion of its chamber. Fixture-based measure-
exists at an instrument’s input and response. If an active device such as ments are challenging because
output connectors. To achieve accu- an amplifier is operating in its linear pathways often include transitions
rate, repeatable measurements—and region—well below its 1-dB compres- from coaxial cables to microstrip-
corrected DUT results—we suggest sion point—during both calibration based shorts, opens and loads. Hint: If
that you push the calibration plane and measurement operations, high quality microstrip elements are
out as close as possible to the DUT. then corrections can be accurately not available it will be necessary to
There are several ways to achieve applied at any power level within measure the fixture with a network
this, whether the pathways are that region. Hint: If the active device analyzer, model the impedance
passive or active and the DUTs are is operating in the nonlinear portion and remove those effects from the
local or remote. of its response then the power level measurements.
used for a measurement must also
be used during calibration to ensure When the DUT is remote, the main
Handling passive pathways accurate correction. If measurements issues are path attenuation in long
As just noted, passive devices have will be made at multiple power levels cable runs and path variation due to
fixed gain and phase shift at any in nonlinear mode, then individual temperature fluctuations and cable
allowed input power level across calibrations must be made at each of flexion. Hint: Characterize path atten-
their bandwidth. However, every those levels and stored for later use. uation either by measuring the entire
connection along a passive path may pathway between the instrument and
Hint: Check the frequency response of the DUT (if possible) or by measuring
have an impedance mismatch, which the active device over the frequency
will cause insertion loss and phase all elements along the path and
range of the DUT. Again, you should using vector math to combine their
shifts (or delays). At high frequencies either measure the entire path at
seemingly simple passive elements complex-valued responses.
specific power levels or characterize
become complex transmission-line the S-parameters of each interface
elements, precluding simple algebraic and use vector math to create a
addition of losses and phase shifts model that can applied after-the-fact
along the path. Hint: Use a VNA to or in real time.
either measure the entire connected
path or characterize the S-param- Hint: To simplify the process of
eters of each element and use vector characterizing and correcting RF
math to model the total loss and signal paths, some system developers
phase shift of the entire path. These minimize the use of active devices.
values can be stored in the system PC This reduces both the calibration
and applied as needed to correct a effort and the chance of errors
measurement, or they can be applied caused by variations in power level
by a network analyzer, for example, when operating in nonlinear mode.
to enable real-time adjustment of
filters and other variable DUTs.
188 21. Six Hints for Enhancing Measurement Integrity in RF/Microwave Test Systems
Hint 4: Be aware of Semi-rigid Avoiding switch-related
As suggested by the name, these problems
everything connected to cables do not easily change shape,
an instrument ensuring excellent performance Switching is central to overall
and repeatability. High quality system functionality, automating
Test equipment manufacturers semi-rigid cables achieve additional the connection of signals and power
specify the performance of every stability during the manufacturing supplies between instrumentation
instrument up to the front-panel process through techniques such as and the DUT. Because most sourced
connectors that source and measure MIL-standard temperature cycling. and measured signals pass through
signals. From there, everything When applied after the forming the switch matrix, any shortcom-
that comes between the instrument process, temperature cycling can ings in its specifications can affect
and the DUT can affect instrument eliminate internal stresses that measurement performance, speed
performance and measurement may cause later deformation of the and repeatability. At high frequen-
repeatability. At RF and microwave preformed cable. cies, three specifications are particu-
frequencies and power levels, the larly important: isolation, VSWR and
three worst offenders are typi- The quality of the dielectric used insertion loss.3
cally cables, switches and signal in these cables also affects their
conditioners. measurement performance. Solid • Maximize isolation. Leakage between
PTFE on is the most common signal paths can make it very
but contributes to insertion loss. difficult to measure low-power
Selecting the right type of cable Expanded PTFE is currently the best signals in the presence of one or
alternative, providing lower insertion more powerful signals. (This is
When specifying a test system you
loss and wider frequency range. All most likely to occur when high-
will need to decide what type of
of this attention to detail is reflected and low-power signals are routed
cabling to use for device interconnec-
in the cost of these cables, which is through a switch matrix simultane-
tion, and you may be able to specify
considerably higher than conform- ously.) Hint: Choose a switch with
the type used within the switch
able or flexible cabling. isolation specifications of 90 dB
matrix. As a general rule, a stable
cable will provide lower insertion or better. This will reduce leakage
Conformable and potentially minimize the need
loss, better VSWR and, therefore, These cables offer less stability
greater measurement repeatability. to route signals through physically
than semi-rigid cables because they separate switch assemblies.
At high frequencies, the three most are easily shaped and reshaped.
commonly used types of cabling are Their flexibility affects measure- • Minimize VSWR. High VSWR can
semi-rigid, conformable and flexible. ment repeatability and long-term cause phase errors and therefore
reliability. affect the accuracy of vector and
modulation measurements.4 VSWR
Flexible in a switch matrix is directly
Sometimes called “instrument-grade related to the VSWR of the coaxial
cables,” these typically offer good switches used within the matrix,
phase stability and low insertion loss and the VSWR of an individual
but at a relatively high price. They switch depends on its mechanical
also tend to be high maintenance, dimensions and tolerances.
requiring careful handling because
severe deformation can alter their
electrical properties and cause inac- 3 For detailed information, please see
curate measurement results. the Agilent Custom Switch Matrices
product note, publication number
5966-2961.
4 Phase repeatability is another impor-
tant specification to consider when
making these measurements.
www.agilent.com/find/open
189
Hint: You can further minimize Evaluating signal conditioners Attenuators
VSWR by using cables that are Electromechanical and electronic
short compared to the required As described in Hint 3, the DUT, its designs provide different levels of
bandwidth. If this is not practical test requirements and its location flexibility and precision in managing
because of wide bandwidth or may dictate the insertion of passive signal levels. Electromechanical
mechanical requirements, the best or active signal conditioners into the attenuators use discrete switches
alternative is to add insertion loss signal paths. These can be standalone that typically provide stepped
to the transmission lines via pads devices or may be built into the resolution of 1 or 10 dB. Electronic
or lossy cables. This will reduce switch matrix. Amplifiers, attenua- attenuators provide virtually
the amplitude of VSWR-induced tors and frequency converters are continuous settings with 0.1 or 0.25
ripples over the frequency range the most commonly used signal dB resolution; however, those that
of interest, but at the expense of conditioning devices. use PIN diode-type switches can
higher overall insertion loss. produce “video leakage” spikes that
Amplifiers may contaminate measurement
• Manage insertion loss. This tends A signal might need additional gain results. Hint: Cascade electrome-
to become a problem at higher if a precise amplitude measurement chanical and electronic attenuators
frequencies and is typically speci- is required or if it is being sent over as needed to provide greater control
fied versus frequency in tabular a long cable run. Several key speci- of attenuation.
or equation form. Hint: As a switch fications will help you determine
ages, its insertion loss may change an amplifier’s suitability for your Hint: Pay attention to the plating
so look for specifications such application. material used on attenuator connec-
as “insertion loss repeatability” tors. As an example, nickel becomes
or “insertion loss stability” that • VSWR. Amplifiers are notorious for nonlinear at high power levels and
are valid through the end of having poor VSWR. Hint: Alleviate will cause intermodulation distor-
the product’s expected lifetime. VSWR problems by connecting an tion. Instead, choose a higher quality
Knowing this type of worst-case attenuator or an isolator (though conductor such as gold.
value can help you manage your these have limited bandwidth) to
error budget. the amplifier output. Frequency converters
When the DUT is remote from the
• Intermodulation. Amplifier test system, you can reduce inser-
bandwidth is important when tion loss in long cable runs by using
measuring intermodulation distor- a downconverter to shift signals to
tion or spurious signals outside the a lower frequency range. Hint: At
bandwidth of the DUT. Hint: Beware the test system, upconversion can
of amplifiers with poor dynamic be used to restore the signal to its
range or a low 1-dB compression original frequency, but it may be
point, which can produce enough necessary to also apply filtering to
intermodulation distortion to remove unwanted frequency compo-
affect harmonic measurements nents created during the conversion
in the presence of a strong processes.
fundamental.
Hint: If multiple signals, paths or
• Spurs. Switching power supplies conversions are used when making
may cause spurs that are related vector or modulation measurements,
to the switching frequency, which some form of phase locking must be
is typically 100-200 kHz. Hint: used to ensure accurate results. You
Avoid using amplifiers or any other can do this by connecting the instru-
devices that contain switching ments and frequency converters to
power supplies. a common frequency reference and
then measuring the phase of each
signal relative to the reference signal.
190 21. Six Hints for Enhancing Measurement Integrity in RF/Microwave Test Systems
Hint 5: Examine the Latching versus non-latching Advanced features: Built-in
operational attributes of Internally, electromechanical
signal conditioning
switches switches use either latching or One advantage of having a switch
non-latching relays. Most latching matrix in a system is that signal
When deciding what type of tech- types need a 100-200 msec pulse conditioning can be built into the
nology to use in a switch matrix, of DC power to open or close the matrix by the manufacturer. As an
it can be helpful to go beyond the relay. Non-latching switches require example, Agilent’s custom switch
electrical performance and look at constant powe—typically 24 V at 200 matrices can be configured with a
operational attributes such as device mA—to maintain contact. In a large variety of devices, including ampli-
longevity, power requirements and switch matrix non-latching switches fiers and attenuators; filters and
fail-safe operation. can generate enough heat within a isolators; and phase- and frequency-
system rack to affect measurement translating devices such as mixers,
performance. Hint: If you choose to
Electromechanical versus use non-latching switches, check
doublers, and dividers. These devices
are permanently connected with
electronic the actual heat rise and be prepared semirigid coaxial cables and no
to include additional cooling in the additional external cabling is needed.
With numerous moving parts and system rack.
physical contacts, electromechanical The result is a compact, convenient,
one-box solution.
switches tend to suffer from rela- Hint: It’s essential to know how
tively rapid degradation, declining either type of switch will behave
repeatability and limited life. In after a power failure or emergency
contrast, electronic switches have no shutdown. For maximum safety,
moving parts so offer longer life and select a switch matrix that returns to
greater repeatability. In practice, the a known condition or configuration
best choice depends in part on the when power is restored. Non-latching
actual number of switching cycles switches are often the default failsafe
a system will require; consider the choice because they open when
number of closures per test, the power is removed and won’t close
number of tests per day, the expected again until power is applied by the
lifetime of the system and so on. test program. However, latching
switches can be made fail-safe if they
Another practical consideration is include hardware and firmware that
the power level of the routed signals. will latch them into a safe mode at
Switching of high power signals will power down.
damage most switches, lowering
repeatability and shortening lifetime.
Hint: To prevent the premature
demise of either electromechanical
Another hint: To minimize power
or electronic switches, program the
requirements, some developers
system instrumentation to reduce program the system to actuate these
signal levels before opening or switches serially or in small batches,
closing any switches in the matrix. though this causes longer total
switching time.
www.agilent.com/find/open
191
Hint 6: Accelerate states as needed. If the system needs This method also reduces wear on
to load arbitrary waveform data test-port connectors and calibration
measurement set up during a test, download the minimum standards.
and execution number of points and use binary
Hint: The application of correc-
format rather than ASCII.
Whether you gauge system perfor- tion data is usually faster when
mance as devices tested per unit Power meters performed inside the analyzer
of time, tests per unit of time or The biggest potential time savings rather than externally in the system
another time-based metric, measure- come from models that offer built-in controller. With most VNAs you
ment speed depends on two essential calibration capabilities that extend can save the calibration curve for a
factors: the time required to set up the cal interval from hours to specific test and recall it as needed.
the system and the time required to months. Hint: Use digitizing power Note that this method is more
perform the measurement. The three meters that offer wide video band- effective when used over a series
major elements of any system— hard- widths and fast data sampling. Some of somewhat narrow frequency
ware, I/O and software—can help or of these units can generate 1,000 or spans than with one extremely wide
hinder both processes. Chapter 7, more corrected readings per second, measurement span.
Maximizing System Throughput improving measurement accuracy
and Optimizing System Deployment, and repeatability through averaging.
offers several useful tips about
software design, system I/O and Spectrum analyzers Conclusion
low-frequency instrumentation. With any spectrum analyzer, the
To complement that material, this three key adjustments are frequency Every test system faces a unique
hint adds new information specific span, points per measurement and set of challenges, but in all cases
to RF/microwave instruments and resolution bandwidth (RBW). Hints: the ability to manage the direct and
systems. Using the fewest necessary points indirect tradeoffs among perfor-
and the widest possible RBW is the mance, speed and repeatability
easiest way to reduce measurement will help you achieve the required
Fine tuning individual time. Utilize a current-generation level of measurement integrity. The
instruments spectrum analyzer that automatically ability to manage crucial tradeoffs
speeds things up by switching to Fast also applies to the selection of
Any configurable device used in a Fourier Transform (FFT) mode when instrumentation, I/O connections
system can become a bottleneck measuring narrow spans. and software elements for your test
that limits measurement speed. The system. Agilent is helping reduce the
latest generations of RF/microwave Hint: To gain maximum benefit, use number of compromises you have
instruments—signal generators, automatic input ranging selectively. to make by offering system-ready
power meters, spectrum analyzers When used to measure signals of instrumentation, PC-standard I/O
and network analyzers—offer flexible rapidly varying amplitude, auto and open software environments.
features and capabilities that can ranging may frequently change the By creating complementary system
minimize bottlenecks and enhance input attenuator settings and slow elements and supporting continually
system performance. the measurement. However, if signal advancing standards such as LXI,
levels are low and relatively constant, Agilent can help you optimize—and
Signal generators auto ranging can improve the even maximize—system performance
Many of these are available with signal-to-noise ratio and also shorten now and in the future.
built-in modulation and arbitrary measurement time by allowing use of
waveform capabilities, potentially wider span and RBW settings.
reducing the number of instruments
in a system, simplifying system Network analyzers
cabling and lessening software Calibration of VNAs can be very time
complexity. Hints: Instrument configu- consuming, especially the manual
ration may be somewhat complex connection of shorts and standards.
and time consuming, but you can Hint: Agilent’s line of electronic cali-
significantly reduce test time by bration or “ECal” modules automates
creating states ahead of time, saving this process, offering faster and more
them in memory and then program- repeatable calibrations on one to four
ming the system to recall the saved ports through a single connection.
192 21. Six Hints for Enhancing Measurement Integrity in RF/Microwave Test Systems
22. Calibrating Signal Paths in RF/Microwave
Test Systems
Introduction with a network analyzer while the This method requires measurements
scalar measurements are typically of both the magnitude and phase
In any RF test system the ability to performed with a signal generator characteristics of the RF path. This
achieve instrument-port accuracy and a power meter or spectrum can be done by either performing
at the device under test (DUT) will analyzer. a network analyzer calibration at
enhance measurement accuracy the DUT’s input and output ports,
and repeatability. Unfortunately, This chapter provides an overview or by using a calibrated network
the non-ideal nature of the cables, of three approaches that can be analyzer to measure the S-param-
components and switches in the used to calibrate RF signal paths eters of an RF path (see sidebar). The
paths between the instruments and and produce accurate, repeatable latter method provides a complete,
the DUT can degrade measurement measurements. It’s important to note complex-valued characterization of
accuracy. Vector or scalar calibration that these calibrations are a comple- the signal path.
is usually required to characterize ment to—not a substitute for—the
and correct for this loss of accuracy. calibration of individual instruments
within a system. Scalar calibration
The proper calibration method
This approach characterizes only
depends on both the type of Understanding vector and the magnitude characteristic of
measurement and the signal path.
For example, measurements of gain scalar calibration the RF path, which is equivalent
to measuring only the magnitude
and phase require complex-valued
Within any test system, common portion of the S21 transmission
vector calibration, which is typically
elements such as fixturing, switching coefficient in a vector calibration.
performed with a network analyzer.
and cabling will introduce offsets and A common technique involves
As another example, measurements
errors that will affect measurement driving one end of the path with a
of power levels and frequency
accuracy. The two types of calibra- signal generator and measuring the
content may be vector measurements
tions used to account for and correct signal at the other end with a power
of modulated signals (accurate
these errors are vector and scalar meter. The magnitude portion of
phase information is essential) or
calibration. the path response is determined by
scalar measurements of continuous
subtracting the source power level
wave (CW) signals. In these cases,
vector measurements are performed Vector calibration
193
setting (in dBm) from the measured Table 22.1. Comparing the advantages and disadvantages of vector and scalar calibration.
power level (also in dBm). This
is repeated at multiple frequen- Calibration Advantages Disadvantages
cies across the band of interest to type
determine the overall magnitude Vector • Enables complete characteriza- • Higher cost than scalar because
characteristic. tion of the path and, therefore, network analyzer is required
more accurate measurements • Doesn’t account for inaccura-
Scalar calibrations can achieve very • Allows adapter embedding and cies internal to instruments
good results as long as high quality de-embedding connected to the signal path
components, adapters and cables are • Provides excellent confidence in
used in the system. This helps mini- path integrity
mize measurement uncertainty and
Scalar • Lower cost approach (network • Not a complete characterization
increase measurement repeatability.
analyzer not required) of the path (magnitude only)
However, when compared to a full
• Can compensate for inaccura- • Doesn’t support adapter embed-
vector calibration, scalar calibration
cies internal to instruments ding or de-embedding
is less likely to detect any changes in
connected to the path, which • Provides less confidence in path
impedance match along a signal path.
may result in better overall integrity
accuracy
Comparing the two methods
The best choice of calibration method
depends on factors such as the test Figure 22.2. The essential elements of a simplified RF/microwave test system
specification and its measurement
and accuracy requirements, the
likelihood of inaccuracies internal to Signal Network Spectrum
generator analyzer analyzer
the measurement instruments, and
the availability of a network analyzer.
The advantages and disadvantages
of each method are summarized in
Table 22.1. RF switching
Performing vector
calibration of network-
www.agilent.com/find/open
195
network-analyzer paths additional benefits come from data after a system calibration is
network-analyzer characterization of completed. This makes it possible to
Non-network-analyzer paths connect the system paths used for modulated account for separately characterized
instruments other than a network DUT measurements. One is greater adapters such as test fixtures or
analyzer to the DUT. The measured confidence in path integrity, which circuit boards that interface to the
signals may be either modulated comes from the ability to easily DUT. Combining these elements with
or CW. measure characteristics such as the existing path data requires that all
return loss of the path (S11 and S22). S-parameters be known for both the
Vector calibration of these paths This allows for a more comprehen- adapter and the path.
is accomplished by connecting a sive self-test of the system and helps
calibrated network analyzer to the
path and measuring its S-parameters.
minimize the uncertainties caused by Performing scalar
input and output mismatches.
Prior to measuring the RF path, the
network analyzer is calibrated in a The other noteworthy advantage
standalone configuration with special is the ability to modify the path
calibration cables. The results of
these path-calibration measurements
are stored in the system controller Figure 22.5. When calibrating paths such as signal-generator-to-DUT-input,
for later recall and application. adapter de-embedding moves the reference plane away from the network
analyzer, ensuring characterization of just the signal path.
RF switching
Feed-through
www.agilent.com/find/open
197
You can do this by (1) using a Conclusion
feed-through to connect the DUT
input cable directly to the DUT The use of vector and scalar calibra-
output cable, (2) setting up the signal tion can increase measurement accu-
generator to output the required racy by helping you correct for errors
range of frequencies and power levels in the RF signal paths. Each method
and (3) making power measurements has advantages and disadvantages,
with the spectrum analyzer. and the choice depends on both the
type of measurements you’re making
The spectrum analyzer should be and the nature of the signal path.
configured just as it will be for DUT
measurements. This is especially This chapter reviewed three different
true of the input attenuator settings, to methods of characterizing the
which often cause wide variations RF path: vector calibration of
in the spectrum analyzer’s input network-analyzer paths, vector
impedance. The resulting calculated calibration of non-network-analyzer
offsets will account for path loss as paths and scalar calibration of
well as some inaccuracies internal to non-network-analyzer paths. When
the spectrum analyzer. performing vector calibration of
network-analyzer paths, the tech-
Note a key assumption here as well: nique of adapter embedding ensures
The accuracy of this calibration characterization of only the signal
depends on the impedance of the paths of interest. Adapter de-embed-
DUT output cable being very similar ding provides the same benefits
to the input impedance of the power when you’re performing vector
meter. It is important to verify these calibration of non-network-analyzer
impedances because a large differ- paths. Within the context of your
ence will cause significant measure- specific measurement needs, each
ment errors. method provides valuable calibration
benefits.
Measuring adapter effects
Accounting for adapters necessary
to perform scalar-path calibrations is
usually accomplished by estimating
or measuring adapter loss at various
frequencies of interest and then
accounting for those losses in the
offset calculations. However, this is
much less accurate than the adapter
embedding and de-embedding
procedures described in the vector
calibration sections.
Adapter — the LAN card and connector Firewall — a hardware device or software Router — a LAN device that joins multiple
that provides an electrical interface to the program (or combination) that protects networks and enables creation of small,
network a computer network from unauthorized private networks
access
ATE — Automated test equipment SI — Synthetic instrumentation
Gateway — a hardware device that
ATS — Automated test system connects devices that use different stan- SIWG — Synthetic Instruments Working
dards and protocols (e.g., LAN to GPIB) Group
AWG — Arbitrary waveform generator
GPIB — General Purpose Interface Bus; SPX — Sequenced Packet eXchange; a
Bridge — a LAN device that connects communication protocol used in the Novell
the dominant 8-bit parallel I/O connection
segments of a network Netware network operating system
for test equipment and test systems
CASS — Consolidated Automatic Support Subnet — a group of connected network
Hub — a multi-port LAN device that
System devices; used to partition networks into
connects multiple devices together, usually
in a star topology segments for easier administration
COTS — Commercial off-the-shelf
ICS — Internet connection sharing Subnet mask — a setting that accom-
DDNS — dynamic domain name server;
panies an IP address and defines the
a service that allows a network device to
IF — Intermediate frequency boundaries of a subnet
establish its host name when it connects
to the network. This lets other devices IP— Internet protocol; requires an address Switch — a LAN device that connects
use that host name with DNS to find the to communicate multiple devices to a single LAN line;
device’s IP address and connect to it. however, unlike a hub, it preserves full
IPX — Internetwork Packet eXchange; a network bandwidth to each device
DHCP — dynamic host configuration communication protocol used in the Novell
protocol; a method of automatically Netware network operating system TCP/IP — Transfer Control Protocol and
obtaining an IP address for a LAN- Internet Protocol; the two standards that
connected device (e.g., PC, router, LAN — local area network provide the data communication foundation
instrument, etc.) of the Internet
LXI — LAN eXtensions for Instrumentation
DMZ — De-militarized zone; a firewall Technology insertion — The introduction
configuration that helps secure the private MAC — media access control; every LAN
of new or improved hardware or software
LAN device has a unique MAC address
capabilities into an existing system
DNS — domain name server; maps NAT — network address translation; maps
TPS —Test program set
specific names to IP addresses, enabling private addresses to one or more public
use of names in place of IP addresses in addresses to enable access to an intranet USB — Universal Serial Bus; designed to
testprograms or the Internet replace the RS-232 and RS-422 serial buses
used in PCs
DoD — United States Department of NetBEUI — NetBios Extended User
Defense Interface; a network communication VME or VMEbus — Versa Module
protocol used in many versions of Windows Eurocard
DUT — device under test; the component,
subassembly or product to be measured by NxTest — Next-generation Automatic Test VXI — VME eXtensions for Instrumentation
the test system Systems
199
Agilent Email Updates
Remove all doubt www.agilent.com
Our repair and calibration services For more information on Agilent
www.agilent.com/find/emailupdates
will get your equipment back to you, Technologies’ products, applications or
Get the latest information on the products
performing like new, when promised. services, please contact your local Agilent
and applications you select.
You will get full value out of your Agilent office. The complete list is available at:
equipment throughout its lifetime. Your www.agilent.com/find/contactus
Agilent Direct equipment will be serviced by Agilent-
trained technicians using the latest
www.agilent.com/find/agilentdirect
factory calibration procedures, auto- Americas
Quickly choose and use your test
mated repair diagnostics and genuine
equipment solutions with confidence. Canada 877 894 4414
parts. You will always have the utmost
Latin America 305 269 7500
Agilent confidence in your measurements.
United States 800 829 4444
Open
Agilent offers a wide range of additional
www.agilent.com/find/open expert test and measurement services Asia Pacific
Agilent Open simplifies the process of for your equipment, including initial Australia 1 800 629 485
connecting and programming test systems start-up assistance onsite education China 800 810 0189
to help engineers design, validate and and training, as well as design, system Hong Kong 800 938 693
manufacture electronic products. Agilent integration, and project management. India 1 800 112 929
offers open connectivity for a broad range Japan 81 426 56 7832
of system-ready instruments, open industry For more information on repair and Korea 080 769 0800
software, PC-standard I/O and global calibration services, go to Malaysia 1 800 888 848
support, which are combined to more Singapore 1 800 375 8100
easily integrate test system development. www.agilent.com/find/removealldoubt Taiwan 0800 047 866
Thailand 1 800 226 008
Europe
www.lxistandard.org
Austria 0820 87 44 11
LXI is the LAN-based successor to
Belgium 32 (0) 2 404 93 40
GPIB, providing faster, more efficient
Denmark 45 70 13 15 15
connectivity. Agilent is a founding
Finland 358 (0) 10 855 2100
member of the LXI consortium.
France 0825 010 700
Germany 01805 24 6333*
*0.14€/minute
Ireland 1890 924 204
Italy 39 02 92 60 8484
Netherlands 31 (0) 20 547 2111
Spain 34 (91) 631 3300
Sweden 0200-88 22 55
Switzerland (French) 41 (21) 8113811 (Opt 2)
Switzerland (German) 0800 80 53 53 (Opt 1)
United Kingdom 44 (0) 118 9276201
Other European Countries:
www.agilent.com/find/contactus
Revised: May 7, 2007