Está en la página 1de 169

THE HANDBOOK OF

MANUFACTURING
ENGINEERING
Second Edition

Assembly
Processes
Finishing, Packaging,
and Automation
DK565X_000_r04.indd 2 11/16/05 10:28:02 AM
THE HANDBOOK OF
MANUFACTURING
ENGINEERING
Second Edition

Assembly
Processes
Finishing, Packaging,
and Automation

EDITED BY

Richard Crowson

Boca Raton London New York

A CRC title, part of the Taylor & Francis imprint, a member of the
Taylor & Francis Group, the academic division of T&F Informa plc.
DK565X_Discl.fm Page 1 Monday, August 8, 2005 11:46 AM

Published in 2006 by
Taylor & Francis Group
270 Madison Avenue
New York, NY 10016
© 2006 by Taylor & Francis Group, LLC
No claim to original U.S. Government works
Printed in the United States of America on acid-free paper
10 9 8 7 6 5 4 3 2 1
International Standard Book Number-10: 0-8493-5565-6 (Hardcover)
International Standard Book Number-13: 978-0-8493-5565-3 (Hardcover)
Library of Congress Card Number 2005020353
This book contains information obtained from authentic and highly regarded sources. Reprinted material is
quoted with permission, and sources are indicated. A wide variety of references are listed. Reasonable efforts
have been made to publish reliable data and information, but the author and the publisher cannot assume
responsibility for the validity of all materials or for the consequences of their use.
No part of this book may be reprinted, reproduced, transmitted, or utilized in any form by any electronic,
mechanical, or other means, now known or hereafter invented, including photocopying, microfilming, and
recording, or in any information storage or retrieval system, without written permission from the publishers.
For permission to photocopy or use material electronically from this work, please access www.copyright.com
(http://www.copyright.com/) or contact the Copyright Clearance Center, Inc. (CCC) 222 Rosewood Drive,
Danvers, MA 01923, 978-750-8400. CCC is a not-for-profit organization that provides licenses and registration
for a variety of users. For organizations that have been granted a photocopy license by the CCC, a separate
system of payment has been arranged.
Trademark Notice: Product or corporate names may be trademarks or registered trademarks, and are used only
for identification and explanation without intent to infringe.

Library of Congress Cataloging-in-Publication Data

Crowson, Richard.
Assembly processes : finishing, packaging, and automation / Richard Crowson and Jack Walker.
p. cm.
Includes bibliographical references and index.
ISBN-13: 978-0-8493-5565-3 (alk. paper)
1. Assembly-line methods. 2. Production planning. I. Walker, Jack M., 1924- II. Title.

TS178.4.C76 2005
670.42'7--dc22 2005020353

Visit the Taylor & Francis Web site at


http://www.taylorandfrancis.com
Taylor & Francis Group
is the Academic Division of T&F Informa plc.
Preface
Handbooks are generally considered to be concise references for specific subjects.
Today’s fast-paced manufacturing culture demands that such reference books provide
the reader with how-to information with no frills. Some use handbooks to impart
buzzwords on a particular technical subject that will allow the uninitiated to gain cred-
ibility when discussing a technical situation with more experienced practitioners.
The second edition of Handbook of Manufacturing Engineering was written to
equip executives, manufacturing professionals, and shop personnel with enough infor-
mation to function at a certain level on a variety of subjects. This level is determined
by the reader.
The final book, Volume IV, deals with the finishing of the product. Packaging
and automation are also discussed. The selection of the assembly process and the
influence of production rate and quality of the product must be considered by the
manufacturing engineer as the productivity of the facility and workers is balanced.
Jack M. Walker, who was unable to participate in the editing of this book, but
who contributed greatly in the last few months of his life, was a pioneer in new ways
of solving old problems.
Jack loved the advent of rapid prototyping. He spent many hours sharing how
rapid prototyping had applications in choosing methods of manufacture or in selecting
materials that could not be selected by mathematics alone. Jack as the manufactur-
ing engineer loved to place prototypes before the persons responsible for making the
final decision in new products. He often called this “touchy, feely” time the point at
which a person would love or hate the design.
Some products lend themselves to hands-on evaluation, and the finish, appear-
ance, and feel are very important in the final choice of a material in this case. But,
as nanometer-level technology develops, the issues of finish and assembly become
much more critical. An engineering science called tribology deals with the inter-
activity of miniscule particles of materials as they come in contact with each other.
Manufacturing engineers must think in terms of this area of assembly and finishing
and ways to relate experience with larger components to the micron- and nanometer-
sized components used in newer technologies today. Thus, this book was edited to
provide the background and working knowledge for the manufacturing professionals
of the next decade.

Richard D. Crowson
SET, CMfgT, CMfgE

DK565X_000_r04.indd 5 11/16/05 10:28:03 AM


DK565X_000_r04.indd 6 11/16/05 10:28:03 AM
Editor
Richard D. Crowson

Richard Crowson is currently a mechanical engineer at Controlled Semiconductor,


Inc., in Orlando, Florida. He has worked in the field of engineering, especially in the
area of lasers and in the development of semiconductor manufacturing equipment,
for over 25 years. He has experience leading multidisciplinary engineering product
development groups for several Fortune 500 companies as well as small and start-up
companies specializing in laser integration and semiconductor equipment manufac-
ture.
Crowson’s formal engineering training includes academic undergraduate and
graduate studies at major universities including the University of Alabama at Bir-
mingham, University of Alabama in Huntsville, and Florida Institute of Technology.
He presented and published technical papers at Display Works and SemiCon in San
Jose, California.
He has served on numerous SEMI task forces and committees as a voting mem-
ber. His past achievements include participating in writing the SEMI S2 specifica-
tion, consulting for the 9th Circuit Court as an expert in laser welding, and sitting on
the ANSI Z136 main committee that regulates laser safety in the United States.

DK565X_000_r04.indd 7 11/16/05 10:28:03 AM


DK565X_000_r04.indd 8 11/16/05 10:28:03 AM
Contributors
Frank Altmayer Aravinda Kar
Scientific Control Laboratories, Inc. University of Central Florida
Chicago, Illinois Orlando, Florida

Shrikar Bhagath Robert L. Lints


Delco Electronics Corporation Quality Assurance Systems
St. Louis, Missouri
Geoffrey Boothroyd
Boothroyd Dewhurst, Inc. John F Maguire
Wakefield, Rhode Island Materials and Structures Division,
Southwest Research Institute
Robert S. Busk San Antonio, Texas
International Magnesium Consultants,
Inc. Timothy L. Murphy
Hilton Head, South Carolina McDonnell Douglas Corp.
Titusville, Florida
Greg Chandler
Manufacturing Engineering, Hubbell Clyde S. Mutter
Premise Wiring, Inc. Titusville, Florida
Wilmington, North Carolina
Michael Pecht
Stephen C. Cimorelli CALC Electronics Packaging Research
Learjet, Inc. Center (EPRC),
Wichita, Kansas University of Maryland
College Park, Maryland
Richard D. Crowson
Melbourne, Florida Robert E. Persson
EG&G
Denise Burkus Harris Cape Canaveral, Florida
Mechanical Design and Developmental
Engineering Department, Allen E. Plogstedt
Westinghouse Corporation McDonnell Douglas Aerospace East
Baltimore, Maryland Titusville, Florida

Alexander Houtzeel Marc Plogstedt†


Houtzeel Manufacturing Systems ITEC
Software, Inc. Orlando, Florida
Waltham, Massachusetts

† Deceased.

DK565X_000_r04.indd 9 11/16/05 10:28:04 AM


Lawrence J. Rhoades Jeffery W. Vincoli
Extrude Hone Corporation J. W. Vincoli and Associates
Irwin, Pennsylvania Titusville, Florida

Paul R. Riedel Jack M. Walker†


Rockledge, Florida Merritt Island, Florida

Thomas J. Rose William L. Walker


Advance Processing Technology/ National High Magnetic Field
Applied Polymer Technology, Inc. Laboratory, Florida State University
Norman, Oklahoma Tallahassee, Florida

Vijay S. Sheth Don Weed


McDonnell Douglas Corporation Southwest Research Institute
Titusville, Florida San Antonio, Texas

John P. Tanner Bruce Wendle


Tanner and Associates Boeing Commercial Airplane Company
Orlando, Florida Seattle, Washington

V. M. Torbilo Kjell Zandin


Ben-Gurion University of the Negev H. P. Maynard & Company, Inc.
Beer-Sheva, Israel Pittsburgh, Pennsylvania

† Deceased.

DK565X_000_r04.indd 10 11/16/05 10:28:04 AM


Contents
Chapter 1 Manual Assembly 1
John P. Tanner and Jack M. Walker
1.0 Introduction to Manual Assembly 1
1.1 Assembly Work Instructions 4
1.2 Assembly Operation Sequences 5
1.3 Workstation and Line Layout 11
1.4 Manufacturing Methods Analysis 12
1.5 Principles of Motion Economy 20
1.6 Standard Manufacturing Processes 26
1.7 Special Manufacturing Instructions 28
References 29

Chapter 2 Assembly Automation 31


Jack M. Walker and Vijay Sheth
2.1 Introduction to Assembly Automation 31
2.2 Assembly Machines in the Factory 32
2.3 Basic Automation Concepts 32
2.4 Types of Automated Assembly Machines 33
2.5 Motion Systems 38
2.6 Justifying Automation 44
2.7 Software Interfaces to Assembly Automation 48
2.8 Design for Automated Production 49
2.9 Automated Material Handling 57
References 72

Chapter 3 Electronics Assembly 73


Michael Pecht, Denise Burkus Harris and Shrikar Bhagath
3.1 Introduction to Electronics Assembly 73
3.2 Typical Package Architecture 74
3.3 Elementary Subassemblies 74
3.4 Chip Carrier Assemblies 91
3.5 Hybrid Microelectronics Assemblies 95
3.6 Printed Wiring Board Assemblies 122
3.7 System Integration 142
References 144
Suggested Readings 144

Index 147

DK565X_000_r04.indd 11 11/16/05 10:28:04 AM


DK565X_000_r04.indd 12 11/16/05 10:28:04 AM
1 Manual Assembly
John P. Tanner

and

Jack M. Walker

1.0 INTRODUCTION TO MANUAL ASSEMBLY


In today’s complex manufacturing world, it is sometimes difficult to remember what
the real purpose of manufacturing a product is—and what the total elements of the
process consist of. If we assume that whatever we manufacture, we will insist on
good quality products, on-time delivery, and complete customer satisfaction, then
we should concentrate on the most economical method to achieve these results. We
should make our manufacturing decisions based on cost. Of course, cost is not a
simple thing to determine. There are a lot of different ways of looking at cost.
For a method of arriving at the lowest cost while maintaining quality, delivery, and
happy customers, perhaps we should examine a one-person business operation. A few
hundred years ago, there were a lot of them—and even today, there are more than most
of us realize. The nation’s small businesses, those with one to ten employees, grew
in numbers during recession-plagued 1991, resisting the downturn experienced by
medium and larger companies, according to a U.S. Census Bureau report. The number
of small businesses increased up to 1% per year between 1987 and 1991. Larger busi-
nesses increased up to 3% per year through 1990, then declined in 1991; those with 10
to 100 employees were down 0.2% in 1991, while those with more than 100 fell 1.7%.
Businesses with more than 100 employees are generally concentrated in the manufac-
turing sector, which as recently as 1970 accounted for 35% of the workforce. By 1991,
manufacturing workers made up less that 20% of the workforce.
In 1991 there were more than 4 million establishments in the United States with
fewer than 10 employees, about 1.5 million with 10–99 employees, and 134,000
companies with more than 100 employees (a total of 6,199,339 establishments). See
Figure 1.1 for a breakdown of the manufacturing industries.
Now let’s get back to our one-person factory. Upon receipt of an order, the owner
makes each of the parts, assembles them, and does the finish painting, packing, and
delivery. The difference between his total income and the amount of money he spent
during the month is his salary, or profit. Of course, he may be paying rent on the build-
ing, buying raw materials and supplies, and even making payments on his equipment

DK565X_001_r04.indd 1 11/16/2005 1:31:36 PM


2 Assembly Process: Finishing, Packaging, and Automation

U.S. MANUFACTURING INDUSTRIES


(X 100) 1963 1967 1972 1977 1982 1987 1990 1995
Total Establishments 312 311 321 360 358 369

UNDER 20 EMPLOYEES 207 199 203 237 230 238


20–99 " 70 74 76 78 84 86
100–249 " 18 20 21 22 21 22
250–999 " 10 11 11 12 11 11
1000 & OVER " 3 3 3 2 2 2
(X MILLION)
TOTAL EMPLOYEES 17 19.3 19 19.6 19.1 18.9 18.8 18.1
PRODUCTION WORKERS 17.3 14.4 13.5 13.7 12.4 12.2 12.1 11.5
PERCENTAGE OF PRODUCTION WORKERS IN ASSEMBLY TASKS: (1967)

MOTOR VEHICLES 45.6


AIRCRAFT 25.6
TELEPHONE, ETC. 58.9
FARM MACHINERY 20.1
HOME REFRIG. & FREEZERS 32.0
OFFICE EQUIPMENT 35.9
HOME COOKING EQUIP. 38.1
MOTORCYCLES, BICYCLES, ETC. 26.3 CENSUS OF MANUFACTURERS
U.S. BUREAU OF CENSUS

FIGURE 1.1 Statistical breakdown of U.S. manufacturing industry.

and machines. Let’s deduct these, and now we have his profit. Whoops—he probably
pays for heating, lighting, insurance of some type—and taxes—and the remainder
was his profit. We can see that even in the one-person factory, real cost is not so easy
to determine.
If business improves, one person may not be able to do everything by working at
a faster pace, or working longer hours—and at some point the owner will have take
some action in order to continue on-time delivery, et cetera. He might decide to buy
the parts and just perform the assembly operation (or vice versa). Another option might
be to add helpers and continue to perform all the operations in-house. In most cases, a
growing company will probably elect to continue to perform the assembly function in
order to have better control of quality, finish, delivery time, and so forth. It will still be
“their” product as far as the company’s customers are concerned, which will keep the
customers satisfied and give the company the opportunity to add additional sales. As
the company continues to grow, the owner might reconsider his make-or-buy decisions
and perhaps add equipment to fabricate critical parts in-house. In the assembly area, the
first step might be to add automated screwdrivers, nut runners, riveters, spot-welding
heads, and perhaps pick-and-place mechanisms. To move parts from the fabrication or
receiving department to the assembly stations, some type of transfer device might be a
logical improvement. The same applies for the transfer of parts and assemblies down
the assembly line. There are an infinite number of options, including redesigning the

DK565X_001_r04.indd 2 11/16/2005 1:31:36 PM


Manual Assembly 3

product to reduce the number of parts required to be fabricated and thereby simplifying
the assembly process.
In the end, the assembly process may well become the key to the owner’s continu-
ing success. The small firm that started with a one-person assembly operation has now
grown to a multiemployee company, probably using many of the same techniques
that were successful in the beginning of the one-person shop. With more orders, and
probably more diverse products, it is decision time again. This is the subject of this
chapter.
There are as many factors influencing the assembly process decisions as there
are products, customers, and factory managers. Geoffery Boothroyd, in Assembly
Automation and Product Design, quotes Henry Ford’s principles of assembly as
follows:

First, place the tools and then the men in the sequence of the operations so that each
part shall travel the least distance whilst in the process of finishing. Second, use work
slides or some other form of carrier so that when a workman completes his operation
he drops the part always in the same place which must always be the most convenient
place to his hand and if possible have gravity carry the part to the next workman. Third,
use sliding assembly lines by which parts to be assembled are delivered at convenient
intervals, spaced to make it easier to work on them.

Assembly operations can be performed manually, automatically, or integrated


in some manner using a combination of systems. If manual assembly is employed,
an operator can adapt to changing conditions such as those brought about by part
variation, mislocation, and product model mix. An operator can compensate for these
changing conditions and, as a result, may not require elaborate tools and fixtures to
perform the assembly tasks. However, operator error and fatigue can result in quality
problems.
When production volumes are high enough, some assembly operations can be
performed automatically with special-purpose machines. These automatic assem-
bly machines consist of workstations grouped along some type of transfer system
for part conveyance. Each station performs one task with the aid of dedicated sta-
tion equipment, jigs, and fixtures. Part variation, misalignment, and product mix
are not readily adapted to, because sensors cannot always be employed efficiently
or economically to guide or monitor the assembly process. Therefore, part varia-
tions and slight misalignments can result in jamming, incomplete operations, and
excessive machine downtime. However, automation can still be justified when the
production volumes are high, product life is long, and assembly tasks are simple.
For an assembly operation to be performed successfully on a repetitive basis, it is
absolutely essential that part variation and location be minimized and consistency
in dimensions and location be maximized. To achieve this in a mass production
environment requires elaborate and costly tooling, fixtures, and the employment of
expensive production controls. Therefore, many assembly operations are performed
manually to resolve some of the problems in mating parts with variations or mislo-
cations, which may result in increased assembly costs and lower productivity.

DK565X_001_r04.indd 3 11/16/2005 1:31:36 PM


4 Assembly Process: Finishing, Packaging, and Automation

1.1 ASSEMBLY WORK INSTRUCTIONS


“Assembly processing” is another way of saying “assembly methods.” Assembly
methods sheets, or work instructions, must describe clearly what is to be done, in
what sequence, and with what tools and materials. Assembly methods sheets should
minimize operator learning time and must be economical to prepare, reproduce, dis-
tribute, and change. Assembly process planning should include an assembly process
summary or process routing, detailed work instructions for each operation called out
in the summary, an operations parts list for each operation, process sketches or visual
aids, and a workplace layout for each operation. The work instructions should call
out all tools necessary to perform the operation, and there should be a standard time
on the process summary for each operation called out, broken down to the level of
setup and run times.
In process planning for fabrication, whether for machining or forming, the
skilled machinist or sheet metal mechanic could work to what amounts to an out-
line process routing supplemented by the engineering drawing of the piece part. As
a skilled worker, he can set up the machine and perform the work with a minimum
of written work instructions. Such is not the case with assembly operations. The
work must be totally and carefully planned by the manufacturing engineer, and
complete work instructions prepared. These are the two extremes. In most manu-
facturing plants today, process planning will fall somewhere in between.
If the plant is a high-volume producer of a single product line, then detailed
assembly work instructions may be unnecessary. Once operators are trained to per-
form a short-cycle assembly operation, little else is needed except possibly some
clear, concise visual aids showing the critical details of the operation in pictorial, or
exploded view, form. However, the manufacturing engineer must plan such produc-
tion down to the most detailed level. He or she must prepare a layout of the assem-
bly lines, show each and every workstation in plan view, plan the assembly tools
required, and write a complete description of the work performed at each station on
the line. The manufacturing engineer must establish standard times and decide where
visual aids are needed and prepare them, and then fine tune or balance the line, assist
in training the operators, and finally shake down or debug the line.
All of the above documentation is necessary when assembly lines are initially
established or set up, and to train the operators. Once the line is flowing smoothly
and the operators are trained, there will be less and less reliance on written work
instructions and even visual aids. This initial planning documentation should always
be available for ready reference and should be kept up to date by the manufacturing
engineer.
If the company manufactures a variety of different product lines in medium to
high volumes, or sets up and produces to a job-order-type system, or does both,
assembly process documentation that is complete and to the greatest level of detail
is especially important. It is a proven fact that good assembly process planning and
documentation significantly reduce operator learning (and relearning) time. This
is especially important when the production run is relatively short. It also teaches
the correct methods to operators and thereby reduces costs of assembly labor.

DK565X_001_r04.indd 4 11/16/2005 1:31:36 PM


Manual Assembly 5

The assembly process documentation package is essential to the operation of an


ongoing production-control and time-keeping system. The assembly process routing
provides the steps or sequences that materials, parts, assemblies, and work in process
must follow to build the product. It provides the time standards for each operation,
and the assembly parts list for each operation provides the information needed by
production control to pull and kit material for production.
In a small plant, where production runs may be small to nonexistent, assembly
process planning with only minimum documentation is required and can be justified
for the reasons mentioned earlier. Even in the case where no formal production con-
trol system exists and the production supervisor draws material from the stockroom
in one batch issue for the entire job, pictorial visual aids, workstation layouts, a tool
list, and an assembly process routing should be provided.

1.2 ASSEMBLY OPERATION SEQUENCES


In assembly process planning, operation sequences usually parallel the indented
parts list or engineering tree chart, because it should represent how the product
goes together or is assembled. This initial assembly process sequence plan should
define an assembly operation for each major and minor subassembly and for the
final assembly (Figure 1.2). It should be emphasized that this is an initial breakdown
and normally will be followed by a more thorough analysis of the steps required to
assemble the various subassemblies and the final assembly. This detailed analysis
is normally done in the preproduction planning phase in the form of an operation

Material Feeding Into The Process

Purchased Purchased Purchased


Purchased
Material Material Material
Material
Subassembly
Or Part
Which Joins Another

Part Or Top Assembly On Which Most Operations


Subassembly Or Part

Subassembly Or Part

Which Joins Another


Subassembly Or Part

Subassembly Or Part
Steps Of Process In Sequential Order

Material On Which
Work Is Performed

Material On Which
Work Is Performed
Are Performed

Material On Which Work Is Performed

Purchased
Material

FIGURE 1.2 Principles of operation process chart construction.

DK565X_001_r04.indd 5 11/16/2005 1:31:37 PM


6 Assembly Process: Finishing, Packaging, and Automation

process chart. Figure 1.3 shows an example of an operation process chart for a Coast
Guard radio receiver.
The assembly process may include soldering, wiring, press fitting, brazing,
shrink fitting, welding, adhesive bonding, riveting, and mechanical fastening. Within
each of these assembly processes a series of sequences is required to accomplish the

FIGURE 1.3 Operation process chart for a radio guard receiver.

DK565X_001_r04.indd 6 11/16/2005 1:31:38 PM


Manual Assembly 7

process, without regard to the product configuration, material, or quantity to be pro-


duced, or the rate of production. For example, many of the steps in creating a circuit
card assembly, a wire harness, the frame of a truck, or in the installation of fittings on
a sailboat are essentially the same. The detailed instructions for the sequence should
spell out the differences peculiar to the product at hand.
Often in assembly work, standard sequences or operations are possible for any
product where these processes of assembly are used. The result is a considerable saving
in manufacturing engineering time and in the elapsed time required to prepare and
release an assembly process plan to the shop. Such standard processes enable preprinted
process planning documentation, which may only require a part number and quantity
to be entered before it is ready for release. A good example of this is in the manufacture
of cables for electrical or electronic equipment, where diagrams are preprinted of the
various electrical connector pin configurations, requiring only that the manufacturing
engineer sketch in the wires terminated to the pins for the specific cable application.
Computer-aided process planning (CAPP), covers the use of a database to accomplish
this task.

1.2.1 Routings, Work Instructions, and Visual Aids


Assembly process routings, such as the one shown in Figure 1.4, list the operations in
the sequence in which they must occur to assemble the item or product called out in the
heading. In addition to listing the operations in their proper sequence, it lists the standard
times for each operation, the performing department, and the latest revision level of
the process instruction sheets. The issue or revision level of the process information
establishes configuration control of the product on the shop floor, because normally the
assembly department does not work to engineering drawings. The importance of this
cannot be emphasized enough. It is the responsibility of the manufacturing engineer
always to have the latest revisions to the engineering drawings incorporated into the
assembly process documentation, especially when the job is active on the production
floor. In many companies the inspection department uses the process documentation
to perform in-process inspections of the product. This is especially true where detailed
process work instructions are used, and the process documentation is also used for shop
configuration control.
As indicated earlier, the assembly process summary or routing is used by production
control to move material or work in process to the next operation or sequence. A copy
of the assembly process summary travels with each batch of parts and material and in
effect becomes a routing sheet or shop traveler. For this to happen, the sequences must
be stamped off, either by inspection or by the operator, as they are completed. If line
production is involved, such a routing or traveler is unnecessary, as the progression on
the assembly line is the routing followed by the assemblies. The process routing is an
especially valuable tool in the job shop, where the shop is laid out by function or pro-
cess, and does not follow the product flow.
Assembly work instructions are the heart of the manufacturing engineering
documentation package, and explain how the product is to be assembled in production.
The assembly work instructions should be available at the operator workstation,

DK565X_001_r04.indd 7 11/16/2005 1:31:38 PM


8 Assembly Process: Finishing, Packaging, and Automation

FIGURE 1.4 Assembly process summary for wiring harness subassembly.

preferably on an easy-to-see holder mounted on the workbench. The assembly pro-


cess routing lists all of the operations for assembly, in the sequence that they must
be performed; the assembly work instructions or assembly methods sheets for each
listed operation explain in detail exactly how to perform the operations. Figure 1.5 is an
example of an assembly methods sheet for the wiring of a connector that becomes
part of a wiring harness in a marine short-wave radio receiver.

DK565X_001_r04.indd 8 11/16/2005 1:31:39 PM


Manual Assembly 9

FIGURE 1.5 Assembly methods sheet for wiring a connector.

An assembly operation parts list should be included with the assembly work instruc-
tions or assembly methods sheets for each operation. This tells production control and
the operator what parts and materials are required to perform the assembly operation
for one unit. Figure 1.6 shows one version of such a parts list. In addition, there must
also be a list of standard and special design assembly tools needed to perform the
operation. In this example it is included in the operation parts list. It can also appear as
a separate call-out on the work instruction or methods sheets.

DK565X_001_r04.indd 9 11/16/2005 1:31:39 PM


10 Assembly Process: Finishing, Packaging, and Automation

FIGURE 1.6 Assembly operations parts list.

An extremely important part of the assembly process documentation is visual aids.


Visual aids can be anything from an actual mockup of the product to a black and white
or color photograph, to a three-dimensional isometric or exploded-view drawing, to a
simple sketch, or to a tracing lifted directly from the engineering drawing. Figure 1.7
shows a light table being used by a manufacturing engineer to trace parts of the engi-
neering drawing in order to make up a visual aid. If regular office copy machines are

DK565X_001_r04.indd 10 11/16/2005 1:31:41 PM


Manual Assembly 11

FIGURE 1.7 Manufacturing engineer using a light table to make a tracing


from an engineering drawing in order to prepare a visual aid.

used, visual aids can be constructed using cut-and-paste methods as shown in Figure 1.8.
Illustrations may also be done by graphic artists or illustrators as shown in Figure 1.9. It
should be kept in mind that all process documentation costs money, and consideration
should be given to the length of the production run, anticipated changes during produc-
tion, and what is really needed to instruct a particular group of assembly operators. One
other very important consideration is that a visual aid supplemented by minimal notes
and instructions is far superior to lengthy written work instructions. As explained earlier,
visual aids that highlight key assembly details are all that are used in many companies.

1.3 WORKSTATION AND LINE LAYOUT


Workstation layouts are important from the standpoint of assembly operator methods.
They tell the assembly supervisor how to set up and configure the individual
workstations for optimum productivity and flow of work. Workstation layouts are usually
in the form of a plan view of the workstation and show where tooling, fixturing, and part
bins should be placed, and include work instructions, tote pans for staging completed
work and for placing incoming work, and any other information pertinent to the opera-
tion and the setup of the workstation. Figure 1.10 shows a workstation layout used in a
small job order shop. Figure 1.11 shows how this layout might look set up in the shop.
Line layouts are used where progressive assembly lines are to be used to build
the product, and the plant layout drawings do not show which workstation goes
where. Again, these are used by supervision in setting up the line to conform to
the assembly process flow and to ensure optimum methods and work flow. See
Figure 1.12 for an example.

DK565X_001_r04.indd 11 11/16/2005 1:31:41 PM


12 Assembly Process: Finishing, Packaging, and Automation

FIGURE 1.8 Inexpensive visual aid prepared by the cut-and-paste method.

1.4 MANUFACTURING METHODS ANALYSIS


In developing the manufacturing process, whether for fabrication, machining, forming,
finishing, or assembly, the manufacturing engineer must specify the most economical
methods for the job and for the work at hand. In order to do this he or she must under-
stand and be able to apply the fundamental techniques of methods analysis, motion
economy, and work simplification. The presumption of good methods and the ability of
the manufacturing engineer in methods analysis is so basic that not to be proficient in
this art is tantamount to being incompetent as a manufacturing engineer. The purpose
and intent of this section is to provide the basic information needed by the manufactur-
ing engineer to gain a degree of proficiency in manufacturing methods analysis.
Even the best and most thorough process planning will sometimes overlook
details or specify methods that can be improved upon later, after the product is in

DK565X_001_r04.indd 12 11/16/2005 1:31:42 PM


Manual Assembly 13

ADAPTER HALVES

NEW SLING

STRAIGHT
HEADED PIN
ADAPTER
FORWARD SLING RETAINER
SLING-HOOK

ADAPTER
INSUL
SLEEVING SIDE VIEW

TOP VIEW ADAPTER

FIGURE 1.9 Visual aid made using formal artwork.

FIGURE 1.10 Assembly workstation layout in diagram form with parts bin setup diagram.

DK565X_001_r04.indd 13 11/16/2005 1:31:45 PM


14 Assembly Process: Finishing, Packaging, and Automation

FIGURE 1.11 Assembly workbench arrange-


ment as it would appear on the shop floor.

FIGURE 1.12 Manual line assembly with manual transfer of the workpieces (a) in a line
arrangement, and (b) in rectangular form. (Courtesy of IPA Stuttgart.)

DK565X_001_r04.indd 14 11/16/2005 1:31:45 PM


Manual Assembly 15

production. Part of the job of the manufacturing engineer is to be alert for these
opportunities to improve the process and the flow.

1.4.1 Work Simplification


Work simplification can be defined as the organized application of common sense
to find easier and simpler ways of doing work. Work simplification provides a sys-
tematic, common-sense approach to make work easier and at the same time to lower
costs. The basic premise of work simplification is that there is always a better way to
do any task. The work simplification pattern includes five basic steps:

1. Selection of a job to be improved


2. Recording of the job details
3. Analysis of the job details
4. Development of the improvements
5. Installation of the improvements

Selecting the job to be improved requires careful consideration and study. Efforts
expended for improvement should be made first where the returns will be the greatest.
Priority should be given to bottlenecks, choke points, trouble spots, jobs that require
excessive amounts of time, or where generally unsatisfactory conditions exist. The
following list provides assistance in making this selection:

1. Greatest cost: work that involves the greatest expenditure of funds, labor
hours, or use of equipment
2. Greatest workload: the largest volume of work being performed by the
activity
3. Number of persons assigned: work that requires large numbers of people
to perform similar tasks
4. Walking: jobs that require a lot of walking around
5. Bottlenecks: work not flowing smoothly
6. Schedules not met: failure to meet deadlines, resulting in work backlogs or
overtime
7. Excessive waste: work that results in wasted materials, in scrap, and in
rework
8. Excessive fatigue: work that requires great physical effort or that is being
done with frequent rest periods
9. Unsafe or unpleasant work practices: work that results in numerous acci-
dents or is undesirable because of extreme conditions such as dust, noise,
fumes, vapors, or extremes of temperature

Every job is made up of three parts:

1. Make ready: the time and effort put into the setup, or getting ready to work
2. Do: the actual work accomplished

DK565X_001_r04.indd 15 11/16/2005 1:31:45 PM


16 Assembly Process: Finishing, Packaging, and Automation

3. Put away: the time and effort put into cleaning up after the “do” part of
the job

For example, if we have a carpenter making a wooden box, we might expect the
job to break down as follows:

1. Make ready: open bin, pick up nails, close bin, pick up hammer, lay nails
on the bench
2. Do: hammer nails
3. Put away: put hammer aside, pick up box, put box aside

Anything that reduces the time required for the “make ready” and “put away” parts
of the job reduces the nonproductive time associated with the job.
Recording of the job details can best be accomplished through the use of process
chart techniques, specifically the flow process chart and the flow diagram. The flow
process chart is a device for recording each step of a job in a compact manner, as a
means of better understanding it and improving it. The chart represents graphically the
separate steps of the events that occur during the performance of work, or during a series
of actions. The process chart may be used to record the flow within a unit, a section,
a department, or between departments. The flow process chart has no bounds.
No matter how complicated or intricate the series of operations may be, a
flow process chart can be constructed if you take one step at a time. The flow
process chart, however, like other methods of graphic representation, may need
to be modified to meet the requirements of a particular situation. For example, it
may show in sequence the total activity of a production operator, or it may show
in sequence the steps that the worker, part, or material goes through. The chart
could be either the operator type or the material type, and the two types should
not be combined.
A careful study and analysis of such a chart, giving a graphic picture of every step
in the process, is almost certain to suggest improvements. It is not uncommon to find
that some operations can be eliminated or that a part of an operation can be eliminated,
that one operation can be combined with another, that better routes for the parts can
be found, more economical machines can be used, delays between operations can be
eliminated, and other improvements can be made, all of which go to produce a better
product at a lower cost.
To make a flow process chart requires careful adherence to the following rules:

1. State the activity being studied. Make certain you are really naming the
activity you have chosen to study.
2. Choose the subject to follow. Decide on a person or a material, and follow
him or it through the entire process. When you have picked a subject, stick
with it.
3. Pick a starting and ending point. This is to make certain you will cover all
the steps you wish to cover, but no more or less.
4. Write a brief description of each detail. Step by step, no matter how short
or temporary, describe each detail.

DK565X_001_r04.indd 16 11/16/2005 1:31:46 PM


Manual Assembly 17

5. Apply the symbols. The description determines each symbol. Draw a con-
necting line between each of the proper symbols.
6. Black in the “do” operation. Shade in the symbols for those operations
that you decide are the “do” operations. This will help you later in your
analysis, when you begin challenging.
7. Enter distances. Whenever there is a transportation, enter the distance
traveled.
8. Enter time if required. This is often not necessary. However, if it will help,
note the time required or elapsed.
9. Summarize. Add up all of the facts and put them in the summary block.
The summary should indicate the total number of operations, transporta-
tions, inspections, delays, storages, and distance that is traveled.

Figure 1.13 shows a flow process chart that is completely filled in, with circled
numbers used to illustrate each of the rules listed above. The elements of the flow
chart are as follows; please see Figure 1.13 or their corresponding symbols:

O Operation. An operation occurs when an object is intentionally changed in


any of its physical or chemical characteristics, is assembled or disassem-
bled from another object, or is arranged or prepared for another operation,
transportation, inspection, or storage. An operation also occurs when infor-
mation is given or received, or when planning or calculating takes place.
⇒ Transportation. A transportation occurs when a person moves from one
workplace to another or when an object is moved, except when such
movements are part of the operation or are caused by the operator at the
workstation during an operation or an inspection.
Inspection. An inspection occurs when an object is examined for identification
or is verified for quality, quantity, or any of its characteristics.
D Delay. A delay occurs to an object when conditions, except those that inten-
tionally change the physical or chemical characteristics of the object, do not
permit or require immediate performance of the next planned action.
∇ Storage. A storage occurs when an object is kept and protected against
unauthorized removal.

It is sometimes helpful to supplement a flow process chart with a flow diagram.


A flow diagram is simply a layout of the area involved in the job being studied,
over which you indicate by a line the path of the object or person followed in the
flow process chart. It is often desirable to indicate the action taking place by using
the same symbols as on the chart. They may, if desired, be keyed to each other by
item numbers. Figure 1.14 shows a flow diagram to accompany the flow process
chart in Figure 1.13.
The third step of the work simplification pattern involves questioning every part,
aspect, or detail of the job. Here we examine each operation and ask some very
pointed questions. Don’t be satisfied until you have asked all possible questions and

DK565X_001_r04.indd 17 11/16/2005 1:31:46 PM


18 Assembly Process: Finishing, Packaging, and Automation

FIGURE 1.13 Flow process chart example.

DK565X_001_r04.indd 18 11/16/2005 1:31:47 PM


Manual Assembly 19

Operations: Pricing and Posting Orders

Subject Charted: Unpriced Orders

PRESENT METHOD LAYOUT

Time stamp

PBX

A
Wooden Box

Files ACCOUNTING
Files
C OFFICES B

Clerk A – Receptionist and PBX Operator


Clerk B – Pricing Clerk
Clerk C – Posting Clerk

FIGURE 1.14 Flow diagram example.

received the related answers. The first thing to do is to question the entire job being
studied. Why is it done? Is the job really necessary? If it is, then question each “do”
operation. If you can eliminate a “do” operation, you also eliminate the “make ready”
and “put away” operations that go with it. The “do” operations are those that add
value to the product or process being studied.
There is always a better way, and our task is to find it.

Why? The overriding question; it establishes the reason for the job. The answer
defines and justifies the purpose of the job.
What? What is done? What are the steps? What does each step do? What
makes the step necessary?

DK565X_001_r04.indd 19 11/16/2005 1:31:48 PM


20 Assembly Process: Finishing, Packaging, and Automation

Where? Where should this step be done? Can it be done more easily? Can it be
done using less time, energy, or transportation, or by changing the location
of employees or equipment?
When? When should this step be done? Is it done in the right place in the
sequence? Can the job be simplified by moving this step ahead or back?
Who? Who should do the job? Is the right person handling it? Would it be
more logical to give the job to someone else?
How? How is the job being done? Can it be made easier? Can the job be done
better with different equipment or a different layout?

The following question-and-answer approach will suggest improved methods:

Why and what lead to elimination.


Where, when, and who lead to combination or sequence change.
How leads to simplicity.

Careful consideration of the possibilities presented in looking for ways to eliminate,


combine, change sequence, and simplify in the questioning approach brings us finally
to the better method and provides us with the answer to “How should the job be done?”
The simplest way is the best way. Figure 1.15 shows a proposed improvement flow
process chart for the pricing and posting of orders in Figure 1.14.

1.5 PRINCIPLES OF MOTION ECONOMY


The following discussion explores the rules or principles of motion economy that
have been and are now being used successfully in manufacturing methods studies.
These principles form a basis, code, or body of rules that, if applied correctly, make
it possible to greatly increase the output of manual factory labor with a minimum of
fatigue. These principles will be examined under the subdivisions of operator tasks
and the workplace, and as applied to tools and equipment.

1.5.1 Operator Tasks


The principles of motion economy as related to the tasks of the operator are as follows:

1. The two hands should begin as well as complete their motions at the
same time.
2. The two hands should not be idle at the same time except during rest
periods.
3. Motions of the arms should be made in opposite and symmetrical directions
and should be made simultaneously.

These three principles are closely related and should be considered together. It seems
natural for most people to work productively with one hand while holding the object

DK565X_001_r04.indd 20 11/16/2005 1:31:48 PM


Manual Assembly 21

FIGURE 1.15 Flow process chart showing improved method plan.

being worked on with the other hand. This is extremely undesirable and should be
avoided. The two hands should work together, each beginning a motion and complet-
ing a motion at the same time. Motions of the two hands should be simultaneous and
symmetrical.
Many kinds of work can be accomplished better using both hands than by
using one hand. For most manufacturing assembly operations, it is advantageous
to arrange similar work on the left- and right-hand sides of the workplace, thus
enabling the left and right hands to move together, each performing the same
motions. The symmetrical movements of the arms tend to balance each other,
reducing the shock and jar on the body and enabling the operator to perform the

DK565X_001_r04.indd 21 11/16/2005 1:31:49 PM


22 Assembly Process: Finishing, Packaging, and Automation

task with less mental and physical effort. There is apparently less body strain when
the hands move symmetrically than when they make nonsymmetrical motions,
because of issues of balance.
The fourth principle of motion economy states that hand and body motions
should be confined to the lowest classification with which it is possible to perform
the work satisfactorily.

1.5.2 Classes of Hand Motions


The five general classes of hand motions emphasize that material and tools should be
located as close as possible to the point of use. The motions of the hands should be as
short as the work permits. In the listing of classifications shown below, the one requiring
the least amount of time and effort is shown first:

1. Finger motions
2. Motions involving fingers and wrist
3. Motions involving fingers, wrist, and forearm
4. Motions involving fingers, wrist, forearm, and upper arm
5. Motions involving fingers, wrist, forearm, upper arm, and shoulder (causes
posture change)

It should be pointed out that finger motions have been found to be less accurate,
slower, and more fatiguing than motions of the forearm. Evidence seems to indicate
that the forearm is the most desirable member for performing light work. In highly
repetitive work, motions about the wrist and elbow are superior to those of the fingers
or shoulders.
The fifth principle of motion economy states that momentum should be
employed to assist the worker wherever possible, and it should be reduced to a
minimum if it must be overcome by muscular effort. The momentum of an object is
defined as its mass multiplied by its velocity. In the factory environment, the total
weight moved by the operator may consist of the weight of the material moved, the
weight of the tools moved, and the weight of the part of the body moved. It should
be a real possibility to employ momentum to advantage when a forcible blow or
stroke is required. The motions of the worker should be so arranged that the blow
is delivered when it reaches its greatest momentum.
The sixth principle of motion economy states that smooth, continuous, curved
motions of the hands are preferable to straight-line motions involving sudden and
sharp changes in direction. Abrupt changes in direction are not only time-consuming
but also fatiguing to the operator.
The seventh principle of motion economy states that ballistic motions are faster,
easier, and more accurate than restricted or controlled movements. Ballistic move-
ments are fast, easy motions caused by a single contraction of a positive muscle
group, with no antagonistic muscle group contracting to oppose it. A ballistic stroke
may be terminated by the contraction of opposing muscles, by an obstacle, or by
dissipation of the momentum of the movement, as in swinging a sledge hammer.

DK565X_001_r04.indd 22 11/16/2005 1:31:49 PM


Manual Assembly 23

Ballistic movements are preferable to restricted or controlled movements and should


be used whenever possible.
The eighth principle of motion economy states that work should be arranged to
permit an easy and natural rhythm wherever possible. Rhythm is essential to the smooth
and automatic performance of any operation. Rhythm, as in a regular sequence of uni-
form motions, aids the operator in performing work. A uniform, easy, and even rate
of work is aided by proper arrangement of the workplace, tools, and materials. Proper
motion sequences help the operator to establish a rhythm that helps make the work a
series of automatic motions where the work is performed without mental effort.
The ninth principle of motion economy states that eye fixations should be as few
and as close together as possible. Where visual perception is required, it is desirable to
arrange the task so that the eyes can direct the work effectively. The workplace should be
laid out so that the eye fixations are as few and as close together as possible.

The Workplace
The first principle of motion economy related to the workplace states that there should
be a definite and fixed place for all tools and materials. The operator should always have
tools and materials in the same location, and finished parts and assembled units should
be placed in fixed positions or locations. For example, in the assembly of mechanical
hardware, the hand should move without mental direction to the bin containing flat
washers, then to the bin containing lock washers, then to the bin containing bolts, and
finally to the bin containing hex nuts. There should be no thinking required on the part
of the operator to do any of this.
The second principle of motion economy related to the workplace states that
tools, materials, and controls should be located close to the point of use. In the hori-
zontal plane, there is a definite and somewhat limited area that the worker can use
with a normal expenditure of effort. This includes a normal working area for the
right hand and one for the left hand for each working separately, and another for
both hands working together. Figure 1.16 shows this and the dimensions of normal
and maximum working areas in the horizontal and vertical planes. Both the standing
and sitting positions are included. It also shows normal bench work surface heights,
which can have a significant adverse effect if they are not correct.
Figure 1.17 shows in greater detail the areas of easiest reach for the left and right
hands, for both hands working together, and the area in which small objects can be
most easily picked up.
The third principle of motion economy related to the workplace states that
gravity-feed bins and containers should be used to deliver the material close to the
point of use. This can sometimes be accomplished by using parts bins with slop-
ing bottoms that feed parts by gravity to the front of the bin, eliminating the need
for the assembly operator to reach down into the bin to grasp parts.
The fourth principle of motion economy related to the workplace states that drop
deliveries should be used wherever possible. This requires configuring the work-
place, for example, so that finished units may be disposed of by releasing them in
the position in which they are completed, delivering them to their next destination by

DK565X_001_r04.indd 23 11/16/2005 1:31:49 PM


24 Assembly Process: Finishing, Packaging, and Automation

Total height

Eye level

Shoulder
28"
14"

Normal Elbow
working 2"
height 9½"

63"
15" Seat

52"
58"
39"
Knee
37"

29"
20"
10"

Maximum work Work position


area
Normal
work
area
"
.5
23
"
14

Edge of work
height 14"
9½"
2"

40"

58"

FIGURE 1.16 Normal and maximum working areas and heights.

gravity. Besides the savings in time, this frees the two hands so that they may begin
the next cycle immediately without breaking the rhythm.
Other principles of motion economy related to the workplace include the
following:

Materials and tools should be located to permit the best sequence of motions.
Provision should be made for adequate lighting.

DK565X_001_r04.indd 24 11/16/2005 1:31:52 PM


Manual Assembly 25

LEFT HAND RIGHT HAND


E
(a)

A C B D

F E G
(b)

A D

F X Y
E G
(c)

W Z
A D

FIGURE 1.17 Horizontal view of areas of easiest reach for each hand and for both hands
working together. (a) Maximum areas of reach for left and right arms. (Broken lines indicate
enclosed area covered by hands when forearm is pivoted on the bent elbow. (b) Area inside in
which small objects are most easily picked up. (c) Area in which the eye can follow both hands
working simultaneously and symmetrically.

The height of the workplace and chair should be arranged so that alternate
sitting and standing at work are easily possible.
A chair of the type and height to permit good posture should be provided for
the operator.

1.5.3 Tools and Equipment


Principles of motion economy as related to the design of tools and equipment include
the following:

1. The hands should be relieved of all work that can be done more effectively
by a jig, fixture, or foot-operated device.
2. Two or more tools should be combined wherever possible.
3. Tools and materials should be prepositioned whenever possible.

DK565X_001_r04.indd 25 11/16/2005 1:31:54 PM


26 Assembly Process: Finishing, Packaging, and Automation

4. Where each finger performs some specific movement, the load should be
distributed in accordance with the inherent finger capacities (arrangement
of typewriter keys).
5. Levers, crossbars, and hand wheels should be located in such positions
that the operator can operate them with the least change in body position,
and with the greatest mechanical advantage.

1.6 STANDARD MANUFACTURING PROCESSES


In addition to operator work instructions, process routings, visual aids, and operation
parts lists, a key ingredient of the documentation package that should be provided
by manufacturing engineering includes standard manufacturing processes. Standard
manufacturing processes include workmanship standards, equipment operating pro-
cedures, and standard repair procedures. These standard processes are common to all
products manufactured in a plant and therefore are either referenced in the product
work instructions, as in “Assemble per standard manufacturing process 367,” or the
appropriate text or illustrations, or both, are copied directly into the product instruc-
tions, thereby eliminating the requirement for the production operator to leave the
work position and look up information in a separate book or document. Obviously,
the latter method is preferred.

1.6.1 Workmanship Standards


Workmanship standards tell the operator what is acceptable work or practice and
what is not. These standards can be in the form line drawings supplemented with
narrative text, color photographs of acceptable and unacceptable work supple-
mented by narrative text, or actual models or prototype units, known as stan-
dards. Figure 1.18 shows an example of a line-drawing workmanship standard for
taping the core of an iron-core transformer in the magnetic components industry.
It should be noted that instructions are superimposed where needed, pointing out
the important points to watch for when tape-wrapping the core. Also, it should be
noted that this workmanship standard tells only the requirements for making an
acceptable tape wrap and that it is necessary to refer to a product specification to
determine the tape material and number of layers required.

1.6.2 Equipment Procedures


Equipment procedures are start-up, operating, and shut-down instructions for a machine
or piece of production equipment. They should be posted at or near the machine or
equipment and are usually prepared by the manufacturing engineer from the manual that
is prepared by the maker of the machine or equipment. These procedures are especially
valuable where process variables controlled by machine or equipment settings can be
critical. Equipment procedures ensure the proper training of new operators and serve as
reminders to the experienced operator. Equipment procedures should be prepared for
everything from vapor degreasers to heat-treat ovens.

DK565X_001_r04.indd 26 11/16/2005 1:31:54 PM


Manual Assembly 27

ONE HALF OVERLAP


ON INSIDE

*LEADS
LESS THAN ONE
HALF OVERLAP ON
OUTSIDE

NO EXCESSIVE BUILD UP
OF TAPE ON INSIDE
CORNERS OF COIL
*LEADS
A. WIND TAPE UP CLOSE TO LEAD
B. FOLD LEAD OVER TAPE
C. CONTINUE WINDING TAPE
SECURE START AND FINISH WITH MYLAR TAPE IF NECESSARY
TAPE MATERIAL AND NUMBER OF LAYERS TO BE SPECIFIED
ON D-SPEC.

FIGURE 1.18 Workmanship standard for taping iron-core transformers.

1.6.3 Standard Repairs


Standard repair procedures provide ready-made work instructions to be used any time
certain types of repair work must be done to product parts, subassemblies, or assem-
blies. These procedures or instructions have been approved in advance by inspection
or quality control, and if necessary, by design engineering and the customer. They
not only tell manufacturing how to repair certain kinds of product defects, but they
also eliminate the need for formal rejection by inspection before the repair procedure
can be implemented.
For example, the standard repair for a mislocated or design-changed hole in a
metal part might read as follows:
1. The maximum number of holes to be plugged or welded in any one part
by this procedure is 20% of the total holes in the part, or eight (8) holes,
whichever is less.
2. The preferred method for repair shall be welding, except where heat would
cause distortion.
3. In applications where surface heat due to welding will cause distortion,
a press fit plug or a formed flat head rivet are acceptable alternatives.

DK565X_001_r04.indd 27 11/16/2005 1:31:55 PM


28 Assembly Process: Finishing, Packaging, and Automation

4. The holes to be filled by weld material shall be prepared by chamfering both


sides of the hole. The chamfer shall be sufficient to insure weld penetration.
5. Surfaces to be welded shall be cleaned in accordance with established
welding practices for the specific type of material involved.
6. The holes shall be completely filled with weld material.
7. Welding shall be accomplished by a certified welder.
8. The surfaces shall be ground or otherwise made flush to eliminate weld
buildup.
9. Holes to be filled by plugging shall be prepared by the incorporation of a
chamfer or lead on the side to accept the plug.
10. The plug shall be of such diameter that a press fit is ensured.
11. After plug insertion, the surfaces shall be ground or otherwise made flush
to eliminate excess material.
12. The plug shall be staked to the part per approved practices.
13. Holes to be plugged with a rivet shall be countersunk on both sides to
accept a flat head rivet.
14. The rivet shall be formed and ground or otherwise made flush with the
mating services.
15. Rivets shall not be loose after forming.

1.7 SPECIAL MANUFACTURING INSTRUCTIONS


Frequently it becomes necessary for the manufacturing engineer to look beyond the
operations in his or her own plant and to make certain that material and parts received
from outside suppliers come into the plant and the manufacturing operations ready to
become part of the product being produced. In the majority of cases, if the vendor meets
the requirements of the purchase order and the engineering drawings (if applicable), the
parts or material will be ready for processing when they arrive at the receiving dock.
In some instances, however, it may be required that the vendor only partially complete
a part or subassembly, or that the vendor produce a certain part on the low side of the
drawing tolerance if the parts are to fit together in assembly.
To ensure that this information is correctly transmitted to the vendor, the manu-
facturing engineer may prepare special manufacturing instructions that are called out
on the purchase order along with engineering drawings and specifications. The special
manufacturing instructions along with the drawings and specifications then become
the acceptance criteria for the parts or material when they go through receiving
inspection.
Special manufacturing instructions are prepared by the engineer at the time the
make-or-buy determination is made and before the detailed internal work instruction
documentation is prepared. These instructions can be in the form of simple handwrit-
ten information given to the buyer to incorporate directly into the purchase order, or
a formal document that becomes an attachment to the purchase order. The important
point to remember is that the manufacturing engineer needs to review the items to be

DK565X_001_r04.indd 28 11/16/2005 1:31:55 PM


Manual Assembly 29

purchased and make certain that if these items are made exactly as the drawings and
specifications say, they can enter directly into production without any problems or
unplanned processing steps.

REFERENCES
Boothroyd, G., Assembly Automation and Product Design, Marcel Dekker, New York, 1992.
Tanner, J. P., Manufacturing Engineering, Marcel Dekker, New York, 1991.

DK565X_001_r04.indd 29 11/16/2005 1:31:55 PM


DK565X_001_r04.indd 30 11/16/2005 1:31:55 PM
2 Assembly Automation
Jack M. Walker

and

Vijay Sheth

2.1 INTRODUCTION TO ASSEMBLY AUTOMATION


In this space age, when society regards everything as possible, there are many mis-
conceptions about automation. Automated technology has been prescribed as the
ultimate cure-all for all technical and social problems. The disinclination of workers
to perform dull, repetitive tasks, considerations of worker health and safety, dramatic
increases in labor rates and fringe benefits, and the increased ability of tool engineers
to devise sophisticated machinery have combined to create a dramatic increase in
automated equipment during the past several decades. Computer-controlled machine
tools now make complex machined parts, electronic components are installed auto-
matically in circuit boards, machines inspect detail parts on a 100% basis, and parts
are positioned, assembled, and checked out automatically in the creation of assem-
blies—untouched by human hands. Robots of all descriptions have become part of
the industrial scene.
It is true that most businesses automate primarily to reduce costs and thereby improve
their competitive position in the market. However, the real objective of this investment
is to make money, not just save money. When compared to manual assembly operations,
the benefits usually derived from automated assembly include the following:

Reduced unit costs


Consistent high quality
Elimination of hazardous manual operations
Increased production standby capacity

The aerospace industry, because of its low production quantities and low-rate
production, has been relatively untouched by automation except in the areas of
complex machined parts and in the assembly of some types of electronic devices,
when the needs for precision and accuracy justify the cost and time of automation.
However, the production quantities of some of the more complex weapons are now
adequate to consider automation, especially where a number of individual parts are
identical on each final assembly. Examples of such weapons are the Dragon rocket

31

DK565X_002_r04.indd 31 11/16/2005 1:29:55 PM


32 Assembly Process: Finishing, Packaging, and Automation

motors and firing circuit boards, all of which have multiple uses on each Dragon
missile, designed and built by McDonnell Douglas.
Semiautomated operations are those in which the worker plays a substantial role
in the activity. The worker’s role exceeds that of supplying the automated equipment
with parts or materials or removing finished parts from the work area. To date the
preponderance of factory operations have actually been semiautomated rather than
fully automated because the worker–machine combination is often the most efficient
and effective in involved tasks.

2.2 ASSEMBLY MACHINES IN THE FACTORY


When they see the term automated factory, many people think of a Japanese factory
containing hundreds of robots. But Japan is not alone in this field; many U.S. corpora-
tions are making similarly impressive strides in boosting factory productivity and effi-
ciency. And robotics, though critical to some manufacturers, represents only a small part
of the overall automation scene. Of major importance are the computer and communica-
tions networks that manage and analyze the information relevant to factory operation.
In addition, every manufacturing facility has its own unique needs. The fabrication and
assembly equipment and the processes that can serve them best can vary considerably.
Although lower product unit costs (and increased overall company earnings) are
the prime drivers in selecting an automated process, the manufacturing process may
be automated in areas where the work is dangerous or monotonous for people, or
where machines can substantially improve the quality of the products or increase the
production rate. Moreover, the products themselves should be designed to facilitate
automated manufacturing, and changes in designs should be easily communicated to
the shop floor. Many existing designs cannot be effectively automated if they were
designed for manual assembly.
Equipment for the factory of the future will reduce the need for some of the skilled
trades as we know them today. The motive on our part will not be to cope with the
shortage of skilled trades, although that already exists in some companies and in some
geographical areas. It will not be to solve the problem of getting young individuals, who
can enter other areas of work for more money, to go into relatively low-paying jobs in
manufacturing. The motive will be the old one: productivity and bottom-line cost.
Automated technology can achieve greater productivity, better quality, reduced
costs, and increased profits—if it is properly applied. However, machinery installa-
tions or business systems may operate as designed but may not solve the ultimate
problem they were created to overcome.
Small details are the stumbling blocks to efficient automatic processing. This
follows from the need for precise control of the movement of the various parts going
into the assembly. There is no single, simple technique for obtaining control.

2.3 BASIC AUTOMATION CONCEPTS


One key rule of automation is that part design must be compatible with the needs
of automatic feeding. Parts are usually introduced to an assembly machine as bulk

DK565X_002_r04.indd 32 11/16/2005 1:29:55 PM


Assembly Automation 33

components. They are placed in a hopper and tracked to a loading station. Whether
the hopper is vibratory, rotary, or oscillatory, it relies on gravity or friction, or both,
for part movement. Some sort of gating or orientation device allows only those parts
in the proper attitude or position to enter the track.
Efficiency is affected by the system used; nonvibratory feeders, for example, are
limited in the number of orientations they can perform. Part geometry is a critical
factor. Soft parts may tangle in the hopper. Bowl driving forces may distort parts to
the point where orienting in the bowl is impossible. Distortion of parts due to stock-
ing and handling can also cause serious difficulties.
Parting line flash from cast and molded parts is an example of a problem that
never appears on a part drawing. The sensitivity of the part to moisture, static elec-
tricity, and residual magnetism may not become apparent until mechanical handling
is attempted. Sometimes a particular surface may be declared critical and must be
protected for subsequent operations. This situation may preclude or restrict use of
some of the automatic feeding methods.
Incomplete molded or broken parts reduce feeding efficiency. Foreign material
in the hopper is certain to alter the level of performance. This includes not only the
presence of material due to machine environment, but contamination of the parts
from previous processes.
Turning to the main part, or body, that will receive the oriented parts, several
questions arise. For example, the body must be strong enough to withstand assem-
bly tooling forces. Typically, forces for assembly are not as high as for machining,
but pressing, sizing, or machining operations might be required on the assembly
machine. High production rates may increase these forces.
The main part must be placed accurately on the assembly machine. This calls for
fixturing, flat locating surfaces, precision locating holes, etcetera. Parts are most effi-
ciently placed in the main body with simple, short, straight-line movement. Grasping
a part may be necessary. This requires some type of actuating force with its atten-
dant timing and control elements. If spearing or grasping the part is impractical, a
vacuum force may be used in transferring parts. Magnetic force is used occasionally
in handling parts, but this is not recommended, because the attraction of metal chips
and dirt degrades equipment performance rapidly. Also, residual magnetism often
cannot be tolerated in the final assembly. The author experienced problems with
the Dragon rocket motor assembly machine due to the cast retainers occasionally
becoming magnetized due to improper heat treatment of the stainless steel. We had
to degauss them all to achieve good hopper feed.

2.4 TYPES OF AUTOMATED ASSEMBLY MACHINES


2.4.1 Standard Machine Bases
There are essentially four types of standard base assembly machines:

Dial indexing machines


In-line machines

DK565X_002_r04.indd 33 11/16/2005 1:29:55 PM


34 Assembly Process: Finishing, Packaging, and Automation

Floating work platform machines


Continuous motion machines

To meet specific assembly automation needs, we add custom tooling. For maxi-
mum cost effectiveness, we can choose from a stock of standard operating stations such
as those for feeding, orienting, inspecting, and acceptance or rejection testing. Standard
base machines and operating stations are outgrowths of the industry’s experience in
designing, building, and using machines for high-volume production programs.

Dial Indexing Machines


The dial indexing automated assembly machines incorporate a mechanical drive that
rotates a circular dial table, or base plate, and indexes with a positive cam action. A cir-
cular, nonrotating table simultaneously raises and lowers a reciprocating upper tooling
plate, usually mounted in the center of the larger rotating base plate. Assembly nests are
installed around the outer edge of the dial table. Parts feeding, assembly, and inspection
stations are installed around or above the assembly nests or on the upper tooling plate.
These machines offer the following advantages:

Greater machine accessibility and minimum floor space. The basic circular
layout of dial-type machines is inherently more compact. High machine
accessibility increases operator efficiency and simplifies maintenance.
Greater adaptability to a variety of operations. The dial types of automated
assembly machines, containing central indexing mechanisms and recip-
rocating tooling plates, offer simplified rotary and up-and-down tooling
motions for high adaptability to many automated assembly operations.

Figure 2.1 shows an elementary rotary indexing machine. The dial type described
above has the added feature of a reciprocating tool table mounted above the indexing
table.

In-Line Machine
In-line automated assembly machines feature a rectangular chassis housing an indexing
mechanism driving an endless transfer chain. Nests that hold and transport the product
during the various assembly operations are fastened to the transfer mechanism. The
parts feeders, workstations, and inspection stations are then arranged along the work
flow. Parts are fed into the assembly nests as required, and work and inspection opera-
tions are performed in sequence along the length of the machine until the product is
completed.
These machines have the following advantages:

Unlimited number of workstations.


Efficient operator loading. The rectangular configurations permit machines to
be placed side by side with an aisle in between. The operator can efficiently
monitor all stations from the central aisle.

DK565X_002_r04.indd 34 11/16/2005 1:29:55 PM


Assembly Automation 35

Parts
feeder

Stationary
workhead

Work carriers

Indexing
table

FIGURE 2.1 Rotary indexing machine. (From


Boothroyd, G., Assembly Automation and Product Design,
Marcel Dekker, New York, 1992. With permission.)

Work can be performed from two or three directions simultaneously.


Figure 2.2 shows an in-line indexing machine.

Floating Work Platform Machines


In floating work platform machines, parts flow into a manifold where they are located,
assembled, and inspected. This system uses divergent flow channels for tandem or par-
allel operations to achieve line balancing and consists of two major elements—a parts
transporting element and a modular assembly element.
The parts transporting element moves the floating work platforms sequentially to
the various modular assembly elements. Each modular assembly element consists of
an independently powered unit containing one or more workstations. Use of a simple
transporting band permits flexibility within the system. Modular assembly elements
can be placed in remote areas such as cubicles, barricaded hazard rooms, holding or
curing rooms, or storage banks and can be returned to the main system by the parts
transportation element for further processing.

DK565X_002_r04.indd 35 11/16/2005 1:29:56 PM


36 Assembly Process: Finishing, Packaging, and Automation

Parts feeder

Stationary
workhead

Completed assembly

Work carriers
indexed

FIGURE 2.2 In-line indexing machine. (From Boothroyd,


G., Assembly Automation and Product Design, Marcel Dekker,
New York, 1992. With permission.)

Continuous band motion also permits the routing of parts onto a constant-motion
machine for nonstop performance, as well as for routing to in-line or dial-type
machines.
The quality of each assembly may be verified by inspection probes placed in
tandem following the work performance at each of the workstations. Stations may be
used for assembly function testing.
This type of machine has the following advantages:

Banks of parts may be accumulated between the workstations to cope with


short station stoppages.
Work can be removed from the system, performed at a hand station, and
returned to the machine.

Continuous-Motion Machines
Continuous-motion automated assembly machines provide for nonstop performance
of operations. Such systems may be capable of up to 1200 assembly operations per
minute. Parts are swept from a conveyor belt, oriented, and fed into the machine.
Following assembly, inspection, and function testing, the assemblies are oriented and
returned to the conveyor belt.
This type of machine has the advantage of higher production rates that can be
achieved with other types of bases. The free-transfer machine shown in Figure 2.3 shows

DK565X_002_r04.indd 36 11/16/2005 1:29:57 PM


Assembly Automation 37

Parts feeders

Stationary
workheads

Work carrier Partly completed assembly


transferring to next station

Buffer stock

FIGURE 2.3 In-line free-transfer machine. (From Boothroyd, G., Assembly


Automation and Product Design, Marcel Dekker, New York, 1992. With permission.)

a buffer position between the two workheads. The buffer parts could be shuttled off-line
to another work position, and then back to the main feed line, in either system.
In general, the higher the production rate, the lower the per-unit cost of a product.
Assembly automation systems are designed to fit production rates to specific needs.
Good machine design considers more than the production rate; it also considers
the overall production capability, including such factors as minimum maintenance,
system efficiency over years of continuous operation, minimum training required
for operators, and production of a consistently high-quality product. Flexibility is
important in every assembly automation system. Modularized workstations, idle sta-
tions, and standardized motions make systems adaptable to product changes with
minimum downtime.

2.4.2 Robots
A robot can be broadly defined as a machine that copies the function of a human being
in one respect or another. Industrial robots are generally equipment with a single arm,
and they are used to perform assembly-line operations and other repetitive tasks such
as feeding parts into another machine.

DK565X_002_r04.indd 37 11/16/2005 1:30:00 PM


38 Assembly Process: Finishing, Packaging, and Automation

The assembly machines we described in the previous paragraphs are designed to


handle large numbers of standard workpieces. These devices could be regarded as
the forerunners of the modern industrial robot, although unlike most robots, they are
controlled by the machine to which they are attached and may not be readily used
to perform another function. The cost of reconfiguring the grippers or end effectors
plus the reprogramming cost may make the robot less flexible in assignment than is
sometimes imagined.
The real differences may be in the eye of the beholder—or in the mind of the
reporter or author. The term robot comes from the Czech work robota, meaning work,
and was first used in a play called “Rossum’s Universal Robots” written in 1920 by the
Czech author Karel Capec. I personally find that the term robot to be rather imprecise
and one that should normally be replaced by a more descriptive term. Perhaps the best
definition is by example. A robot consists of three basic assemblies:

1. Motion system
2. Controller system
3. Heads and work tools

2.5 MOTION SYSTEMS


The motion systems start with the basic two-axis linear-rotary models as sketched
in Figure 2.4. These high-speed, servo-controlled robot systems are designed for a
variety of material handling and pick-and-place operations. The robot consists of a
high-speed linear table, rotational table, arm, and work tool to provide high-speed
positioning over a large work area. An optional Z-axis could be added (between the
arm and -axis) to permit raising or lowering parts from one height to another.
Providing three axes of motion (X, Y, Z), the robots shown in Figure 2.5 can per-
form various functions within a large volume. High positional accuracy of the X–Y–Z

O X

FIGURE 2.4 Two-axis motion system (X, R).

DK565X_002_r04.indd 38 11/16/2005 1:30:01 PM


Assembly Automation 39

tables make this robot ideally suited for intricate tasks, such as stuffing components
onto PC boards. The Z-axis capability permits components to be selected and posi-
tioned at various vertical levels.
The basic X–Y–Z rotary model is shown in Figure 2.6. These robots, featur-
ing X–Y–Z– motion plus a special work turret, offer exceptional versatility for
robotic applications. The X–Y–Z– motion permits work to be performed on a three-
dimensional workpiece from the top as well as two sides. The turret may contain
an electrooptical sensor for position location and autocentering. For performing
multiple tasks, other turret positions may include a variety of tools, such as a screw-
driver, Allen wrench, and grippers.
Extended-work-area robots, featuring five axes of motion (X, Y, Z, 1, 2), provide
ultimate flexibility in servicing a 360° area around the robot in a large volume. When
employing the same turret as the four-axis models, these robots can be programmed
to provide a single robot assembly station with multiple task capability. The robot
may also be placed to service an assembly line on either side of it. See Figure 2.7 for
a schematic.
Robots (and assembly machines) today are normally designed by assembling
standard components and assemblies to fit specific needs. The X–Y tables can position
parts to 0.001 in., at speeds as low as 0.001 in./sec or as high as 60 in./sec with
accelerations to 4g.
The controls may be a programmable positioning controller consisting of a
microprocessor-based computer numerical control (CNC) system with built-in
hardware calculators. This permits simple manual programming via a keyboard.
Program entry may also be from floppy disk or direct from a central or local area
network (LAN) computer. Simultaneous interpolation up to six axes linear and three
axes circular are becoming standard. The controller will accept positional feedback

X Y

FIGURE 2.5 Three-axis motion system (X, Y, Z ).

DK565X_002_r04.indd 39 11/16/2005 1:30:03 PM


40 Assembly Process: Finishing, Packaging, and Automation

from optical encoders, interferometers, inductsyns, and resolvers. Transducer out-


puts, such as pressure, torque, and power levels, may also be input. Other input/
output functions available are usable in tuning operations, such as tuning a coil
to its peak, or adjusting a potentiometer to a desired output level. Here the work
tool motion is controlled by a feedback signal to perform the adjustment to the
desired device output signal. Most robots today also have a teach pendant, which
can be moved manually through the desired positions to program the controller or

Z
Y
X u

FIGURE 2.6 Four-axis motion system (X, Y, Z, ).

Z
Y
u1
X u2

FIGURE 2.7 Five-axis motion system (X, Y, Z, 1, 2).

DK565X_002_r04.indd 40 11/16/2005 1:30:08 PM


Assembly Automation 41

computer. The program generated by a teach pendant can then be printed out, and
minor reprogramming steps can be added to smooth the motions.
Figure 2.8 shows an unmanned robotic cell in a flexible manufacturing system,
built around robots that handle parts and material, perform assembly tasks, oper-
ate tools, and perform other manufacturing operations. This cell is built around a
Cincinnati Milicron T3 material handling robot. The robot controller, the material
handling system and its controller, the fine-resolution system and its controller, and
the automatic riveter and its controller receive signals through a multiplexer from a
central DEC PDP-11/34 computer.
In designing a robotic system for manufacturing, the flexible system concept
described above should be the base on which the largely unmanned robotic system
should be built. Figure 2.9 shows the floor layout for an application developed by the
National Institute of Standards and Technology in Gaithersburg, Maryland. A deburring
cell has been developed, which demonstrates the deburring of parts based on their
description and a graphically developed process plan. This cell, a cleaning and deburr-
ing workstation, consists of a workstation controller, two robots, various quick-change
deburring tools, a rotary vise for part fixturing, and a part transfer station.
The working head may include a turret, similar to a lathe. The variety of mechanical
grippers, vacuum grippers, screwdrivers, nut runners, and so forth, is almost endless.
The type most similar to a human hand, with multiple linkages and force feedback

General Electric FRS Fine Resolution


TN 2200 ctor Controller System (FRS)
End Effe
Solid State System
Array Camera

End Effector
Controller
Cincinnati
Milacran
T3 Robot

Functional
Functional CNC RC VER2 Sensors
Sensors Robot
Controller
Material GEMCOR G-400B
DEC PDP-11/03 Automatic
Handling
Microcomputer Riveter
System

Material Precautionary Riveter


Handling Sensor Controller
Controller System

MUX
Fiber Optics (Typ)

To
DEC PDP-11/34A

FIGURE 2.8 Manufacturing cell schematic. (Courtesy of McDonnell Douglas Corp.)

DK565X_002_r04.indd 41 11/16/2005 1:30:14 PM


42 Assembly Process: Finishing, Packaging, and Automation

FIGURE 2.9 Cleaning and deburring workstation utilizing robots. (Courtesy of U.S.
Department of Commerce.)

servos, is often pictured in the press as standard, but in fact it is very expensive and
more commonly seen as a laboratory experiment than a factory working tool.

2.5.1 Inspection and Testing during Automated Assembly


Assurance that parts are within design tolerance before they are assembled, that parts
are assembled correctly, and that the product works reliably help make assembly
automation less expensive and more reliable than hand assembly. The goal is to build
quality into the product rather than just inspecting defects out. Experience has shown
that product quality improves through automation. Not only does the per-unit cost
decrease, the percentage of defective units also decreases with automation, reducing
total costs even more. Generally, a machine will turn out a good part every time—or
reject a faulty assembly—or stop.

DK565X_002_r04.indd 42 11/16/2005 1:30:15 PM


Assembly Automation 43

A machine can include unique memory systems for product quality inspection
and function testing. The assembly automation system knows design tolerances
and assembly processes. It can make certain that the product works the way it was
designed, by rejecting defective units.
A typical assembly automation system with inspection capability performs the
following operations:

Checks the presence and position of internal parts in an assembly.


Rejects defective units automatically rejected in salvable form.
Sense, through probe stations, the condition of the product at critical stages in
the assembly process. If required, component dimensions can be checked
100% immediately prior to or following assembly.
In the event of missing, surplus, or improperly oriented components or assem-
blies, the machine will either reject the defective part for salvage or stop until
it receives corrective action by the operator. A light on the control panel indi-
cates the nest location of any problems.

The machines are also capable of testing products as a final acceptance or rejec-
tion decision. A simple example, which was installed at McDonnell Douglas by the
author, torqued two parts together and then measured the total height of the finished
assembly. This was a difficult design to work with, because the correct torque and the
correct height were both considered critical to the performance of the product. Had
redesign been possible at the time, one of the criteria should have been dominant. One
of the most complex functional test examples was observed at the Austin, Texas, plant
of IBM, where they assemble laptop computers. The completed computers are trans-
ferred to the test area via a powered transfer line that has slack built into the system to
act as a buffer prior to test. The following operations should then be performed:

Open the computer lid (containing the screen).


Remove the cardboard floppy disk that is put in place for protection during
shipping.
Pick up and insert a test diskette, and close the top.
Plug a test cable into the back of the machine.
Transfer the unit to the powered conveyor that carries the computer through
the test process.
The computer then proceeds through a complex multilevel module that runs a
continuous functional test while applying heat and vibration for 24 hours.
As the computer leaves the test module, it removes the test diskette, reinserts
the cardboard shipping diskette, unplugs the test cable, and closes the top.
If the test shows an anomaly, a description of the test failure is printed out, as
well as the instructions for which component should be repaired or replaced.
The failed computer is then shunted to a rework and repair station, where a
mechanic accomplishes the required repairs and reenters it into the test cycle.
The good computers (which are nearly all of them) continue on down the line
for packaging and so forth.

DK565X_002_r04.indd 43 11/16/2005 1:30:15 PM


44 Assembly Process: Finishing, Packaging, and Automation

2.5.2 Machine/Operator Relationships


Because the repetitive assembly operations are performed automatically, operator
fatigue errors are practically nonexistent. The assembly automation system moni-
tors its own function through its memory system so that personnel can be quickly
and easily trained as operators. Regardless of the operator’s experience, system effi-
ciency and product quality remain at a high level.
Built-in sensors detect and forewarn the operator to replenish the supply of parts
at each of the automated feeding stations, thus assuring uninterrupted operation.
Techniques have been developed to assure safe machine operation and provide
operator protection. These techniques, including shields, barricades, and grounding
conductors, protect against personnel hazards and machine damage.

2.6 JUSTIFYING AUTOMATION


Calculating automation implementation payback is a subject that continues to gener-
ate controversy—and, more often than not, confusion—throughout the manufactur-
ing world. Moreover, the issue of how management will justify the vast expenditure
of automation against the total needs of the business is still one of the hottest
around.
The National Association of Electrical Manufacturers (NEMA) recently
reported the results of a survey on automation. Financial payback was classified
as very important by 91% of the respondents. What’s more, 78% of another survey
felt that “Most businesses in the United States will remain so tied to traditional
investment criteria that they will be unable to realistically evaluate the potential of
computer-aided manufacturing options.” In other words, businesspeople are using
formulas that include only traditional benefits to run the numbers, and as a result
they are procrastinating—they are not putting these technologies to work.
In the automation systems business, we have been developing new ben-
efit considerations—new manufacturing economics—that go beyond the limited
return-on-investment evaluations of the past to take into account the total impact of
automation on the entire business. This is because automation technologies do not
change the factory floor; they change the entire business in new and important ways.
The point is that businesspeople should be skeptical of conventional wisdom in
this area because old guidelines often do not work today. Automation investments
represent strategic decisions with far-reaching competitive implications and should
be evaluated as such. However, there still are financial calculations that can be made
to validate the appropriateness of the investment.
So far we have talked about financial justification with words such as traditional,
limited, and conventional. It’s interesting that under the so-called enlightened con-
cept of financial justification—which is outlined below—we are learning there really
is no black magic involved—no voodoo economics.
In fact, we still use many of the evaluation techniques that have been around for
years. We still try to estimate a net present value and a payback period—these are
basic decision criteria. In addition, discounted rate of return—some people call it

DK565X_002_r04.indd 44 11/16/2005 1:30:15 PM


Assembly Automation 45

internal rate of return—is still an important measurement. The only difference now
is that we have to expand our horizons and think more globally. This means that
we must project and incorporate, in our calculations, benefits that accrue business-
wide—benefits that previously would not be considered at all. It also means that the
long-term, macro view is the field of measurement, taking into account the benefits
of automation over the long haul.

2.6.1 Benefits
Let’s talk first about businesswide benefits. Up to now, whenever we evaluated the
advantages of a plant or equipment investment, we identified potential savings in
such terms as lower direct material costs, direct labor reductions, or improved indi-
vidual machine utilization. Now, even though these benefits can—and still do—come
from automation investments, we find that they pale by comparison to savings being
generated in other areas. For example, we can substantially reduce indirect manufac-
turing expenses for such things as fork trucks, fork truck drivers, material handlers,
and employees in packing and shipping. Further, automation frequently reduces the
number of machines required in a factory, and a direct result of this is a reduction
in maintenance and perishable tooling costs. At the McDonnell Douglas Missile
Assembly Plant, we saw a reduction in the overhead areas of operating supplies and
hazardous waste disposal by reducing the so-called touch labor in the SMAW mis-
sile encasement department with the implementation of our semiautomated braiding
system. This is real savings.
Fewer machines also means that less floor space is required. This, in turn, drives
down such related expenses as heat, light, taxes, and insurance.
One of the strongest leverage factors we count on today is quality costs. The positive
impact that automation can have on quality will most assuredly affect the costs associ-
ated with scrap, rework, warranty claims, quality inspectors, and after-sales service. The
ounce of prevention that improved quality builds into our products can therefore lead to
across-the-board reductions in these areas—well worth many pounds of cure.
Up to now, we have been talking about lowering costs and expenses that appear
on the operating statement. It is also important to leverage the balance sheet. One of
the largest single items on any balance sheet is inventory. Therefore, to the extent that
inventory can be reduced, cash is made available for other purposes.
We have learned that automation in general—and computer-integrated manu-
facturing systems in particular—reduce plantwide manufacturing cycles. These
reductions usually free up large amounts of cash—cash that previously was tied up
in inventory or work in process (WIP). This, in effect, becomes the down payment
for automation.
The payoffs we have talked about so far are quite tangible and easily measured.
We call these the ripple effects of automation. Ripple effects are bonus points, if you
will: effects such as higher productivity from increased worker satisfaction, higher
market share from improved quality, higher market share from faster response time
to shifts in marketplace demand. Though these advantages are difficult to measure,
they are even more difficult to forecast.

DK565X_002_r04.indd 45 11/16/2005 1:30:16 PM


46 Assembly Process: Finishing, Packaging, and Automation

All these benefits are more than theoretical: they are empirical. They are results
that really can be and have been achieved by many companies. General Electric
(GE) experienced this initially at one of its own facilities, the Locomotive Products
Division in Erie, Pennsylvania. Over several years, GE implemented a large, mul-
tiphased project that covered a broad spectrum of automation technologies. The
benefits from this project were realized across several areas. Impressive savings
were generated in direct material and direct labor costs, while equally important
results were obtained in increased capacity, higher product quality, and reduced cash
lockup in inventory. These benefits not only leveraged earnings, but also established
the foundation for survival in tougher times—the smaller, more competitive world
market the locomotive division is dealing with today.
Let’s explore some of the disciplines for using this information in the justifica-
tion process. Before we start, a rule: We cannot demand large short-term paybacks.
The payback may come in phase three of a three-phase project. History has proven
that synergy really does come into play here: When all the phases are complete, the
total system benefit is larger than the sum of its individual parts.
The payback computation requires two sets of data. These take the form of a
5- or 10-year pro forma cash flow for the business. The pro forma should be com-
piled assuming (1) the automation alternative is undertaken, and (2) a do-nothing
or as-is scenario, which requires a candid answer to a very tough question: “What
will happen to my competitive position if we don’t automate?” Often, the truth will
not reflect the status quo. Instead, it may include a possible erosion of market share
because of the likelihood that one’s competitors will not stand still. In this case, the
competitor’s lower cost structure and higher quality product will allow it to gain
market share.
Once the cash flow has been determined for these two alternative scenarios,
the cash flow difference between the two should be used in the automation payoff
calculations. For example, a corporation is evaluating whether or not to proceed
with a $3 million automation project (see Table 2.1). For starters, we have assumed
that the status quo will prevail if the project is not adopted. However, if the proj-
ect is undertaken, the income will increase $200,000 each year—and cash flow
will increase $2 million in the first year because of inventory reduction, and half a
million each year thereafter. The key now is to plug the difference in cash flow into
the financial calculations.
The 10-year rate of return on the investment is 24%. If we had mistakenly
considered only the operating cost efficiencies and overlooked the inventory reduc-
tion, the return would have come out a meager 2%—with the result being project
disapproval.
Consider another case. Let’s assume a 5% erosion of market share in a
$100 million market if the project is not undertaken. The resulting decrease in sales,
net income, and cash flow now magnifies the benefit of automation and generates
a discounted rate of return of 62% (Table 2.1). So the lesson to be learned is that
inventory reductions and share gains or losses usually have dramatic leverage on the
payoff to be earned.

DK565X_002_r04.indd 46 11/16/2005 1:30:16 PM


Assembly Automation 47

TABLE 2.1 Ten-Year Pro-Forma Cash Flow Analysis of Operating Results of


a Product Line with and without a $3 Million Investment in Automation.
(a) Considers traditional and nontraditional benefits, with no loss of market
share without automation. (b) Considers traditional and nontraditional
benefits, and assumes a 5% loss of market share without automation.
(c) Considers traditional benefits only, and no loss in market share without
automation.
TRADITIONAL BENEFITS
Ten Year Pro Forma
Direct Material
Operating Results (Millions) Direct Labor
Machine Utilization
No Automation $3.0 Automation

Sales/yr. $25.0 $25.0

Net Income/yr. $ 2.8 $ 3.0 NONTRADITIONAL BENEFITS

Less Inventory
Cash Flow/yr. $ 3.0 $ 5.0 (1st yr.)
Indirect Cost Reductions
$ 3.5 (yr. 2–10) Fewer Machines
Less Floor Space
(a) Lower Costs from Higher Quality

Ten Year Pro Forma Sensitivity Analysis

Operating Results (Millions) Operating Results (Millions)


No Automation $3.0 Automation No Automation $3.0 Automation

Sales/yr. $20.0 $25.0 Sales/yr. $25.0 $25.0


Net Income/yr. $ 2.0 $ 3.0 Net Income/yr. $ 2.8 $ 3.0

Cash Flow/yr. $ 3.0 $ 5.0 (1st yr.) Cash Flow/yr. $ 3.0 $ 3.5
$ 3.5 (yr. 2–10)
(Project DCRR 62%) (Project DCRR 11%)
(b) (c)

With regard to the ripple effects we discussed above, we should deal with these
softer measurements by means of a sensitivity analysis. Returning to our exam-
ple, let’s change it slightly by assuming no $2 million inventory reduction. The
discounted rate of return falls to 11% (Table 2.1). Suppose management has set a
hurdle rate of 20%. That means the discounted rate of return must be at 20% before
the project can be approved. We can easily determine that an additional $200,000
of cash would be needed each year if we wanted a 20% return—and this, in turn,
translates into one additional point of market share. Management can then make a
gut feel decision as to whether or not this gain in share is feasible—and whether or
not to proceed.

DK565X_002_r04.indd 47 11/16/2005 1:30:16 PM


48 Assembly Process: Finishing, Packaging, and Automation

2.7 SOFTWARE INTERFACES TO ASSEMBLY


AUTOMATION
Manufacturing automation needs to solve many technical problems before it is
accepted by the majority of manufacturers. The federal government has put our tax
dollars to work on clearing these barriers.
Chief among the efforts at the Automated Manufacturing Research Facilities
(AMRF) is an attempt to link disparate computers, robots, and machine tools
in a seamless network. At stake, says Howard M. Bloom, chief of the Factory
Automation Systems Division, which contributes personnel to the AMRF, is the
acceptance of computer-integrated manufacturing (CIM). “What we do here will
have a dramatic effect on how manufacturers choose systems and modernize their
factories,” he says.
If CIM made all the elements in a production facility, in effect, plug compatible,
even the smallest factories and machine shops in the United States could build the
factory of the future little by little. As yet, only a handful of large American corpora-
tions with substantial resources have started automating their production facilities.
Studies commissioned by the National Bureau of Standards (NBS), where AMRF is
housed, have shown that about 90% of the discrete-parts producers, those who make
parts in small batches, are organizations with fewer than 50 employees. At the same
time, they are responsible for about three fourths of the trade in manufactured goods,
and many of them subcontract from major corporations.
You can buy a computer from IBM and an operating system from another
vendor because IBM has provided specifications for interfacing. You can also buy
languages from different vendors because well-defined interfaces allow compilers
to match the operating system. The situation is different with factory systems. You
cannot readily buy software or a factory system with well-defined interfaces. Even
when you can, the interface may be a unique, internal interface that will not allow
you to connect equipment from another vendor. To modernize factories today, you
often replace everything. Moreover, factory interfaces call for a lot of wiring, and
interfaces consume 30–50% of expenditures for automation.
AMRF is developing software that will translate data so they can be understood
by robots, machine tools, and sensors anywhere in the facility. To ensure practical
results, AMRF has chosen to work with many commonly available commercial com-
puters and software packages. These include nine kinds of processors: the IBM 4341;
Digital Equipment Corp.’s VAX 11/780; Hewlett-Packard Co.’s 9000/9920/9836; Sun
Microsystems’ line of Sun workstations; the 4404 AI Workstation from Tektronix;
Symbolics LISP Machine; and Iris from Silicon Graphics. They also include nine
computer languages in various versions: C, Pascal, LISP, Prolog, Fortran, BASIC,
Assembly, Forth, and Praxis.
To control the machine tools, robots, sensors, and other equipment, AMRF
personnel have designed control system software that runs on IBM PC-compatible
micros and micros based on Motorola 68000 microprocessors. For robot control, the
software uses Forth and LISP because they are transportable, easy to use, and efficient
as interpreters and operating languages, says Bloom. C is used for programming the

DK565X_002_r04.indd 48 11/16/2005 1:30:16 PM


Assembly Automation 49

Integrated Manufacturing Database Administration System (IMDAS), which AMRF


developed.

2.8 DESIGN FOR AUTOMATED PRODUCTION


The human operator’s hands have infinite control and flexibility to compensate for
the many variations encountered in manufacturing processes. The automatic assem-
bly machine can produce quality and quantity of parts only under strict, inflexible
parameters. The ability to adapt automation to fabrication or assembly depends
largely on the detail design of the individual elements and the complexity—or lack
of it—in their assembly into the end product. The objective of this section is to pres-
ent some of the basic principles of piece-part design that will enable products to be
assembled automatically with a minimum of difficulty.
Parts are usually introduced to the assembly machine as bulk components. They
are placed in a hopper and tracked to a loading station. Whether the hopper is vibra-
tory, rotary, or oscillatory, it relies on gravity, friction, or both, for part movement.
Some sort of gating or orientation device allows only those parts in the proper attitude
or position to enter the track.
Efficiency is affected by the system used; nonvibratory feeders, for example, are
limited in the number of orientations they can perform. Part geometry is a critical
factor. Soft parts may tangle in the hopper. Bowl driving forces may distort parts to
the point where orienting in the bowl is impossible. Distortion of parts due to stack-
ing and handling can also cause serious difficulty. Parting line flash from cast and
molded parts is an example of a problem that never appears on a print. The sensitivity
of the part to moisture, static electricity, and residual magnetism may not become
apparent until mechanical handling is attempted. Sometimes a particular surface
may be declared critical and must be protected for subsequent operations. This situa-
tion may preclude or restrict use of automatic feeding methods. Use of symmetry in
part design makes orientation either unnecessary (ball bearings, etc.) or very simple
(plain rods, disks, etc.). Asymmetrical parts can sometimes be made symmetrical,
and the added manufacturing cost is often insignificant when compared to orientation
and sensing costs involved during assembly (Figure 2.10).
Parts can be designed with distinct polar properties. Asymmetrical parts can be
oriented by geometry, weight, or both. Obvious orientation features are exploited in
vibratory feeders, as shown in Figure 2.11. The example shows drilling a hole in one
end of the part to make it off-balance to aid in orientation in the feeder and the addi-
tion of a flat to aid in sensing the hole position. Proper application of gravitational
and frictional forces will move parts from the bowl onto the rising spiral track in
random orientation. Improperly oriented parts passing sensing stations on the track
are forced off the track and back into the feeder bowl, leaving only parts in the
desired orientation moving into the pickup station.
Vibrating action in the feeder provides the forces needed to orient the piece parts.
Track design features that detect shape or weight differences in parts to be rejected
from the track are shown in Figure 2.12. If rejected by gravity, they fall off the track
because of the weight distribution. If rejected by shape, removal is accomplished

DK565X_002_r04.indd 49 11/16/2005 1:30:16 PM


50 Assembly Process: Finishing, Packaging, and Automation

ASYMMETRICAL SYMMETRICAL
Asymmetrical parts
can often be redesigned as
symmetrical to facilitate
assembly. The same tooling
could make two cutoffs,
grooves, chambers etc. as
easily as one on most parts.
Symmetry is very important ASYMMETRICAL SYMMETRICAL
where off center grooves or
ridges would make orientation
difficult and expensive.
SYMMETRICAL
ASYMMETRICAL

OR

FIGURE 2.10 Examples of symmetrical versus asymmetrical parts.

NONFUNCTIONAL
FEATURE

(a) (b)

External feature, a flat, is added for sensing hold position


(a) or the center of gravity is relocated by drilling a hole in one end of
a part (b).

FIGURE 2.11 Addition of nonfunctional features to aid in feeding.

with cutouts in the track cams along the track surface, or sensing feeders. Parts are
then rejected by air jets or mechanical actions.
The following techniques simplify, improve, and in some cases make feeder
orientation possible.

Minimize the number of different orientations. Vibratory feed rates depend


on successful orientation. Sensing features may be repeated on a part to
improve the probability of correct orientation (Figure 2.13). For instance, if
the track feed rate is 600 pieces/min and the orientation probability is only 1
in 10, we are effectively limited to 60 pieces/min for assembly. The example
part, a cube, is modified to increase orienting probability from 1 in 24 to 1
in 6, which will increase feeder output by a factor of 4.
Avoid parts that tangle. Springs are probably the best example. For instance,
a simple open-coil spring may be redesigned with close-wound coils at both

DK565X_002_r04.indd 50 11/16/2005 1:30:18 PM


Assembly Automation 51

Flats on the sides make it


Difficult to orient with much easier to orient with
respect to small holes respect to the small holes

No feature sufficicently When correctly oriented


significant for orientation will hang from rail

Triangular shape of Nonfunctional shoulder permits


part makes automatic proper orientation to be established
hole orientation in a vibratory feeder and maintained
difficult in transport rails

FIGURE 2.12 Examples of features that aid in


feeding. (From Boothroyd, G., Assembly Auto-
mation and Product Design, Marcel Dekker,
New York, 1992.)

Cube with off center hole


(a) has a 1 in 24 probability of correct
orientaion. On center hole (b) has 1 in 5
probability. 100% probability of correct
orientation is obtained with centered holes
(a) (b)
on all faces or no holes at all.

Sphere (a) modified to bead


shape (b) now has an effective short axis
that aids orientaiton.
(a) (b)

FIGURE 2.13 Additional features that aid in part orientation.

ends and in the center. This design often permits handling by conventional
hoppers and feed systems. Springs may also be fabricated in the automatic
assembly machine to avoid this problem. Slotted parts may also be
redesigned to minimize tangling, as shown in Figure 2.14.
Avoid using parts that nest, shingle, or climb. Matching tapered surfaces
will cause parts to lock or nest. Internal ribs, projections, or increasing the

DK565X_002_r04.indd 51 11/16/2005 1:30:20 PM


52 Assembly Process: Finishing, Packaging, and Automation

angle well above the locking angle will avoid the problem. Avoid shingling/
climbing parts. Parts clinging onto other parts will often cause jamming in
feeder tracks. If anticipated, this potential malfunction in the feeder can
be avoided by redesign. Parts tend to climb over thin or shallow beveled
contact surfaces. (Figure 2.15).
Determine critical tolerances. It is essential to determine dimensions that are
critical for proper parts feeding and yet may be noncritical for piece-part
function if the parts feeding operation is to remain trouble free.

Straight slot will tangle. Crank slot will not tangle.

Closed-ended spring will


Open-ended spring will tangle.
tangle only under pressure.

FIGURE 2.14 Examples of part tangling. (From Tanner,


J. P., Manufacturing Engineering. Marcel Dekker,
New York, 1991. With permission.)

difficult to feed - parts overlap

easy to feed
FIGURE 2.15 Example of parts climbing or shingling. (From
Boothroyd, Assembly Automation and Product Design, Marcel Dekker,
New York, 1992. With permission.)

DK565X_002_r04.indd 52 11/16/2005 1:30:23 PM


Assembly Automation 53

Assume for now that parts are oriented and coming down the feeder track;
the job is now half done. There are, however, additional important design
considerations if parts are to be successfully integrated in a trouble-free
assembly operation.
Design parts for easy assembly. Design parts with guides for proper location
during assembly. Clearances should be carefully considered, and the liberal
use of chamfers and radii (Figure 2.16), will allow the parts to fall together
in place unassisted. Sufficient clearance should be allowed for feeder fin-
gers or other tooling needed to handle the parts.

We have not yet reached a point where we need a decision process for automate
or conventional build, comparable to make or buy, but it is coming. Meanwhile,
whenever the designer encounters a part or an assembly whose repetitive require-
ments to make or assemble exceeds 200,000 (project plan quantity times items per
end product), he or she should immediately confer with manufacturing engineering
to determine any anticipated plans to automate. When the quantity exceeds 1 million,
it is almost certain that automation must be used.
Data and books on automation are available in manufacturing and elsewhere.
These are illustrated and should be reviewed and studied by designers who encounter
parts and assemblies that can or should be automated.
When parts pose unusually difficult handling and orientation problems,
experimental mockup stations should be built to verify parts handling processes prior
to building the assembly machine. The product designer, the manufacturing engineer,
and the tool designer should work as a team through this development state.
Though each part or assembly presents a special problem in automation, some
general recommendations and requirements apply.

1. Parts should have at least one axis of symmetry, with ability to fabricate,
feed, and assembly in respect to the axis. A sphere is, of course, ideal.
2. Parts should be suitable for sorting and alignment for feeding on equip-
ment such as Syntron or Vibron unit.
3. Dimensioning should be such that inspection gauging can occur from one
reference point or axis and one plane of reference.

Old design New design

FIGURE 2.16 Design parts for ease of assembly. (From Boothroyd, G., Assembly Automation
and Product Design, Marcel Dekker, New York, 1992. With permission.)

DK565X_002_r04.indd 53 11/16/2005 1:30:23 PM


54 Assembly Process: Finishing, Packaging, and Automation

4. Parts being assembled automatically must be designed to engage with or


snap on to mating parts so that they will not shake loose during subsequent
assembly sequences. It is impractical to balance parts on one another when
using automated equipment.
5. Units being torqued together should have a shaped region such as a square
or hexagon to permit grasping and torquing.
6. Messy materials, such as lubricants and adhesives, create machine mainte-
nance and cleanup problems unless the parts receiving them are designed
to accept leak-free feed orifices with no-drip, point-of-application cutoff.
7. Semiautomated operations should always be considered whenever the
discriminatory capabilities of a human being will result in a significant
simplification of the automation equipment.
8. In designing for automatic assembly, the designer should ask the following
questions:
Can parts be made symmetrical to avoid orientation problems?
Do symmetrical parts have clearly defined polarity features?
Are the number of significant part orientations minimized?
Will parts tangle?
Will parts nest or interlock, thereby causing problems?
Will the part design cause shingling?
Are critical dimensions and tolerances clearly defined?
Do the parts lend themselves to easy location and assembly?
Does the design have a datum surface for accurate parts location during
assembly?
Have all unnecessary handling requirements, such as turning over parts,
assemblies, or both, been avoided?
Does the assembly have components that are buried or difficult to reach
and position?
Has the design been simplified and standardized as much as possible?
Have excessive burrs and flash been eliminated?
Can difficult-to-handle parts be assembled in an automatic system?

2.8.1 Machine/Operator Relationships


Because the repetitive assembly operations are performed automatically, operator
fatigue errors are practically nonexistent. The assembly automation system monitors
its own function through its memory system, so personnel can be quickly and easily
trained as operators. Regardless of the operator’s experience, system efficiency and
product quality remain at a high level.
Built-in sensors detect and forewarn the operator to replenish the supply of parts
at each of the automated feeding stations, thus assuring uninterrupted operation.
Techniques have been developed to assure safe machine operation and provide
operator protection. These techniques—including shields, barricades, and grounding
conductors—protect against personnel hazards and machine damage.

DK565X_002_r04.indd 54 11/16/2005 1:30:24 PM


Assembly Automation 55

We will find that we need a little psychology when our goal is to get the best per-
formance possible from an automatic assembly machine. No, not on the machine—
on the people who must interface with the machine to really make it work. One of
the psychological problems in getting superior performance can show up early in
the game. Companies that isolate line and operating personnel from the equipment
prior to its delivery will be confronted by an extended learning curve in bringing the
machine on line. Among the shop people there will be an explicit or implicit attitude
that “Tooling bought the machine, let them make it run.”
What’s the solution? Management must get line supervision, machine opera-
tors, and maintenance personnel involved in the final construction and debugging
of the machine. It is not expensive. When operating, an assembly machine may
return $100 or more per hour; one day’s operation in the future will more than
cover these costs.
Another psychological problem in the introduction of automatic or semiauto-
matic machines is created by negative or skeptical attitudes on the part of manage-
ment. If operators believe that top management questions the technical feasibility
or operating capability of the machine, this will be reflected in the attitudes of line
personnel. The result will be a substantial delay in realizing the potential of the
equipment.
The solution lies in management understanding the debugging phase of machine
use. In trying to debug and refine an assembly machine, one is attempting to pull
oneself up by the bootstraps. The first plateau of productivity is reached relatively
easily. Each subsequent plateau generally is a lesser gain and is harder to realize.
For example, in the beginning, during rough debugging, very little attention is
paid to hopper inadequacies other than absolute failure. As machine development
continues, hopper performances that might be adequate in off-and-on machine
operation may be marginal under full production conditions. The hoppers might have
been developed using preproduction samples or without full exposure to the possible
variations that occur during production. Effort will be required to bring hopper per-
formance up to par.
To realize the full potential of automated equipment, it is important not to show
disappointment or dissatisfaction when problems become apparent. If the fundamental
machine design is sound, these problems will act as stepping stones to final produc-
tivity. They should be discussed and considered as advances in debugging, not as
setbacks.
We cannot overemphasize the importance of positive attitudes in the ultimate
success of the system: We must express these positive attitudes to those charged with
operating the equipment.
There is another hurdle to superior performance that has a psychological twist.
The general evidence of the problem is that the user becomes dissatisfied months after
a machine is put into production. Production goals are not being met, supplementary
hand lines must be maintained, and financial management is up in arms. The irate
user wants everything else dropped, and maintenance staff, engineers, managers—
anybody—to fix a machine that just doesn’t work. Usually, enough production

DK565X_002_r04.indd 55 11/16/2005 1:30:24 PM


56 Assembly Process: Finishing, Packaging, and Automation

has been run at the builder’s plant and the user’s plant that the machine truly was
production-worthy at some point. A general complaint about machine productivity,
rather than specific complaints, is typically evidence of either underanalysis or over-
analysis of problems.
For example, some people cannot control the urge to play with adjustments to
see if the machine can be made to work better. It is not at all uncommon to find every
conceivable screw, nut, knob, turnbuckle, and compliant link completely out of its
normal position. Or, when something goes wrong, instead of analyzing the problem,
someone immediately turns a screw or adjusts a locknut. No one ever remembers
where they were originally set, and things go from bad to worse. All that is required
to solve the problem is to return the machine to its original state.
The overanalysis syndrome stems from an inability of machine operators,
machine maintenance personnel, or operating supervisors to accept the limitations
inherent in any unique, special-purpose machine. There are limits to the tolerances
that can be imposed on component parts. There are limits to the amount of machine
development that is practical before turning a machine over for production use. What
is to be expected from such equipment is that the net production realized is within a
range that permits reasonable justification for the purchase of the equipment. Some
users appear to expect and aim for perfect performance. Every time the machine stops,
three or four people stand around discussing the situation, with no one attempting to
return the machine to operation.
Good assembly machines are always self-inspecting. They inspect for the proper
action of their own elements, and for the presence and position of the component parts
of the assembly being processed. Upon detecting some malfunction, they either shut
themselves down or switch to an alternate control mode. In any reasonably designed
machine, there are a few random self-clearing failures, but most failures to assemble a
product indicate a problem in the machine action, or jams caused by foreign material,
or improperly toleranced parts in the part feeders. Regardless of the control mode,
someone must correct the deficiencies and return the machine to operation.
There comes a point where continued analysis of the cause of such failures is
counterproductive. We must realize that such problems occur—regularly. We must
judge when the machine has reached a mature state of development—and turn from
debugging to running the machine and attempting to obtain the maximum production
results.
The problem is real, and the solution is a simple change in outlook. In several
instances, assembly machine servicepeople report they have completely cured such
a problem by offering operators a cash gift. In return, the operators were to stop
diagnosing the problems of that machine and attempt to operate it to meet a specific
production goal. In every single instance, this change in orientation and motivation
resulted in immediate production capacities that satisfied the customer.
It is interesting to note that most of the start-up problems discussed in this section
actually occurred at McDonnell Douglas when we introduced a propellant blanking
machine, a weight and skive machine, a propellant grain heating and rolling station,
an insertion machine for propellant scrolls into igniter cups, and rocket motor final
assembly machines on the Dragon missile high-rate production program. We learned

DK565X_002_r04.indd 56 11/16/2005 1:30:24 PM


Assembly Automation 57

a lot at that time, and after a very successful implementation program, produced over
3 million live rocket motors with very high reliability and at our projected cost. The
major change that solved most of our start-up problems was to decrease the machine
rate from a 4-second cycle to a 6-second cycle, which still met our production rates
and allowed machine maintenance and repair as needed. All machines have a speed
where everything works best. Any speed slower or faster than this “natural” rate
causes problems in feed, synchronization, or something. We found that the opera-
tors learned the sound of the machine in operation and were the first ones to suspect
an impending failure. One important fact came out of our experience with machine
design and maintenance. We had one machine that contained cams and gears, electri-
cal circuits, electronics, fluidic sensors, and computer controls. When the machine ran
perfectly, there was no problem. However, when a minor glitch occurred—whom did
you call for diagnostics and for repair? The machine had too many different systems
that required one or two engineers to make the diagnosis. The maintenance mechan-
ics were experts in only one or more of the systems, but not all. Our next machines
had fewer systems, and our supervisor of maintenance was a graduate engineer.

2.9 AUTOMATED MATERIAL HANDLING


2.9.1 Introduction to Automated Material Handling
Automation of a manual process or machine operation has been regarded as a step-
ping stone to gradually automating the entire factory operation. This is a common
approach, as it replaces the direct labor quite visibly and is entirely measurable.
However, automation of the material handling function is another matter. Because it
is a non-value-added function, not quite as visible as a significant cost contributor
to the product it supports, many companies have refrained from seeking innovative
approaches to automation of the material handling function. It also quite often requires
much higher capital outlays.
With the modern concepts of just-in-time, zero inventory balance, continuous-
flow manufacturing, and materials requirements planning (MRP), more emphasis is
now placed on eliminating the cost of material handling, or automating it to provide
low-cost, consistent, reliable material flow to the production areas. Essentially, an
automatic material handling system does just that—positions the parts or equipment
correctly, and at the right time for the next operation.

2.9.2 Characteristics of Material Handling Automation


Automated equipment is designed and built to fill a specific requirement. It is
necessary to define the need and coverage of the project. Following are some of
the characteristics that need consideration in order to define the automated material
handling equipment:

Definition of the beginning and end points served


Product configuration and physical changes occurring during the process
Production rate and fluctuation

DK565X_002_r04.indd 57 11/16/2005 1:30:24 PM


58 Assembly Process: Finishing, Packaging, and Automation

Quality and remedial activities in process


Process variation
Safety of workers around the equipment
Coordination with support functions and overall control desired
Avoidance of product damage and hazards due to breakdown in automated
system
Progressive time-phased plan for a large, capital-intensive installation
Alternative plans in case of prolonged breakdown of handling equipment

Advantages and Disadvantages of Automation


Automation provides many advantages over manual handling of material. Though
not all are cost–related, some of the intangible advantages provide valuable informa-
tion in making decisions. The advantages are the following:

Reduction of safety and health concerns of material handlers


Avoidance of production downtime with on-time deliveries
Reduced cost of the handling function
Avoidance of damage to material and equipment transported
Savings in floor space by continuously flowing material, rather than manual
batch transfer
Increased productivity, with automated transport lines acting as pacing devices
for progressive assembly operations

Obviously, there are also deterrents to automation. Some of them are as follows:
High installation cost
Reduced flexibility in schedule
Obsolescence and costly reconfiguration possible if process is altered
Major breakdown in system can cause shutdown of entire plant

2.9.3 Approaches to Material Handling Automation


There are two ways of approaching automation of the material handling function
in a production operation. One method is to automate portions of the plant progres-
sively, with a plan to interconnect them in the final phase. The other method is the
complete installation of automation equipment in the entire plant at one time. Both
approaches are widely used, based on the type of project and the commitment of
management.

Islands of Automation
An approach to total factory materials handling systems is through a progressive
implementation of stand-alone processes, a few at a time. These processes can involve
the use of numerical control (NC) or CNC machines, automatic tool changers, and
specialized conveyors to serve a few select machines and operations. In this case, the

DK565X_002_r04.indd 58 11/16/2005 1:30:24 PM


Assembly Automation 59

general material flow is converted into individual manufacturing cells with some degree
of automation. There are two important points in consideration of this approach:

1. Before implementing any such system, a completed overall plan is required


to make sure that the systems already implemented are compatible with
future islands of automation.
2. Because the overall objective is an effective materials handling system for
the entire plant, there should be a cohesiveness among all the subsystems
that leads to total integration.

Many companies fail to realize these requirements in the rush to implement.


Though they are able to justify the subsystems installed, they fail to realize the long-
term potential in terms of profit, quality, and longevity possible by integrating the
entire system. Sometimes it becomes difficult to justify the investment for equipment
to integrate subsystems already installed.
Islands of automation exist in plants with many separate processing depart-
ments. Figure 2.17 shows an example of this concept. A system can also be easily
expanded to include prior operations, or operations between existing islands.

WORK
MONORAIL SYSTEM
RACKS STATIONS
OVERHEAD CONVEYOR SYSTEM
LATHE WORK STATIONS
#1
CART
LATHE

TANKS
LOAD
#2

STAGING/MASKING
TEST
CART
‘B’
‘C’
‘A’ IN OUT
IN
OUT TRANSPORTER
SYSTEM
WORK STATIONS
FUTURE
PARTS
HANDLING FUTURE CAROUSEL ‘D’
SYSTEM PARTS
1995 HANDLING
‘A’ TO ‘B’ FUTURE
SYSTEM
PARTS
1996
HANDLING
‘B’ TO ‘C’
SYSTEM
1996
‘C’ TO ‘D’

PRODUCT FLOW

‘A’ : AUTOMATED ‘B’ : CHEMICAL ‘C’ : SUBASSEMBLIES ‘D’ : FINAL ASSEMBLY


TURNING
TREATMENT AND TEST
CENTER

FIGURE 2.17 Islands of automation. Departments A, B, C, and D are considered


islands for product flow inside the departments. They are to be interconnected by a
materials handling system. The advantage is in spreading out investment costs over
the years. However, individual systems must be compatible with the interconnecting
systems of conveyance and batch processing requirements. (From Sheth, V., Facilities
Planning and Materials Handling, Marcel Dekker, New York, 1995. With permission.)

DK565X_002_r04.indd 59 11/16/2005 1:30:25 PM


60 Assembly Process: Finishing, Packaging, and Automation

Product configuration and its development through each stage of operation plays
an important part in determining the installation of this system.
The progressive installation of this system requires consideration of the
automation of only one process at a time. This approach is more easily understood
by management and financial executives than the installation of a complex, plant-
wide materials handling system that carries a much higher price tag. Another advan-
tage is that the total materials handling project can be spread over several years, thus
causing less anxiety in lean times and reducing the up-front financial burden.
The progressive automation technique is useful when the following conditions
are present:

Product design is not fully matured


Process improvements are likely to occur in the near future due to technological
advances—in either the product, or in factory systems and equipment
Automation is difficult to justify as a whole, but phases can be justified
independently
Installations are conceived as an interim arrangement until final plans can be
implemented

Total System Installation


Many large, established corporations have considered it worthwhile to replace out-
dated facilities with new automated plants for products that have demand. The auto-
mobile and appliance industries fall into this category, where production is fairly
continuous for many years. Recent examples are the GM Saturn and IBM PC and
Proprinter automated facilities, which have replaced the manual material handling
functions because of the cost, quality, and production rate advantages. This approach
is sound and profitable provided the products and the processes are established. It also
requires heavy dependence on the automation supplier for quality and field support.
The use of computers and electronics in linking information flow has helped tremen-
dously in simplifying automation processes.

2.9.4 Storage Equipment for Material Handling


Automation
In the last two decades, more emphasis has been placed on the carrying cost of inven-
tory and handling systems for storage equipment, because these are the two highest
contributors to the cost of warehousing and storage in general. The final objective,
as both of these are non-value-added costs, is to totally eliminate them. Many new sys-
tems and equipment choices have been introduced. They have been further enhanced
by the computer technology now available to provide accuracy, flexibility, and effi-
ciency. Some of this equipment, now widely used in industry, is described below.

Automated Storage and Retrieval Systems


An automated storage and retrieval system (AS/RS) has two objectives. The first is
to faithfully support warehousing efforts. An AS/RS provides materials handling

DK565X_002_r04.indd 60 11/16/2005 1:30:25 PM


Assembly Automation 61

requirements with maximum efficiency, offering a unique and cost-effective way of


storing and handling materials. The second objective is to provide data in real time
to the cost collection and accounting function within the plant—or sometimes to
the home office, because parts inventory and issues are tracked in digital form. For
example, material issued has a contract or product code assigned along with the date
and quantity. This can be converted to a dollar figure and charged to the finished
product cost or contract. On some contracts, this may then be billed to the govern-
ment or other customer as WIP. Analysis of such data also is used to update pricing
data for bidding new contracts, because some of the items have common usage, or
are allocated as part of the overhead cost.
Quite often a minicomputer is used for AS/RS operation to feed these data into
the mainframe cost accounting computers. This means that networking involving a
language conversion is involved. If all elements in the computer system are using
one of the newer fourth-generation languages (4GL) and are compatible with open
computer architecture, the problems are minimized. However, trying to tie in with
older computers and languages can sometimes create serious problems in accuracy
and cause delays in the retrieval, delivery, and reorder of material.
Figure 2.18 shows an AS/RS system layout that is widely used. It consists of a
vertical stacking robot with a platform that travels up and down the aisle on a rail
with an extended arm for safe, smooth, and precise storage and retrieval of inven-
tory loads. On both sides of the aisle are racks full of bins, each assigned a particular
location value. At a computer command, the robot travels down the aisle, adjusts
to a specific height, and retrieves a bin from the rack. It then moves up the aisle
and places the load at the operator’s workstation or conveyor. Next, it moves to an

RACK WITH BINS

WORK

STATION

RETRIEVAL PICK

PLATFORM POSITION
WORK

WORK

STATION

RACK WITH BINS

FIGURE 2.18 Typical AS/RS layout for high output and improved inventory control.
(From Eastman, 1987. With permission.)

DK565X_002_r04.indd 61 11/16/2005 1:30:29 PM


62 Assembly Process: Finishing, Packaging, and Automation

adjacent station, picks up the previously delivered bin, and deposits it onto the rack
at its proper location. It begins the next cycle by picking another bin for the operator
while he is still getting material out of the bin that was just delivered. The computer
system tracks all possible storage locations and tells the robot where to store each
load. It makes the necessary updates to the inventory record every time the bin is
served. The system can be linked to a higher-level host computer for material plan-
ning and purchase order placement. Benefits of this system are the following:

Real-time inventory accuracy (often 100%)


Reduction in space requirements (as high as 80% reduction)
Faster order picking (up to four times the manual method)
Consolidation (makes off-site warehouses seem part of a central system)
Improvement of human factors (eliminates forklifts, climbing, bending,
lifting, etc.)
Increased pilferage control (controlled access to material by designated
employees)
Error-free operation (eliminates order picking errors)

Proper planning is required to install an AS/RS. Because most units are custom
made, including the computer software applications, it is necessary to do the following:

Make a complete list of parts to be stored and ensure that all aspects of the
AS/RS are sized correctly.
Make sure that the throughput of the AS/RS will meet your requirements.
Be specific about AS/RS computer and software requirements, if it is to be
interfaced with the network system.
Discuss performance, drawbacks, and potential problems with the users of a
similar operation. It is a major investment and downtime can paralyze the
entire plant operation.
Plan item placement carefully. Frequently used categories should be stored
in the front to reduce travel time. The number of picks per hour is greatly
dependent on travel time back and forth from the aisle to the workstation
and up and down the height of the racks. In most cases, the operator will be
idle waiting for the next load.
Load-test the system before implementing it to make sure that it is assembled
properly and can support the weight of the inventory stored.
Provide for a contingency plan in case the computer system breaks down or a load
causes a jam-up. Two- or three day-breaks in the operation have been experi-
enced with computer breakdowns. This may cripple an entire manufacturing
operation if the system is based on small batch sizes with short cycle times.
Installation and start up are major tasks. It is necessary to decide on its loca-
tion with the understanding that a future move for the AS/RS will be costly
and cause production downtime.

While the AS/RS is used to store and retrieve loads on pallets and in containers,
there is another version that is commonly known as the miniload system for

DK565X_002_r04.indd 62 11/16/2005 1:30:29 PM


Assembly Automation 63

handling and storing small parts in pans, drawers, and toteboxes. This system can
efficiently automate small parts storage and handling of raw materials, finished
products, kitting, and WIP materials. The working principle is the same, but it often
requires the workstation layout to suit a particular application based on whether it
is a kitting operation, production requirement, or warehousing operation.
Another variation of AS/RS is the application for high-density storage, in which
the robot’s extendable arm is replaced with a shuttle car. This car travels on tracks
underneath the loads supported on the rack. With this arrangement, it is possible
to serve unit loads on the racks deeper than one unit, thereby reducing the number
of aisles and robots required to serve the system. This system provides maximum
building cube utilization and maximum storage density.

Carousel Storage and Retrieval System


The carousel storage and retrieval system (CS/RS) has been in use for years in the
garment industry. We have seen a moving garment conveyor application in dry clean-
ing shops, where a clerk activates a power conveyor supporting the garments to bring
the order to the front counter. It is only in the last few years that this principle has
been extended to industrial facilities for controlling inventory and handling costs.
The addition of a programmable controller is useful in most cases.
Carousels are either horizontal or vertical. A horizontal carousel moves the prod-
uct horizontally for use at floor level. In a vertical carousel, the product moves in a
vertical loop and often can be serviced from more that one floor level.
A horizontal carousel is a closed-loop storage system comprised of a long, narrow
framework supporting vertically placed storage baskets, which rotate on a horizontal
plane. The baskets are connected by a common power drive system so all the bas-
kets move at the same time as seen in Figure 2.19. The operator at the workstation
activates the control to bring the desired basket to the workstation. Various applica-
tions of the horizontal carousels are shown in Figure 2.20. Many types of products
may be stored with this system, including hardware, electrical parts, appliances, and
subassemblies. It can be used in warehouse operations, burn-in for electrical equip-
ment, assembly-line operations, kitting, and WIP storage, among others.
Variations of the horizontal carousel include a twin-bin system, which allows
an operator to pick parts from the sides of two adjacent bin columns instead of from
only the front of the bin. A slightly different version is the rotary rack, in which each
of the tiers rotates independently of the others. In this way, two tiers can be simulta-
neously rotated to a desired location for assembly-line operation. This system can be
further enhanced by automating the storage and pick-up functions with the use of a
fixed-location robot with an extending arm. This concept is useful in supporting parts
requirements for a flexible assembly workstation.
A vertical carousel, also know as a vertical storage and retrieval system (VS/RS),
operates much the same as the horizontal system, but in vertical rotation as shown in
Figure 2.21. It has the following features:

Allows full use of the building cube


Provides flexibility of storage

DK565X_002_r04.indd 63 11/16/2005 1:30:30 PM


64 Assembly Process: Finishing, Packaging, and Automation

FIGURE 2.19 Horizontal carousel for random storage with computerized inventory and
in-process control. (From Sheth, V., Facilities Planning and Materials Handling, Marcel
Dekker, New York, 1995.)

Reduces downtime, delays, and damage to products


Allows for small batch sizes
Reduces set-up time (in countertype set up, such as tool cribs)

Figure 2.22 shows vertical carousel layouts for a variety of situations.

2.9.5 Transportation Equipment


Transportation equipment for materials automation consists mainly of transporters,
automatic guided vehicle systems, and power and free conveyor systems.

DK565X_002_r04.indd 64 11/16/2005 1:30:30 PM


Assembly Automation 65

Order picking
Shipping objectives are met by
freeing operators to do more
work in less time for unsur-
passed productivity!

Buffer storage
Provides instant adjustment to
peaks and valleys in produc-
tion processes.

Burn-in and Test


ACCUTRIEVETM provides
efficient and effective burn-in
testing for electronics Order Picking
manufacturing.
10% TO
Records Storage
BUFFER STORAGE
Combines the space-saving 90% TO
advantages of high-density ASSEMBLY
archival storage with fast, easy
access.

Buffer Storage

Burn-in and Test

Records Storage

FIGURE 2.20 Various applications of horizontal carousels. (Courtesy of


Richards-Wilcox Manufacturing Company.)

DK565X_002_r04.indd 65 11/16/2005 1:30:32 PM


66 Assembly Process: Finishing, Packaging, and Automation

FIGURE 2.21 Vertical carousel. (From Sheth, V., Facilities Planning and
Materials Handling, Marcel Dekker, New York, 1995. With permission.)

Transporters
One commonly used derivative of a floor-supported belt or roller conveyor is a
transporter. It is a system comprised of sections of interconnected conveyors with
diversion gates leading to various assembly work stations. Figure 2.23 shows a
typical configuration of a transporter.
An effective manufacturing cell arrangement can be installed by combining a
transporter and a horizontal carousel. The dispatcher loads the transporter to deliver a
totebox from the carousel and signals the transporter to deliver the totebox to the avail-
able workstation. The power conveyor moves the load until the power-activated arm
diverts the load to the desired station. After completing the operation, the workstation

DK565X_002_r04.indd 66 11/16/2005 1:30:32 PM


Assembly Automation 67

Carousel in vertical design


Carousel with loading nose
in horizontal quadruple design

Carousel with loading nose


in vertical double-loop design

Carousel in multi-level construction


with 3 operating stations and loading noses
and vertical double-loop design

Carousel in multi-level construction


with loading nose
Carousel with loading nose
in vertical triple-loop design

FIGURE 2.22 Vertical carousels can be configured to take advantage of tight spaces
where operator or handling equipment access is difficult. (Courtesy of Baumann, Inc.)

operator transfers the totebox to the dispatcher through the bottom return conveyor,
which is traveling in the opposite direction. The entire system is computer controlled
to also keep track of the inventory count, work order status, and the storage informa-
tion on the carousel. The system is advantageous because it does the following:

Eliminates the line balance problem, because the dispatcher controls the WIP
inventory for every station on the line.

DK565X_002_r04.indd 67 11/16/2005 1:30:40 PM


68 Assembly Process: Finishing, Packaging, and Automation

FIGURE 2.23 The transporter provides flexibility in tasks perfomed at workstations.


It eliminates line balancing requirements since workstations are set independent of
one another for tasks performed. (Courtesy of Speedways Conveyors, Inc.)

Minimizes the supervision job of keeping workers busy, as they always


have work in front of them. The employee’s task at each station may
vary, depending on the schedule needs and experience of the individual
operators.
Flexibility exists to change product lines and to perform various product oper-
ations, without having to rebalance the line or reconfigure the conveyor.

DK565X_002_r04.indd 68 11/16/2005 1:30:41 PM


Assembly Automation 69

Schedule changes can be accommodated by adding more workstations for the


same operations.
Manual operations are a simple interface, and the system can also be used as
a buffer.

Automated Guided Vehicle System


The automated guided vehicle system (AVGS), much like a conveyor, brings work to
the operator in a driverless vehicle as shown in Figure 2.24. However, there are distinct
differences between the two systems. Load sizes are usually bigger for the AGVS, and
frequency of service is quite different. It is relatively easy to change the path of the
vehicle with the improved track devices available. A unique use of the AGVS is by
Volvo in Sweden, which uses the AGVS as a progressive assembly-line conveyor for
its automobiles, with flexibility in the line and workstation layout.
The AGVS has seen overwhelming growth in the last decade, with the increas-
ing use of computer and microprocessor technology in material handling equipment.
Earlier models, developed over 30 years ago, had limited applications as to load
capacity and the length of the path traveled. With the development of the AS/RS, the
carousel, and the emphasis on flexible manufacturing systems, the role of the AGVS
has been crucial in bridging so-called islands of automation. An AVGS offers many
benefits not limited to just the factory floor. Other uses in the factory include such
mundane tasks as delivering the mail.
One system consists of a fixed path laid out on the plant floor by cutting a groove
in the concrete with a diamond saw one-eighth-inch wide by one-half-inch deep
along the center of the entire path. A continuous insulated metal conductor is laid
into the groove and covered with epoxy. Both ends of the conductor are connected
to an electrical device to generate low-level electrical signals. The AGV is equipped
with sensors that pick up the electrical signals to keep it on the path within a range
of one-half inch. The vehicle is usually propelled by a rechargeable battery. Vehicles
are also equipped with computer and microprocessor controls to identify the loca-
tion and duration of their stops. Automatic load and unload robots at each station
may also be installed. Other systems track a strip of metallized tape or a metallic
paint stripe on the floor.
Storage AGVSs, unlike stackers, are capable of stacking parts at different heights
as well as transporting materials to various locations in the plant. The very-narrow-aisle
version (5 ft) is gaining popularity in industrial warehouse operation because of its
excellent load-handling capability of medium-sized loads, making it possible to raise
the storage height of the warehouse bins.

Power and Free Conveyors


One version of an industrial plant conveyor is a dual-track conveyor system referred
to as a power and free conveyor and shown in Figure 2.25. The system consists of two
parallel tracks, one housing the power chain and the other containing free carriers that
support the load. Latch mechanisms are attached to the power chain at a predetermined

DK565X_002_r04.indd 69 11/16/2005 1:30:41 PM


70 Assembly Process: Finishing, Packaging, and Automation

FIGURE 2.24 Automated guided vehicle system. (Courtesy of Automated Systems


Division, and Howell.)

spacing that engage the carriers to propel them through the free track. A carrier can be
stopped by a stationary cam stop mechanism that disengages the latching mechanism.
Power and free conveyors are available in overhead and inverted styles. In gen-
eral, with power and free conveyors, the product is supported and transported in three
different ways: on the floor, flush with the floor, or by the overhead system. For an
overhead system, the load is suspended from the overhead conveyor with hangers.

DK565X_002_r04.indd 70 11/16/2005 1:30:41 PM


Assembly Automation 71

FIGURE 2.25 Power and free conveyor showing two-track system. One track supports a
power chain with pushers and another supports trolley for product movement. (Courtesy of
Richards-Wilcox Manufacturing Company.)

This allows freedom of operation around the product from the top down. The con-
veyor can also be elevated so the floor area under the conveyor can be utilized for
other productive work. This type is often required in the case of equipment assembly
lines using manual labor or equipment and robots.
Figure 2.26 shows a typical layout of a power and free conveyor system. The man-
ufacturing engineer should be able to lay out the system within specific requirements
so the conveyor manufacturer can design a system to suit the customer’s needs.
Benefits of power and free conveyors include the following:

They provide on-line accumulation at a desired location without disrupting


the flow of other carriers in the system.
Unlike the conventional conveyor, the product on a power and free system
rides on a free track. This provides an extra opportunity to enhance the per-
formance by adding switches, line-full sensing devices, lifting and dipping
possibilities, and automatic inventory counts.
Unlike the conventional conveyor, where the speed has to be synchronized
with operations of the same time duration for the progressive assembly
line, the power and free system provides flexibility to provide accumulation
sections, bypass and diversion, switches, empty carrier storage, and varying
throughput requirements.
Support functions such as testing, inspection, or rework can be tied into the
same conveyor with free tracks.
Future expansion is possible by add-on sections and carriers rather than
installing a totally independent system.

DK565X_002_r04.indd 71 11/16/2005 1:30:41 PM


72 Assembly Process: Finishing, Packaging, and Automation

Receiving
Parts and products are loaded onto the
Pre-Assembly Storage conveyor, with each carrier’s destination properly
Parts and products not needed immediately are encoded and fed into the central computer before
sent to overhead storage spurs, where they are it is sent into the main loop.
indexed by product type and called for by the
central computer when needed in production.

Shipping
Products are unloaded from the
conveyor and sent to shipping
docks. The destination codes are
cleared from the computer and
the carriers are sent to recelving
to begin the proccess again, with
new loads and new instructions.
Production & Assembly
Here the conveyor supplies workers located
along multiple spurs with the necessary parts and
products. Carriers are designed for easy, quick Final Assembly
access, and variable speed drive units allow the Products are readied for
production flow to be adjusted to optimum levels shipping, with those requ
of efficiency and productivity. added drying time sent
overhead spurs for short-
storage.

Inspection
Products are tested for flaws and other quality
control considerations. Manual and automatic
switches enable the workers to send products
onto finishing and final assembly or route them
back to the production area for repair.

Finishing
Those products that require plating
or painting are switched into the
finishing area, where the conveyor
moves the loads through various ovens
and dipping stations at precise
Warehousing speeds.
Products not scheduled for immediate finishing
and shipping are stored overhead in parallel
spurs, which allow random carriers to stack
large produces in a minimum amount of space.

FIGURE 2.26 Layout of power and free conveyor system for a product going through
various activites in a plant with varying output. (Courtesy of Richards-Wilcox Manufacturing
Company.)

REFERENCES
Boothroyd, G., Assembly Automation and Product Design, Marcel Dekker, New York, 1992.
Sheth, V., Facilities Planning and Material Handling, Marcel Dekker, New York, 1995.
Tanner, J. P., Manufacturing Engineering, Marcel Dekker, New York, 1991.

DK565X_002_r04.indd 72 11/16/2005 1:30:45 PM


3 Electronics Assembly
Michael Pecht,

Denise Burkus Harris, and

Shrikar Bhagath

3.1 INTRODUCTION TO ELECTRONICS ASSEMBLY


Electronic assemblies consist of any number of electronic and nonelectronic devices
(screws, heat sinks, carrier plates, etc.) grouped together to perform a function. Each
assembly is a self-contained production item, built within a particular facility in
which it follows its own process flow. For example, an insertion-mounted printed
wiring board (PWB) can be built in the same facility as a surface-mounted PWB;
however, different equipment and processes are used. Both boards can be populated
by the same pick-and-place machine, but a wave solder system is commonly used for
insertion boards, while surface-mount boards use solder paste or thick solder plate
reflowed in a vapor-phase or infrared (IR) furnace.
This chapter discusses the basic subassemblies commonly used within deliver-
able electronic units and the assembly processing associated with each. It describes
how these subassemblies are grouped together to form new assemblies, which are in
turn the building blocks for larger assemblies, and the integration of these assemblies
to form the final system.

3.1.1 System Breakdown


A given system may have many levels of assembly. The final system, or top assembly,
can contain many subassemblies; each of these can contain its own subassemblies, and
so on. Most subassemblies are manufactured as deliverable, self-contained units, built
at various facilities. The system house obtains all of these subassemblies and integrates
them into the final system. A crucial first step is creating what is variously called a
system breakdown, system parts list, or ultimately an indented bill of material.
The list is started with the final system at the top and the major subassemblies
as branches below. These subassemblies are branched out further with their own

73

DK565X_003_r04.indd 73 11/11/2005 7:40:06 PM


74 Assembly Process: Finishing, Packaging, and Automation

subassemblies and components. A subassembly is denoted by indenting the list at


each new level.
In the example in Figure 3.1, each indentation indicates a lower level of assem-
bly, or a subassembly within the given heading.1 The quantities given are for one
assembly. For example, there are 32 transmitter modules in the radar. Each transmit-
ter module has one module housing, one cover, one PWB assembly, and so on. Thus,
you would order (or build) 32 module housings, 32 covers, 32 PWB assemblies, and
so on for each complete assembly.

3.2 TYPICAL PACKAGE ARCHITECTURE


Single-chip packages can be broadly classified into plastic packages, ceramic pack-
ages, and metal packages, depending on the kind of casing used. Figure 3.2 shows one
of the simplest component packages, encapsulated in plastic. Figure 3.3a–c shows
schematics of additional packages. A ceramic package typically has a substrate, to
which the die is fixed using a die-attach. The die has conductor pads on its periphery
to which wirebonds (or other interconnections) are attached. The other end of the
wirebond is attached to the package leads, which form the basic input/output (I/O)
of the package. Lead seals are used to fill the gap between the leads and the package
case. The lid of the package is affixed to the case with a lid seal. Table 3.1 gives some
of the commonly used materials for the package elements mentioned above.

3.3 ELEMENTARY SUBASSEMBLIES


3.3.1 Die Assemblies
Die assemblies are the simplest of subassemblies and consist of a bare die, such as
a transistor or diode, mounted on a tab with a conductive surface. Figure 3.3 shows
examples of typical die assemblies. Die assemblies are mounted on a substrate that
is protectively overcoated (in commercial applications) or encased in a hermetically
sealed package or hybrid (in military applications). Building your own assemblies
has several advantages:

Die assemblies need be only slightly bigger than the dies themselves, making
them much smaller than a discretely packaged or canned diode or transistor.
Although direct mounting on the higher assembly’s substrate (a common
practice with hybrid assemblies) would further decrease the size needed,
mounting the dies first on tabs allows functional testing before committing
them to the more expensive higher assembly.
Pretested die assemblies can be used for low-yield dies, such as application-
specific integrated circuits (ASICs); high-power dies, such as field-effect

1. This figure is typical of a breakdown used early during the design phase of a product so that the
designer can keep track of the materials first selected for the design. It would have part numbers and other
information assigned prior to appearing on a finished drawing or CAD file.

DK565X_003_r04.indd 74 11/11/2005 7:40:08 PM


Electronics Assembly 75

Final System: Transmitter for a Radar


20 #6 Screws
1 Rack
1 Cover
1 Power supply module
REF 234553 Test Spec
1 Module housing
1 Cover
1 Power supply PWB assembly
1 PWB
25 3µ capacitors
10 0.1µ capacitors
1 Transformer
1 Power hybrid
1 Case
1 Cover
1 Solder preform
2 FET assemblies
1 FET
1 BeO tab, thin-film metallized
AR 0.010-in.-diameter Al wire
1 Logic substrate assembly
1 Substrate
4 2N2222 transistors
2 Rectifier chips (ICs)
3 0.01µ capacitors
AR conductive epoxy
AR 0.001-in.-diameter Au wire
6 LM139 DIPs
8 Com-04 cer paks
1 Matrix plate
32 8-pin connectors
1 10-pin connector
15 ft 1/4-in. tubing
33 Cables
32 Transmitter modules
1 Module housing
1 Cover
1 T/R PWB assembly
1 PWB
15 2µ capacitors
1 Gate array
1 Gate array die
1 84-pin leaded chip carrier
1 Stripline
1 Combiner
4 Regulator hybrids
1 Case
1 Cover
1 Epoxy preform
1 Substrate assembly
1 Substrate
2 10-kΩ resistor
3 IN3600 diodes
3 0.033µ capacitors
AR conductive epoxy
AR 0.001-in.-diameter Au wire
1 Epoxy preform
1 Cold plate
64 Cables

FIGURE 3.1 Example of systems parts list or Christmas tree. AR, as required (a term used
for materials purchased in bulk quantities); REF, reference.

DK565X_003_r04.indd 75 11/11/2005 7:40:08 PM


76 Assembly Process: Finishing, Packaging, and Automation

Plastic case Die

Wire Die attach

Wirebond Leadframe

FIGURE 3.2 Cross section of typical plastic-encapsulated single-chip


package.

transistors (FETs); or when it is essential to match pairs of dies for func-


tional reasons.
The tab allows easier rework, because die assemblies supported by tabs lift
up in one piece. To replace the die, the substrate assembly must be heated
to reflow the solder or soften the epoxy with which the die is mounted. The
die is then extracted from the substrate. Because most dies are only 0.01-
inch-thick silicon, they tend to come off in pieces. These pieces must be
carefully pried off one at a time. Extreme care must be taken not to damage
the neighboring components.

Die Assembly Materials


The tabs used for die assemblies are metallized ceramic or metal plated to allow for
proper mounting and wirebonding. Metal tabs give the backplane of the die assembly
the same electrical function as the backplane of the die. Thus, for example, the tab on
which a transistor is mounted becomes the collector, allowing the back-plane I/O to
be accessed by connection to the backplane of the die assembly or the top of the tab.
The most common metal tabs, Kovar and molybdenum, have high thermal conductivi-
ties and coefficients of thermal expansion (CTEs) between those of the silicon dies
and the ceramic substrates on which they are mounted. The metal tabs are plated, first
with 0.0001 to 0.0003 in. of nickel to provide a corrosion barrier metal, and then with
0.00005 to 0.00015 in. of gold to allow for wirebonding and solderability. If aluminum
wirebonding is used, the minimum recommended thickness of gold is 0.0001 in.
Ceramic tabs are employed when the backplane of the die assembly should be
inactive. The surface metallization of the tab becomes the collector when a transistor is

DK565X_003_r04.indd 76 11/11/2005 7:40:09 PM


Electronics Assembly 77

ANODE CATHODE ANODE CATHODE

ANODE CATHODE ANODE CATHODE

(a) (b)

BASE

EMITTER
COLLECTOR BASE EMITTER
COLLECTOR

(c) (d)

EMITTER BASE COLLECTOR

(e) (f)
METAL/METALLIZATION

FIGURE 3.3 Die assembly examples. (a) Diode on metal tab; (b) diode on ceramic tab;
(c) transistor on metal tab; (d) transistor on ceramic tab; (e) transistor on ceramic tab with emitter
bonded; (f) transistor and diode on ceramic tab with base connected to anode and collector
connected to cathode.

DK565X_003_r04.indd 77 11/11/2005 7:40:09 PM


78 Assembly Process: Finishing, Packaging, and Automation

TABLE 3.1 Commonly Used Materials for Various Package Elements


Package Element Materials
Chip carriers and substrates Ceramic: alumina, beryllia, silicon carbide, aluminum nitride,
silicon nitride, forsterite
Metal: Molybdenum, copper-tungsten, tungsten, Kovar, stainless
steel
Semiconductor die Silicon, germanium, silicon carbide, gallium arsenide, cadmium
sulfide, indium antimonide, lead sulfide, II–V and II–VI
compounds
Die attach and adhesive Silicone, polyurethane, acrylic, epoxy Novolac, epoxy phenolic,
epoxy polyamide, solder
Wires for bonding Gold and aluminum
Leads Kovar
Lead seals Glass
Lids Same as case/chip carrier materials
Lid seal Welding, brazing, and solder materials
Encapsulants Silicones, epoxies, polyimides, polyamides, polyurethanes,
fluorocarbons, acrylics, diallyl phthlate, polyvinyl chloride,
parylene

mounted on a ceramic tab. Ceramic tabs are metallized with thick-film gold alloys or
thin-film metallization. If 96% alumina is used, the ceramic substrate is laser drilled to
create a perforated array of the desired tab dimensions, called a snapstrate. The thick
film is printed and fired onto the snapstrate, creating an array of bonding pads within
the perforations. When snapstrate is snapped apart, individual tabs result. The back
side of the snapstrate may be left as bare ceramic if the die assembly is to be dielec-
tric epoxy mounted, or it may be metallized to allow for solder mounting. If thin-film
metallization is desired, a 99% alumina substrate is used, with a base metal, such as
chromium, titanium, tungsten, or copper, deposited on it. The substrate is then nickel-
gold plated like metal tabs. After plating, the substrates are diamond sawed into tabs of
the desired dimensions. If beryllium–oxygen tabs are required to provide high thermal
conductivity, they are metallized using thin-film methods, because thick film does not
provide good adhesion to the beryllium–oxygen.

Die Assembly Processing


Die Mounting The dies are mounted to the tabs using either solder or eutectic
mounting. Figure 3.4 shows a typical phase diagram for a solder. The x-axis of the
diagram gives the composition of the alloy; the y-axis is temperature. The liquidus lines
indicate the temperature at which a given composition will completely melt or be in a
liquid phase. The solidus lines indicate the temperature at which a given composition

DK565X_003_r04.indd 78 11/11/2005 7:40:11 PM


Electronics Assembly 79

LIQUID PHASE
LIQUIDUS PLASTIC
MELTING
ZONE
POINT OF
METAL B
MELTING
TEMPERATURE

POINT OF
METAL A

SOLIDUS

SOLID PHASE

0 20 40 60 80 100 %A
100 80 60 40 20 0 % B

EUTECTIC POINT
(50% A/50% B)
FIGURE 3.4 Eutectic alloy phase diagram.

will crystallize or become completely solid. The areas between the liquidus and solidus
lines represent the plastic zones of a given composition, or the zone in which both liquid
and solid phases are present. In other words, as the temperature is increased for a given
composition, that composition starts to melt when the temperature matches that of the
solidus and has completely melted when the temperature matches that of the liquidus.
The eutectic is the point at which the solidus and liquidus intersect. A eutectic alloy is a
composition with the lowest possible melting point, going from solid to liquid at one
temperature (that is, a eutectic alloy has no plastic zone). Eutectic mounting is used for
dies that use a eutectic alloy for attachment. This method is very labor intensive and has
a high potential for damaging the die during the rubbing, or scrubbing, process.
In eutectic mounting without a solder preform, a silicon die is placed on the gold-
plated tab. The tab is placed on the hot plate of a die-bonding station. Dry nitrogen
is blown over the die-mounting area (alternatively, the entire process is performed in
an environmental chamber filled with dry nitrogen). The temperatures of the tab and

DK565X_003_r04.indd 79 11/11/2005 7:40:11 PM


80 Assembly Process: Finishing, Packaging, and Automation

die are brought to approximately 400–420°C (about 20–40°C above the gold–silicon
eutectic point). The die is rubbed gently back and forth into the gold plating until
the gold and silicon melt together to form the eutectic alloy. At this point, the die is
eutectically bonded to the tab.
Eutectic mounting with a preform, the most common die-mounting method, uses
a solder preform of the desired eutectic alloy. An alloy is chosen depending on (1) the
metallization of the substrate and the die backplane; (2) the desired reflow temperature;
and (3) the formation of intermetallics, which can affect the long-term reliability of the
unit. For example, 80% gold–20% tin has a eutectic point of 280°C. This temperature is
high enough to surface mount the substrate with eutectic lead–tin (63% lead–37% tin,
which reflows at 183°C) without reflowing the die mount, but it is lower than the eutectic
point of some other commonly used gold eutectics; this allows limiting the exposure of
the die to extreme temperature. This alloy also has good wetting properties on both the
gold metallization commonly found on the substrate and the diffused gold on the die
backplane. However, gold–tin intermetallics readily formed at relatively low tempera-
tures are very brittle and can lead to mechanical fracture failures. Silver eutectics are not
brittle, but silver can migrate, leading to small changes in resistivity over time. Silver is
also notoriously susceptible to corrosion problems. Table 3.2 lists some commonly used
solder alloys.
After an alloy is selected, a preform is punched out with a cookie cutter die, or
simply cut out of the alloy foil with scissors. The preform is placed between the die and
the mounting surface, and the die is mounted using the scrubbing method discussed
earlier.
For high-volume production, furnace mounting or fluxless solder reflow is
used. The die and preform are placed on the tab or substrate and sent through
a furnace filled with forming gas (approximately 5% hydrogen and 95% nitro-
gen). Forming gas provides a reducing atmosphere, which prevents corrosion and
contamination while allowing wetting to occur. Industrial furnaces normally have
three to nine zones; each is programmed to maintain a specified gas flow and tem-
perature. The parts travel on a belt at a programmed speed through the furnace.
The combination of gas flow, belt speed, and zone temperatures yields a profile of
the temperature the part will reach. Attaching a thermocouple to a number of parts
and running them through the furnace determines the profile. A typical profile is
shown in Figure 3.5.

Interconnection
Interconnection to the die assembly is most commonly accomplished with wire-
bonds. The standard wire is 0.0007- or 0.001-inch-diameter gold wire. When better
current-carrying capacity is required, 0.005- to 0.01-inch-diameter aluminum wire
is used to avoid using costly, larger-diameter gold wires. Table 3.3 lists the current-
carrying capabilities and fusing limits of some standard-size wires.
Gold wirebonds are made using thermocompression or thermosonic bonding. In
both cases, the wire is fed through a capillary. Heat is applied to melt the wire and
form a ball of material at its end. This ball is thermocompressed into place on the

DK565X_003_r04.indd 80 11/11/2005 7:40:11 PM


Electronics Assembly 81

TABLE 3.2 Solder Alloys


Temperature (ºC) Elements (%)
Solidus Liquidus Eutectic Pb Sn Ag Ln Sb Bi Cd Au Zn
— — 93 — 42.0 — 44.0 — — 14.0 — —
— — 95 32.0 15.5 — — — 52.5 — — —
— — 117 — 48.0 — 52.0 — — — — —
118 125 — — 50.0 — 50.0 — — — — —
— — 125 43.5 — — — — 56.5 — — —
— — 125 9.6 15.0 — 70.0 — — 5.4 — —
— — 138 — — — 42.0 — 58.0 — — —
— — 139 — 43.0 — — — 57.0 — — —
— — 143 — — 3.0 97.0 — — — — —
— — 145 30.6 51.2 — — — — 18.2 — —
155 149 — 15.0 — 5.0 80.0 — — — — —
— — 146 32.0 50.0 — — — — 18.0 — —
— — 162 18.0 70.0 — 12.0 — — — — —
144 163 — 43.0 43.0 — — — 14.0 — — —
156 165 — 25.0 — — 75.0 — — — — —
160 174 — 30.0 — — 70.0 — — — — —
— — 177 — 67.8 — — — — 32.2 — —
— — 179 36.1 62.5 1.4 — — — — — —
— — 183 38.1 61.9 — — — — — — —
183 186 — 35.0 65.0 — — — — — — —
179 189 — 36.0 62.0 2.0 — — — — — —
183 190 — 40.0 60.0 — — — — — — —
183 192 — 30.0 70.0 — — — — — — —
183 195 — 25.0 75.0 — — — — — — —
183 195 — 42.0 58.0 — — — — — — —
— — 198 — 91.0 — — — — — — 9.0
183 202 — 20.0 80.0 — — — — — — —
183 203 — 45.0 55.0 — — — — — — —
180 209 — 50.0 — — 50.0 — — — — —
183 209 — 15.0 85.0 — — — — — — —
183 214 — 50.0 50.0 — — — — — — —
183 215 — 10.0 90.0 — — — — — — —
— — 215 85.0 — 15.0 — — — — — —

(Continued)

DK565X_003_r04.indd 81 11/11/2005 7:40:11 PM


82 Assembly Process: Finishing, Packaging, and Automation

TABLE 3.2 Solder Alloys (Continued)


Temperature (ºC) Elements (%)
Solidus Liquidus Eutectic Pb Sn Ag Ln Sb Bi Cd Au Zn
183 218 — 48.0 52.0 — — — — — — —
— — 221 — 96.5 3.5 — — — — — —
183 224 — 5.0 95.0 — — — — — — —
183 225 — 55.0 45.0 — — — — — — —
179 227 — 35.5 61.5 3.0 — — — — — —
221 229 — — 96.0 4.0 — — — — — —
183 238 — 60.0 40.0 — — — — — — —
233 240 — — 95.0 — — 5.0 — — — —
183 242 — 62.0 38.0 — — — — — — —
221 245 — — 95.0 5.0 — — — — — —
179 246 — 36.0 60.0 4.0 — — — — — —
— — 246 89.5 — — — 10.5 — — — —
183 247 — 65.0 35.0 — — — — — — —
— — 248 82.6 — — — — — 17.4 — —
— — 252 88.9 — — — 11.1 — — — —
183 258 — 70.0 30.0 — — — — — — —
250 264 — 75.0 — — 25.0 — — — — —
183 268 — 75.0 25.0 — — — — — — —
— — 280 — 20.0 — — — — — 80.0 —
183 280 — 80.0 20.0 — — — — — — —
225 290 — 85.0 15.0 — — — — — — —
268 299 — 88.0 10.0 2.0 — — — — — —
296 301 — 93.5 5.0 1.5 — — — — — —
268 302 — 90.0 10.0 — — — — — — —
— — 304 97.5 — 2.5 — — — — — —
— — 309 97.5 1.0 1.5 — — — — — —
301 314 — 95.0 5.0 — — — — — — —
— — 318 99.5 — — — — — — — 0.5
316 322 — 98.0 2.0 — — — — — — —
304 365 — 94.5 — 5.5 — — — — — —

Courtesy of the International Society for Hybrid Microelectronics.

DK565X_003_r04.indd 82 11/11/2005 7:40:11 PM


Electronics Assembly 83

PEAK
TEMP.
REFLOW
TEMP.
TEMPERATURE

TIME AT
PEAK TEMP. RAMP OR COOL DOWN

RAMP UP

ROOM
TEMP.

ENTER 1 2 3 4 5 6 7 8 9 EXIT
TIME
THIS EXAMPLE SHOWS A TYPICAL 9 ZONE FURNACE PROFILE FOR THICK FILM PASTE. THE PARTS ENTER AT ROOM TEMPERATURE.
THE FIRST 3 ZONES ARE PROGRAMMED SO THAT THEY WILL RAMP UP THE TEMPERATURE QUICKLY ENOUGH FOR THE PASTE TO
REACH ITS REFLOW TEMPERATURE BEFORE THE ACTIVE SOLVENTS TOTALLY VOLATILIZE. ZONE 4, SET A PEAK TEMPERATURE
(APPROXIMATELY 30–50°C) HIGHER THAN THE REFLOW TEMPERATURE) GUARANTEES THAT THE PASTE WILL REACH AND MAINTAIN
THE REFLOW TEMPERATURE FOR THE PROPER DURATION TO ACCOMPLISH TOTAL REFLOW WITHOUT OVER FIRING. THE REMAINING
ZONES ARE PROGRAMMED TO COOL DOWN THE PARTS FAST ENOUGH TO ESTABLISH PROPER GRAIN GROWTH. BUT SLOW ENOUGH
TO PREVENT HIGH THERMAL STRESSING. THE PARTS EXIT THE FURNACE AT A SLIGHTLY ELEVATED TEMPERATURE , BUT COOL
ENOUGH THAT HITTING ROOM TEMPERATURE AIR WILL NOT THERMALLY STRESS THEM.

FIGURE 3.5 Assembly methods sheet for wiring a connector.

TABLE 3.3 Wirebond Fusing Currents


Wire Diameter (in.) Nonfusing Current Fusing Current Maximum Wire Length (in.)
0.0007 (gold) 160 mA — 0.070
0.001 (gold) 250 mA 1A 0.090
0.005 (aluminum) 3A 5–7 A 0.400
0.010 (aluminum) 10 A 17 A 0.500

die, forming a thermosonic ball bond, and heat and pressure are applied while the
wire is vibrated at ultrasonic frequencies. The wire is pulled through the capillary
and looped over to the bond pad on the substrate. The capillary is used to press or
scrub the wire into the gold-metallized bonding pad on the substrate, forming the
stitch. The wire is then broken, leaving a tail. Figure 3.6a shows the steps in forming
a ball bond.
Thermal bonding methods cannot be used to form aluminum wirebonds. When
gold (substrate metallization) and aluminum (wire) are placed together and sufficient
heat is applied, gold–aluminum intermetallics can form; these brittle intermetallics
can greatly weaken the strength of the wirebond. Thus, ultrasonic bonding techniques
are used for aluminum wirebonds. Ultrasonic techniques are similar to thermosonic
methods but are performed at room temperature. Because no heat is applied, no ball
is formed; both ends of the wirebond are stitched. These stitches, also called wedges,
produce wedge bonds. Figure 3.6b depicts the formation of a wedge bond.
Military and space applications specify that bonds be tested. In pull testing, the
common military standard test, the wires are pulled until they fail. The type of fail-
ure is documented (ball lift, break in wire loop, break at stitch, etc.), and the force at
which the failure occurred is noted. For a nondestructive pull test, each wire is pulled

DK565X_003_r04.indd 83 11/11/2005 7:40:12 PM


84 Assembly Process: Finishing, Packaging, and Automation

WIREBONDER HEAD

CAPILLARY

GOLD WIRE

1. WIRE IS FED THROUGH THE 2. HEAT IS APPLIED TO THE 3. THE BALL IS FORMED.
CAPILLARY. WIRE TO MELT IT AND
FORM A BALL.

4. THE BALL IS PULLED BACK 5. THE CAPILLARY IS LOWERED AND 6. THE CAPILLARY IS RAISED WHILE WIRE
TO PLACE IT AT THE END THE BALL IS PRESSED OR SCRUBBED IS FED THROUGH IT.
OF THE CAPILLARY. ONTO THE DIE’S BOND PAD.

7. THE CAPILLARY IS MOVED OVER AND LOWERED TO 8. THE CAPILLARY RAISES BREAKING THE WIRE OFF
THE SUBSTRATE BOND PAD, FORMING A LOOP. THE AT THE STITCH. THE WIREBOND IS COMPLETED.
CAPILLARY THEN PRESSES OR SCRUBS THE WIRE ON
THE BOND PAD FORMING THE STITCH.

(a)

FIGURE 3.6(a)

to a specified force. If the wirebond does not break or fail, the unit passes the test.
Figure 3.7 shows a wirebond being pull tested.

Applications
Military and space applications typically require maximum reductions in weight, area,
and volume, along with the highest possible reliability. They also often require the use
of state-of-the-art or leading-edge technology. For example, a radar for a fighter jet
must have the most advanced capabilities yet be small and lightweight enough to fit
into the nose of the aircraft. Use of discrete packages is very uncommon because of
the extra volume and weight they impose. Despite their higher cost, bare-die assem-
blies (i.e., hybrids) are commonly used to meet the size limitations of the system.
In addition to increased packing densities, military applications generally have
higher power densities. These often require that certain transistors be tested at high
temperatures prior to being mounted within the sealed unit, or that these transistors
be mounted on a high-spreading tab (e.g., a metallized beryllium–oxygen tab) to
prevent thermal stressing of the dies during high-temperature operations and thermal

DK565X_003_r04.indd 84 11/11/2005 7:40:12 PM


Electronics Assembly 85

WIREBONDER HEAD

CAPILLARY

GOLD WIRE

1. WIRE IS FED THROUGH THE 2. THE CAPILLARY IS LOWERED AND


CAPILLARY. THE WIRE IS ULTRASONICALLY SCRUBBED
ONTO THE DIE’S BOND PAD.

3. THE CAPILLARY IS RAISED WHILE 4. THE CAPILLARY IS MOVED OVER AND LOWERED TO
WIRE IS FED THROUGH IT. THE SUBSTRATE BOND PAD, FORMING A LOOP,. THE
CAPILLARY THEN ULTRASONICALLY SCRUBS THE WIRE ON
THE BOND PAD FORMING THE STITCH.

A A

5. THE CAPILLARY RAISES BREAKING THE WIRE OFF VIEW A–A


AT THE STITCH FORMING A TAIL. THE WIREBOND
IS COMPLETED.
(b)

FIGURE 3.6(b) Examples of the formation of (a) a ball bond and (b) a wedge bond.

cycling tests. Furthermore, the advanced functionality of the military system often
requires the use of a new die that is still immature in its design. The yields of such
units are usually low, so it is necessary to test them functionally prior to placing them
in the hybrid. Because of the die’s active backplane, this testing can be accomplished
only when the die is mounted to a conductive surface.
Die assemblies are not commonly used in commercial applications, which gen-
erally do not need the restricted sizes that merit the extra expense of additional
subassemblies. Therefore, mature, high-production-volume, discretely packaged
dies are commonly used commercially. These devices are very inexpensive (transis-
tors can be as little as $0.30 each when bought in large quantities) and have high
yield rates because of their simplicity and maturity; they have been produced for so
long in such high volumes that the processes used to produce them and the products
themselves have been fully debugged and optimized.
Die assemblies are used in commercial products with high-temperature environ-
ments; for example, car ignitions require the use of FET transistors. But discretely

DK565X_003_r04.indd 85 11/11/2005 7:40:12 PM


86 Assembly Process: Finishing, Packaging, and Automation

FIGURE 3.7 Wirebond pull test. (Courtesy of Westinghouse Electronics Systems Group.)

packaged FETs, with the extra interfaces of the discrete package, its mounting material,
and its board material, do not dissipate heat enough to maintain the proper operating
junction temperature. Therefore, a die assembly is used that consists of the necessary
FET, eutectically mounted to a highly thermally conductive tab, such as molybdenum.
This assembly is then encapsulated with a high-temperature-protective overcoat.
Another commercial application of die assemblies is the electronic sensing
devices used in oil drilling. These miniature electronic devices, located at the tip of
the drill, sense the conditions of the drill site—for example, pressure, temperature,
and viscosity—then transmit this information through the drill rod to the surface,
where it is analyzed to determine how close the drill is to the oil deposit. The envi-
ronment in this case can exceed 200°C, due to the heat of the earth and the friction
experienced by the drill.

3.3.2 Capacitor Banks


Introduction to Capacitor Banks
Capacitors are commonly used in electronic systems to regulate current and to store
energy for the electronic assembly. Capacitors can constitute the bulk of the parts list
for a power supply; power supplies, in turn, are needed to provide and regulate the

DK565X_003_r04.indd 86 11/11/2005 7:40:13 PM


Electronics Assembly 87

necessary power for virtually all systems. In high-power applications, higher farad
values are needed to ensure device functionality; in most cases, several capacitors
must be placed in parallel to obtain the necessary capacitance.

Calculating Capacitance
For capacitors in series, the total capacitance (CT) is the reciprocal of the sum of the
reciprocal capacitances of the individual capacitors:

1
CT 
1 / C1 + 1 / C2 + 1 / C3 +"1 / Cn

For capacitors in parallel, the total capacitance is the sum of the individual capacitances,

= CT (5F) = = C1(1F) = C2(1F) = C3(1F) = C4(2F)

where CT is the total capacitance; C1, C2, C3, … , Cn are the values of the individual
capacitors; F is farads, the standard measure of capacitance; and the standard symbol
for a capacitor is

Producing Capacitor Banks


Because placing multiple capacitors flat on a board to be hooked up in parallel can
take a great deal of board real estate, a capacitor (cap) bank subassembly may be the
best alternative. To produce a cap bank, the capacitors are placed on end, side by side.
The bodies of the caps are bonded together with dielectric epoxy to prevent shorting
the end terminations of any one cap. The end terminations of one side of the bank
are bonded together with conductive epoxy or a low-temperature solder (e.g., 60/40
SnPb or eutectic SnPb). Some applications require electrical isolation between end
terminations of different caps.

Example
Figure 3.8 shows an example of a cap bank. All the negative terminations are
grounded to the floor of the module housing with conductive epoxy. To achieve

DK565X_003_r04.indd 87 11/11/2005 7:40:13 PM


88 Assembly Process: Finishing, Packaging, and Automation

A A

METALLIZATION

FUSES OR WIREBONDS
(4 PLACES)

KAPTON TAPE

FUSE BOARD

KAPTON TAPE

VIEW A–A

FIGURE 3.8 Capacitor bank with interconnecting board and fuses.

electrical isolation between the positive terminations, small pieces of Kapton tape
are placed between each cap with dielectric epoxy. If the caps were ceramic with
gold-plated terminations, the connections to the positive terminations could be
accomplished with wirebonds or by attaching a gold ribbon with conductive epoxy.
If the caps were tinned or had solder-coated terminations, a jumper wire could be
hand-soldered to the appropriate termination. In Figure 3.8, a small ceramic board
with a metallized pattern is placed over the caps; tolerancing has been calculated to
allow the cap leads to fit into the drilled holes of the board. Because the metallized
pattern is only on the top surface of the board, the ceramic board material in direct
contact with the caps acts as an insulator. The leads are connected to the ring of
metal around the holes with solder or conductive epoxy. Attached to these rings are
the bonding pads needed for making the connections to the appropriate I/Os on the
associated substrate. Fuse wires can also be added to this design, as needed.

3.3.3 Microwave and Radio-Frequency Subassemblies


Microwave functionality is governed as much by mechanical design as by electrical
design. The placement, orientation, and proximity of components to one another can
greatly influence device electromagnetic fields and thus affect system functionality
at operating frequencies associated with microwave applications. At radio frequen-
cies (RFs), signal speeds are very fast. The signal or current travels along the metal
surface; if the path length is too long, it can slow down the signal. Thus, when signal

DK565X_003_r04.indd 88 11/11/2005 7:40:13 PM


Electronics Assembly 89

speed is the top priority, the signal path must be shortened by reducing the resistance.
Resistance can be reduced by using a very electrically conductive material for the
signal path, by using highly conductive die-attach materials, or by shortening the
physical length of the paths. In some extreme cases, the only configuration that will
meet the operating rise times required by the design is to stack the components on
top of each other. In other cases, electromagnetic interference (EMI) shielding is
provided by metal barriers added to the die assembly or integrated into the module
housing.
Just as component placement affects electromagnetic fields, so does the wirebond
configuration. Wirebonds are formed with a loop between the connection points that
acts like the loop in an inductor coil. This loop, with its imposed inductance and
capacitance, can change the functionality of the circuit design. Consequently, ribbon
bonds are commonly used in microwave applications. A ribbon bond is not looped
and thus does not add unwanted inductance to the connection. The potential induc-
tance can be further limited by tacking the ribbon down along the connection path,
either by welding it at multiple locations or by attaching it with conductive epoxy.
The ribbon is connected to the die and the substrate metallization with a welding tip
split in half; one tip welds at each end of the width of the ribbon. Power is applied
and the ribbon is welded across its width to the surface metallization.
Microwave subassemblies are highly customized and unique to the electrical
requirements of the device in which they will be utilized. Figure 3.9 gives some
examples of microwave subassemblies.

3.3.4 Summary
Elementary subassemblies are used when area restrictions are critical or when special
pretesting is required. Common applications are in military and space equipment,
specifically avionics or satellite applications, where weight and size are of utmost
importance. Elementary subassemblies are typically composed of simple discrete
functions; a die assembly is usually only a diode or transistor, and a cap bank is sev-
eral caps assembled together to form one subassembly that is functionally equivalent
to one large farad-valued capacitor. The need for elementary subassemblies is very
design dependent: If the design requires spreading heat under a high-power FET,
a die assembly with a beryllium–oxygen tab might solve the problem; if fast rise
times are required, stacking components may be the answer. These subassemblies
can be built or purchased beforehand as individual products, or built as an integral
part of a higher assembly, such as a hybrid.
In high-voltage applications, cross-talk and arcing potentials are critical. To prevent
such problems, adequate spacing is designed into the assembly. If the real estate for
such spacing is not available, spacing between two tracks can be increased by break-
ing up the substrate into different subassemblies. Because the current or charge travels
on the surface, the path of potential arcing must travel down the side of one substrate,
across the floor of the case, and up the side of another substrate. Thus, without an
increase in surface area, the spacing between two high-voltage potentials is increased
threefold.

DK565X_003_r04.indd 89 11/11/2005 7:40:14 PM


90 Assembly Process: Finishing, Packaging, and Automation

FIGURE 3.9 Examples of microwave subassemblies.

DK565X_003_r04.indd 90 11/11/2005 7:40:14 PM


Electronics Assembly 91

3.4 CHIP CARRIER ASSEMBLIES


Chip carriers are plastic or ceramic packages that carry a single chip or die. The
assembly processes used to make chip carriers are very similar to those used to make
dual in-line packages (DIPs) for plastic chip carriers or ceramic packages (cerpaks)
for ceramic chip carriers.

3.4.1 Plastic Chip Carrier Fabrication and Design


A plastic chip carrier package is fabricated by first punching or etching a lead-
frame pattern into sheet metal, typically Kovar plated with nickel and gold. The
die is then epoxied to the center of the leadframe and wirebonded accordingly. The
leadframe is encapsulated, with die and wirebonds, within a plastic mold (plastic
chip carriers are always leaded). The leads extending out of the molded body can
have a gull-wing or DIP configuration, or be shaped like the letter J (J-leaded chip
carriers). There are several molding techniques available for the encapsulation of
plastic packages. The most widely used are transfer molding, injection molding,
and reaction-injection molding.

Fabrication Using the Transfer Molding Process


In transfer molding, a plunger pushes the molding compound into a mold after the
compound has been preheated to a set temperature, typically 90–95°C, for between
20 and 40 sec. The compound is then inserted into a transfer pot whose cylindrical
cavity is maintained at around 170–175°C. The transfer plunger is depressed, which
pushes the softened compound into the mold. The viscous molding compound flows
over the chips, wirebonds, and leadframes and encapsulates the device.
There are basically two types of molds: the cavity chase mold (split-mold
design with ejector pins, runners, gates, and vents) and the aperture plate mold.
The latter, specifically designed for encapsulating microelectronic packages, is
assembled from a series of stacked plates. The sides of the package body are
formed with cutouts in the two aperture plates, and the leadframes are loaded
between these two plates. The runner system is in the plate above the aperture
plates. On the top surface of the molded body is the surface finish plate. In a
standard press without microprocessor control, the transfer pressure is controlled
by throttling the higher packing pressure through a speed-control valve. Once the
mold fills, the transfer pressure builds up to the packing pressure, because the flow
through the throttling valve slows to a stop (Manzione, 1990).
Immediately following the molding process, the package is postcured in large
ovens for 4–8 h at 170–180°C to enable the molding compound to undergo com-
plete conversion. The packages are then deflashed using pneumatic means or solvent
deflashers (flash is the molding compound that inadvertently flows onto the lead-
frame). The leads are trimmed and formed (bent into shape), then solder-dipped and
plated to facilitate solder attachment to the circuit board.

DK565X_003_r04.indd 91 11/11/2005 7:40:16 PM


92 Assembly Process: Finishing, Packaging, and Automation

Fabrication Using the Injection Molding Process


The injection molding process was developed for thermoplastic materials. In this
process, plastic, in the form of small pellets, falls from a feed hopper into a screw
that forces the material forward through several heating zones. As the plastic is
transported, it is heated to a highly viscous liquid state, arriving at the front zone
fully melted, de-aired, dried, and compressed up to 20,000 lb/in.2. The flights of the
screw vary in each zone. When enough melt has accumulated, the screw is forced
forward hydraulically as a piston, pushing the melt through a nozzle and down run-
ners to a mold that is cooled, usually with chilled water, to freeze the plastic as
quickly as possible. The removal of heat from the plastic is the time-limiting factor,
which determines the overall cycle time, from seconds to minutes. The thermo-
plastic molding materials are not altered substantially by this process, so sprues,
runners, and scrap parts can be reground and blended with virgin material for reuse
(Harper, 1991).
Injection-molded parts are extremely low in cost because of the automated nature
of the process and the very low cycle time. However, the molds are extremely com-
plex, must withstand very high pressures, and are very costly. Production must be
considerable to justify the mold cost and setup expense.

Reaction Injection Molding Process


Reaction injection molding is a type of injection molding used to produce thermoset
shapes, usually of large sizes. Polyurethane and polyester resins are used. Reactive
liquid components are prepared separately, then pumped into a mixing head,
where they are thoroughly combined. The mixed liquid resins are then forced into
a heated mold that may also contain fiber-mat reinforcement. The part cures like a
compression-molded part. Advantages are low cost and large parts. Disadvantages
are the limited choice of materials, the rather low-performance physical properties,
and the difficulty of controlling the reaction process (Harper, 1991).

3.4.2 Ceramic Chip Carrier Fabrication and Design


Ceramic chip carriers are made of cofired ceramic with tungsten metallization.
The ceramic body is available in two styles, slam or flatpack, with a cavity for the die.
The slam chip carrier can be single or multilayered. The die mount pad is in the center,
surrounded by the wirebond pads. The bond pads via down to the buried layer, fan out
toward the edge of the chip carrier, or both, where they either via up to the top of the
castellation (single layer) or via down to the solder-mount pad/castellation (multilayer).
On the surface, surrounding the wirebond pads, is a ring frame of metallization used to
solder-mount a dome-shaped lid (multilayer) or a domed lid glass (single layer).
Multilayered ceramic chip carriers are cofired ceramic layers with windows cut out
in the top layers. When these layers are fired together, the windows form a cavity. The
bottom layers of the chip carrier form the bottom of the package. The next layers have
small windows that form the lower cavity, where the die is mounted. The third set of
layers, with a slightly larger window, form the wirebonding ledge within the chip carrier.

DK565X_003_r04.indd 92 11/11/2005 7:40:16 PM


Electronics Assembly 93

The top layer has a metallization ring surrounding the window, which is used for solder-
sealing the lid onto the chip carrier. Figure 3.10 shows multilayered chip carriers.
Chip carriers are available in a cavity-up or cavity-down configuration. In the
former, the cavity faces up from the surface mount pads on the bottom of the package.
That is, the lid goes on the top of the package and the bottom is mounted downward
to the PWB. In a cavity-down package, the lid is on the bottom of the package and
is placed down on the PWB. A cavity-up package is used for traditional mounting
configurations in which both routing and thermal paths are directed down through
the PWB. A cavity-down package also allows the routing to be directed down to the
PWB, but the thermal path is directed toward the top of the package, away from the
PWB. This thermal design is sometimes used in convection-cooled systems.
Ceramic chip carriers are available in a variety of lead configurations. Leadless
chip carriers have castellations. When the chip carriers are fabricated, the green tape
used to make them is larger than the desired package, even considering the tape shrink-
age during firing. Holes are punched into all the tape layers around the outer perimeter
of the carrier, and the edges of these holes are metallized with tungsten. After the
layers are laminated together, the prefired stack-up is cut to size. The outer rim of
green tape is cut off through the holes, leaving evenly spaced semicircular columns
around the outer perimeter of the carrier. These metallized half-circles form the cas-
tellations. The castellations provide a column of metal that allows the solder to flow

FIGURE 3.10 Examples of multilayered chip carriers. (Courtesy of Westinghouse


Electronics Systems Group.)

DK565X_003_r04.indd 93 11/11/2005 7:40:16 PM


94 Assembly Process: Finishing, Packaging, and Automation

up, forming the solder fillet during the surface-mounting process. Figure 3.11 shows
a leadless chip carrier mounted to a board. Note the solder fillets that have formed in
the castellations. Figure 3.12 shows a leaded chip carrier, also mounted to a board, on
which the leads have been top-brazed to the package to give the standard gull-wing
leaded style. The leads have been formed to step down to the board, providing a shock
absorber. As the board temperature fluctuates, the chip carrier expands at a different
rate than the board. The bent lead will give, absorbing the thermal stress between the
carrier and the board.
Leaded chip carriers are fabricated like leadless ones, with the additional step
of brazing on leads after the exposed tungsten metallization is plated. The leads,
Kovar with nickel and gold plating, are attached using a high-temperature braze,
such as a eutectic copper–silver. The leads can be brazed to the top of the package
in a gull-wing configuration; to the bottom of the package; to the side of the pack-
age, for insertion mounting; or to the side of the package in a J-lead form for surface
mounting.

3.4.3 Pin-Grid Array Packaging


Pin-grid array (PGA) packages are high-lead-count chip carriers. To accommodate the
large number of leads without increasing the perimeter of the chip carrier, the I/Os are

FIGURE 3.11 Leadless chip carrier (LCC) mounted to a PWB. (Courtesy of


Westinghouse Electronics Systems Group.)

DK565X_003_r04.indd 94 11/11/2005 7:40:18 PM


Electronics Assembly 95

designed as an array of pads on the bottom of the package, instead of in a single row
along the package perimeter. The leads are pedestals brazed onto the pad array, yielding
a set of pins for insertion mounting on the PWB. Likewise, to accommodate the high
I/O count, dual ledges for wirebonding are typically used inside a PGA. PGAs are
available in both cavity-up and cavity-down configurations.

3.5 HYBRID MICROELECTRONICS ASSEMBLIES


3.5.1 Definition of a Hybrid
A hybrid microelectronics device (hybrid) with a customized electronic function consists
of two or more electronic components mounted and interconnected via a substrate. The
substrate is made of a dielectric material patterned with metallized signals or tracks.
Hybrids are known by many names: multicircuit packages, multichip packages
(MCPs), multicircuit hybrid packages (MHPs), and power hybrid packages (PHPs)
(for hybrids with power densities greater than 10 W/in.2). Some microwave hybrids
are also known as integrated microwave assemblies (IMAs), but IMA is the name
used for any microwave assembly; all microwave hybrids are IMAs, but not all
IMAs are hybrids. The essential characteristic of a hybrid is that it contains multiple
components.

FIGURE 3.12 Gull-wing chip carrier mounted to a PWB. (Courtesy of Westinghouse


Electronics Systems Group.)

DK565X_003_r04.indd 95 11/11/2005 7:40:19 PM


96 Assembly Process: Finishing, Packaging, and Automation

3.5.2 Hybrid Design


Unlike PWBs, in which all components are individually packaged and then mounted
to the board, hybrids use bare dies packaged together in a single hermetic case.
A given schematic, with a parts list of needed components, can be built with individ-
ually packaged and leaded components mounted on an organic PWB, which provides
the necessary routing among components. The same device can be built using bare
dies, chip components (e.g., capacitors), and thick-film resistors, all mounted to a
multilayer ceramic substrate that provides the internal routing; the substrate assembly,
mounted inside a hermetically sealed metal package, forms the hybrid.
A hybrid can be as simple as two diodes in one chip carrier or as complicated as
a multiple-channel amplifier, regulator, or analog-to-digital (A/D) converter. Hybrids
can be digital, analog, or a combination of both.
In the most common hybrid design, bare dies and chip components are mounted
on a multilayer ceramic substrate. A hybrid substrate is typically 96% alumina with
thick-film metallization; however, green tape, cofired ceramic, and thin-film mul-
tilayering using polyamide as the dielectric may also be used, depending on the
particular requirements of the design. Because bare dies are mounted, the substrate
assembly must be packaged or sealed to protect them. The substrate assembly is
either mounted inside a leaded case or encapsulated within a protective overcoat, as
shown in Figure 3.13.

FIGURE 3.13 Encapsulated hybrid. (Courtesy of Westinghouse Electronics Systems


Group.)

DK565X_003_r04.indd 96 11/11/2005 7:40:19 PM


Electronics Assembly 97

The advantages of hybrids are a vast reduction in volume, area, and weight; better
thermal management; increased functional densities; increased frequency capabilities;
and improved electrical performance. The disadvantage is increased cost.
Eliminating individual component packaging is the major factor in size reduction,
allowing bare die-mounting directly onto the substrate. Components can be placed much
closer, and interconnections between components are more direct. Instead of wirebond-
ing the bare die to the package’s internal leads and channeling the signal through the
feedthrough to the external leads (which then have to be soldered to the PWB, which in
turn routes the signal to the solder joint of another packaged die, etc.), the die on a hybrid
can be wirebonded to the substrate, which connects directly to the wirebond pad of the
other die. Thus, electrical performance and frequencies are improved.
Thermal interfaces are also eliminated. On an organic PWB, the heat must travel
from the die through the attaching media, the die package, the package leads, the
solder, and finally through the PWB to a heat sink. In a hybrid, the heat travels from
the die through the attaching material to the substrate and through the hybrid case to
the heat sink. Not only are there fewer thermal interfaces, the thermal conductivities
of the hybrid materials are higher and can dissipate more heat. Thus, thermal manage-
ment is improved.
The disadvantage of using a hybrid is the increase in cost and turnaround time
from electrical design to finished product. Organic PWBs use polyimide with copper
metallization. Hybrids use ceramics with gold, silver, or gold alloy metallization, all
of which are higher in cost.
Dies, whether discrete ones such as diodes and transistors, hor integrated circuits
(ICs), can be purchased in two forms: prepackaged or bare. Packaged dies come in
many varieties: leadless chip carriers (LCCs) for surface mounting, DIPs for insertion
mounting, plastic J-leaded or gull-winged chip carriers for surface mounting, or canned
dies such as the three-leaded metal can packages commonly used for transistors. All
packaged dies are designed for mounting on PWBs. Suppliers package the dies and
then test and sort them. For example, in the case of a diode, all the bare dies are probed
while still in wafer form; this go–no go test indicates only that the dies work like diodes
or not at all. After being packaged, the diodes that passed the test are tested under full-
power operating conditions. They are then sorted into bins—for example, there might
be three different bins for a particular diode: one for diodes that operate only under 3
V, one for 3–7 V, and one for 7–12 V. These units can also be fully screened accord-
ing to military standards, including burn-in to weed out any infant failures. Because
packaged dies are in greater demand than bare dies and are manufactured in very large
quantities using automated techniques, their cost is only pennies more than that of
bare dies. In some cases, bare dies cost even more than their packaged counterparts
because of increased handling difficulties, electrostatic discharge (ESD) damage, and
decreased yields (because bare dies cannot be completely tested and screened prior to
packaging). For some ICs, especially those used in microwave applications, packag-
ing increases yield, due to the EMI and ESD protection it offers. Bare dies are more
susceptible to damage than their packaged counterparts; their surface metallization can
be scratched or contaminated, damaged by ESD, or chipped by tweezers, or they can
simply be lost because of their tiny size (a transistor can be as small as 0.011  0.007 in.).

DK565X_003_r04.indd 97 11/11/2005 7:40:20 PM


98 Assembly Process: Finishing, Packaging, and Automation

Packaged dies have high postmounted yields, typically over 98%, while bare dies have
lower yields because of their vulnerability to handling or ESD during mounting. Bare
die yields can be as low as 60%, depending on the complexity, sensitivity, and process
control used in their mounting, but, under proper conditions, a bare die yield of 90–97%
can be expected.
The yield of a PWB assembly is normally quite high because all elements and
components used in the assembly are fully tested prior to commitment to the PWB.
Hybrids, however, have lower yields. Because they use several bare dies, the overall
yield problem of the end product is compounded.
Hybrids can also be expensive in terms of design time. As with PWBs, layout
and routing are done by computer. After being checked, the information from the
design station is downloaded into a laser artwork generator (LAG), which draws out
the layers on a film such as Mylar. This artwork is used to pattern the metal on the
PWB. For hybrids, this artwork must first be reduced and then converted into screens
and stencils to pattern the metallization onto the substrate. After board fabrication,
the PWB can be assembled by mounting the components and reflowing the solder.
For hybrids, components must be wirebonded before the substrate assembly can be
packaged. Thus, hybrids are more labor intensive than PWBs.
The lead time needed to order and receive components is also typically longer for
hybrids. Because hybrids are used only when size limitations are extreme, they are
typically not a high-volume production item. Bare dies and hybrid chip components
are therefore not in as great demand as surface-mount components, and costs and
delivery times are correspondingly greater.
Although extra manufacturing steps, component availability and lead times, mate-
rials and component costs, turnaround times, and yield factors can all lead to higher
costs, for hybrids these cost impacts can be reduced through several procedures:

Die placement and interconnection can be automated (this requires some capital
expenditure).
Full functional testing of dies can be conducted prior to mounting (requires
sophisticated universal die-testing equipment that can cost approximately
$2 million).
Proper precautions can be employed to reduce possible ESD damage.
Process control and operator training can be used to reduce handling damages.
Rework procedures can be used to increase final yields.
Design guidelines and checklists can be developed to reduce design cycle time.
Design engineers can be educated about the processing capabilities of the
facility.
Manufacturers can be kept aware of designers’ needs.
Design and manufacturing engineers can be updated on system trends and
design demands.
The overall tasks of design and fabrication can be carefully planned. (Defining
piece parts early and using expeditious purchasing can eliminate bottle-
necks. Good planning can also keep the electrical design from going to
layout prior to simulation and debugging.)

DK565X_003_r04.indd 98 11/11/2005 7:40:20 PM


Electronics Assembly 99

Concurrent engineering, involving all engineering fields from design through


production can be utilized. (System, electrical, and mechanical designers must
work with manufacturing, reliability, and production engineers to design a
producible product; engineers must work with purchasing and management
to coordinate efficient design and building efforts.)

3.5.3 Hybrid Processing


Component Mounting
Introduction to Component Mounting The most common method of mounting
components on the substrate is to attach them with epoxy. In the 1960s and 1970s,
dielectric (i.e., nonconductive) epoxies were used. However, with advances in polymer
engineering, epoxy adhesion was greatly improved, as were techniques for filling
epoxies with metals to make them electrically conductive. To obtain a low enough
resistance to allow for proper conduction, the epoxies must be 70–80% gold or silver.
Because of cost, silver is the more commonly used material. Other metals are either
not conductive enough, too expensive, too corrosive, or not conducive to filling.
In recent years, the use of silver-filled epoxies in military applications has been
questioned. The silver in epoxies can oxidize and reduce conductivity over time;
silver can also migrate over lengthy periods, collecting in one area of the epoxy joint,
leaving behind an insulating layer. Data showing gradual resistance changes in the
epoxy support these two concerns. However, recent studies have shown that if the
epoxy is properly processed and is contained in a hermetic environment away from
potential contaminants and oxidizing agents, such as air and moisture, resistance
changes are insignificant. Silver migration still occurs, but its effect is negligible.
Proper Storage The organics within the epoxies can spoil if not stored properly.
Premixed epoxies are kept frozen or refrigerated. The shelf life of frozen epoxy is
one year; of refrigerated epoxy, six months; and of epoxy stored at room temperature,
one to three months. When the epoxy is to be used, it is taken out of the freezer and
allowed to thaw before the container is opened; if the container is opened before the
material reaches room temperature, humidity in the air can condense in the material
and contaminate it. The material has to be protected by using only clean stainless
steel utensils to mix or apply it; wood tongue depressors can leave fibers that can
react with the epoxy organics. Old material left on the screen after printing should
never be mixed back in with the fresh epoxy; this material should be gathered in its
own container and recertified prior to use on deliverable products, or it should be set
aside for engineering or prototyping use only.
Conductivity If the device is mounted with conductive epoxy, the epoxy acts as
the electrical connection; it should only be used under end terminations to avoid
shorting. In high-current or microwave applications, in which the resistivity of the
conductive epoxy or the minute changes of resistance over time do not meet the
electrical performance criteria of the design, the capacitors should be mounted
with dielectric epoxy under the entire length of the capacitor body, and the end

DK565X_003_r04.indd 99 11/11/2005 7:40:20 PM


100 Assembly Process: Finishing, Packaging, and Automation

terminations should be wirebonded to provide the electrical connection. Though


the same component mounting techniques can be used for both dielectric and
conductive epoxies, using dielectric epoxy dictates the need for additional
interconnections, usually in the form of wirebonds.

Applying Adhesives Epoxy can be applied as either a film or a paste. Some epoxy
films use a fiberglass mesh support in the prepreg or B stage to laminate the layers
of an organic PWB. Epoxy films can also be formed by printing a layer of epoxy on
a nonstick, inert surface (e.g., Mylar tape) and partially curing it, leaving it flexible
but no longer fluid. Film adhesives can be purchased in approximately 12 in.  12
in. sheets or can be preformed by cutting or punching. Film adhesives are from
0.003 to 0.012 in. thick. Most film adhesives require the application of pressure
during cure.
Film adhesives are used for mounting large-area devices (that is, mounting LCCs
approximately 0.3 in.2 to larger die assemblies or chip carriers with higher I/O counts),
for mounting substrates to the case floor, for mounting the case to the PWB, or for
attaching the cold plate to the module housing. Preforms are not practical for mounting
small dies or chip components—pastes are used for this purpose. Epoxy pastes should
be hand-applied only when building prototypes and then using a stainless steel spatula
or dental picks. If the facility uses automatic dispensing, the epoxy paste is placed
in the dispenser gun and pressure is used to push it out. With a manual dispenser,
the pressure is controlled through a trigger or foot pedal. More advanced equipment
applies a set pressure pulse automatically, giving uniform amounts of epoxy. Some
automatic dispensers can be programmed to deposit epoxy in predetermined loca-
tions, much as pick-and-place machines are programmed to place components.
The most common method of epoxy application is printing the epoxy, just as thick-
film paste is screen-printed, using a larger mesh (typically 80 count) and a thicker
emulsion (0.002–0.01 in.). The larger mesh accommodates the epoxy’s larger particle
size and higher viscosity. The thicker emulsion is used to regulate the thickness of the
epoxy deposit. Stencil printing can also be used, as in solder paste printing.

Soldering Other component mounting methods use solders. The die can be scrubbed
down using eutectic alloy formation between the die and the substrate metallization or
eutectic solder alloy preforms. Furnace mounting, using eutectic alloy preforms, is
another process used with hybrids.
Chip components can also be mounted using solder paste or solder wire. The
solders used for these applications are normally lead–tin alloys rather than the gold
alloys used for die attachment. They reflow at lower temperatures and are not neces-
sarily eutectic, though they can be. To get good uniform adhesion, or wetting, of the
solder to the surface metal, oxides must be removed. This can be accomplished in
two ways. In the first method, the solder is reflowed in a reducing atmosphere in a
furnace filled with forming gas (a mixture of inert nitrogen and approximately 5%
hydrogen). The nitrogen prevents the introduction of any new oxides, and the hydro-
gen, at the reflow temperature, reduces the oxides that have formed on the solder.
This method is used for high-temperature or eutectic solders.

DK565X_003_r04.indd 100 11/11/2005 7:40:20 PM


Electronics Assembly 101

The other method utilizes flux to increase wetting. The flux chemically reduces
any oxides on the metal surface or in the solder and can be applied to the surface or
component prior to mounting. For example, flux can be printed on a pretinned or thick-
solder plated board, providing a tacky surface that temporarily holds components in
place prior to solder reflow. Pretinned components can also be dipped into flux before
placing them on the board surface, but this labor-intensive method is not frequently
used. Solder pastes and most solder wires have flux in them. Solder pastes suspend par-
ticles of the solder alloy in a solution containing flux, emulsifiers, and volatile solvents,
which burn off during reflow. Solder wires have a rosin core that contains the flux.
Solder pastes or wires are reflowed by applying heat locally with a soldering
wand (soldering gun) or hot-air gun. Solder paste is applied to the substrate by hand
or by screening; the component is placed in the solder and the solder is reflowed.
The solder paste can be reflowed locally by submitting the entire substrate to the
necessary heating cycle. The substrate is placed on top of a hot plate or on a reflow
system that employs a belt that travels the length of the reflow equipment through
zones heated conductively (by coils under the belt), convectively (by hot air or nitro-
gen blown down onto the belt), or both.
Furnace reflow can involve fluxless solder reflow, as in eutectic die attach, or
fluxed solder. In the case of fluxed solder, the reflow is performed in a nitrogen atmo-
sphere; a reducing atmosphere is not needed because the flux provides the reducing
or wetting agents. Another way to provide the energy needed to reflow the solder
exposes the substrate in the furnace to infrared radiation (IR furnace).
Finally, in the vapor-phase method, the substrate is lowered into a chamber filled
with Freon vapor; the boiling point of the Freon determines the temperature of the
vapor. As the part is lowered into the vapor, the solder reflows as the vapor condenses
on the surface of the substrate and transfers its latent condensation energy to the solder.
Figure 3.14 shows the conceptual construction of typical vapor-phase systems.

FIGURE 3.14 Conceptual outline of vapor-phase reflow. (Courtesy of Westinghouse


Electronics Systems Group.)

DK565X_003_r04.indd 101 11/11/2005 7:40:20 PM


102 Assembly Process: Finishing, Packaging, and Automation

When using these methods, the substrates must be defluxed or cleaned prior to
mounting any bare dies. It is best to deflux the unit before it has a chance to cool
down fully, or the flux may bake onto the board and penetrate the surface. The board
must be cleaned while it is still warm and the flux is still tacky to minimize flux residue,
which can outgas later and damage the bare die. Flux damages bare dies, so they
cannot be mounted using flux methods.
Dies are passivated (given a protective overcoat of a glass or an oxide) only at
the wafer level. When the wafer is sawn into individual dies, the passivation is also
cut. Thus, only the die surface, not its sides, is protected. Flux and flux fumes can
deposit on the die, migrate under the passivation at the sides, then attack the oxides
and metallization, causing loss of functionality over time.

Interconnection
Wirebonding and Ribbonbonding The most common interconnections in a hybrid
are wirebonds. In high-power or microwave applications, a single wirebond connection
does not always meet thermal or electrical requirements. In high-power FETs, the
current, and thus the heat, is dissipated through the gate connection; a single wirebond
used for this connection, even if it has a large diameter (0.005 in.), might not be able
to handle the thermal transfer. When the heat travels from the bond pad surface of the
gate to the much smaller surface area of the wirebond, the power is bottlenecked and
cannot transfer fast enough, resulting in a burned-out gate (this power loss should not
be confused with the power loss in the silicon due to FET switching). Consequently,
multiple wirebonds are often used, although the wire itself may be rated for the carried
current. Sometimes it is better to use three or four smaller-diameter gold wirebonds
instead of a single larger-diameter aluminum wirebond; the increased surface area of
multiple bonds, coupled with the higher thermal conductivity of the gold wire, can
greatly decrease the potential for a thermal failure under high-power conditions.
There are several different kinds of wirebonds, including the wedge bond, the ball
bond, and the stitch bond (Harper, 1970). Wedge bonds are made with a wedge or
chisel-shaped tool that applies pressure to the lead wire on a preheated bonding pad.
Difficulties with wedge bonding include imprecise temperature control, poor wire
quality, inadequately mounted silicon chips, and a poorly finished bonding tool.
Ball bonding is a process in which a small ball is formed on the end of the wire
by severing the wire with a flame; the ball is then deformed under pressure against
the pad area on the silicon chip. The number of steps in this bonding operation is
small, and the strength of the bond obtained is strong. Aluminum wire cannot be
used because of its inability to form a ball when melted with a flame. However, gold
wire, which is an excellent electric conductor and more ductile than aluminum, can
be used. A disadvantage of ball bonding is that a relatively large bond pad is required.
Figure 3.6a shows the steps in forming a ball bond.
Stitch bonding combines some of the advantages of both wedge and ball bonding.
The wire is fed through the bonding capillary, but the bonding area is smaller than
for ball bonds, and no hydrogen flame is required. Both gold and aluminum wires
can be bonded at a high rate. Figure 3.6b depicts the formation of a wedge bond.

DK565X_003_r04.indd 102 11/11/2005 7:40:20 PM


Electronics Assembly 103

Gold wire is typically bonded using thermocompression; aluminum wire is typi-


cally bonded using an ultrasonic process. Thermocompression wirebonding depends
on heat and pressure. In general, the bonding equipment includes a microscope,
a heated stage, and a heated wedge or capillary that applies pressure to the wire at
the interface with the bonding surface. In addition, a wire-feed mechanism is required,
as is some mechanism for manipulation and control. Three primary parameters affect-
ing thermocompression bonding are force, temperature, and time. These parameters
are interdependent and are affected by other conditions and factors. Minor changes in
these variables can cause significant differences in bond characteristics. Low bonding
temperature is desirable to avoid degradation of wire bonds due to gold–aluminum
interactions. Low pressures avoid fracturing or otherwise damaging the silicon beneath
the bond. The bonding tool used in the process may be of tungsten carbide, titanium
carbide, sapphire, or ceramic (Schafft, 1972).
In a ball bond the weakest link occurs in the annealed wire leading to the bond. In
a stitch or wedge bond, it occurs in the region of the wire in which the cross section
has been reduced by the bonding tool.
Ultrasonic wirebonding also involves heat and pressure, but heat is supplied
by ultrasonic energy rather than by a heated stage or capillaries. Pressure is also
used but is incidental to the effect of the ultrasonic energy. Three primary factors
affecting ultrasonic wirebonds are force, time, and ultrasonic power. The ultra-
sonic power available for making the bond is dependent on the power setting of
the oscillator power supply and the frequency adjustment of the tool. The force
used is generally of the order of tens of grams and is large enough to hold the
wire in place without slipping and to channel the ultrasonic energy into the bond-
ing site without causing deformation of the wire. High power and short bonding
time are usually preferred to avoid metal fatigue and to prevent the initiation of
internal cracks. Lower power nevertheless gives a good surface finish and a large
pull strength.
The third bonding method is a combination of ultrasonic and thermocompres-
sion wirebonding. In ultrasonic ball bonders, the ultrasonic heat is identical to that in
ordinary ultrasonic bonders, but a straight-wire capillary is used to feed the wire, as
on the thermocompression bonder. Also included is the flame-off device necessary to
form the ball on the gold wire (Pecht, 1991).
In ribbon bonding, a gold ribbon is split-tip welded to the die and substrate.
Ribbons range from 0.005 to 0.05 in. wide and from 0.001 to 0.005 in. thick. The
cross-sectional area of a ribbon can offer more current-carrying capacity than a
wirebond, and the larger surface area allows power transfer from the die to the
ribbon. An example comparison of cross-sectional and bonding surface area is
given below.

Cross-sectional area Bond surface


(105 in.2) (105 in.2)
0.005-inch-diameter wire 1.9 2.0
0.01 in.  0.002 in. ribbon 2.0 5.0

DK565X_003_r04.indd 103 11/11/2005 7:40:20 PM


104 Assembly Process: Finishing, Packaging, and Automation

As can be seen in this tabulation, a ribbon gives 2.5 times the heat-transferring area
for approximately the same cross-sectional area. This factor can be further improved by
tacking down the ribbon at multiple places on the pad site.
Tape Automated Bonding Another interconnection method used in hybrids is tape
automated bonding (TAB). With TAB, copper metallization is built up on top of a
sheet or tape of Kapton. The copper is patterned using photolithography or etching.
If the design requires multilayering, another layer of Kapton is placed on top of the
patterned copper and the etching is repeated for the second layer. The copper between
the layers of Kapton is sealed as protection against corrosion; the exposed copper,
including the leads, is nickel- and gold-plated to generate a corrosion barrier and a
bondable surface, respectively.
Figure 3.15 shows a close-up of tape that has been patterned and plated. Bumped
tab leads are shown in Figure 3.16. The die and the substrate or chip carrier are
connected to bumps of copper and/or plating at the ends of the leads by a process
similar to resistance welding. After the die has been bonded, it can be probed and
tested while still in tape form. Figure 3.17 shows a tape-automated-bonded die. Note
that the leads fan out to larger probing pads, which enable the die to be functionally
tested. These probe pads are cut off when the die and its bonds are punched out of the
tape. After removal from the tape, the leads must be formed, or bent, to be mounted
and bonded to the chip carrier or substrates.
TAB has several advantages: The bonding area is much larger than that of a ball
or wedge bond; the lead itself provides a larger cross-sectional area for carrying the
current and power; and the copper also gives one of the highest possible thermal
conductivities and current capacities. The disadvantages of TAB are the time and
cost of designing and fabricating the tape and the capital expense of the bonding

FIGURE 3.15 Patterned tape for TAB. (Courtesy of Westinghouse Electronics Systems
Group.)

DK565X_003_r04.indd 104 11/11/2005 7:40:20 PM


Electronics Assembly 105

FIGURE 3.16 Bumped TAB leads. (Courtesy of Westinghouse Electronics Systems


Group.)

FIGURE 3.17 Tape–automated-bonded die. (Courtesy of Westinghouse Electronics


Systems Group.)

DK565X_003_r04.indd 105 11/11/2005 7:40:21 PM


106 Assembly Process: Finishing, Packaging, and Automation

equipment. Each die must have its own tape, patterned for its bonding configuration,
and each TAB design requires its own equipment program and setup. Consequently,
TAB has typically been limited to high-volume production applications.

Flip-Chip Bonding Flip-chip technology for interconnection of dies is gaining


popularity owing to the fact that the number of I/O connections per chip achieved using
flip-chip bonding is much greater than can be achieved using wirebonding or TAB.
The flip-chip uses a solder bump to connect the bond pad to the substrate
(Figure 3.18). The flip-chip assembly consists of a ball-limiting metallurgy (BLM)
on the chip bond pads, the solder bump, and a top surface metallurgy (TSM) on the
substrate bond pads. BLM is a multilayered structure, with an adhesive layer, a
barrier layer, and a bonding layer. Common materials for the BLM structure include
chromium and titanium for the adhesion layer; copper, palladium, platinum, and
nickel for the barrier layer; and gold for the bonding layer. Common materials for
TSM structures include nickel for the lower layer and gold for the upper bonding
layer. Common materials for the solder bump include 95 Pb/5 Sn, and 50 Pb/50

Al Pad
Die Cr

50/50 Cu/Cr
Cu
Solder
Die bump
Passivation Au

Ball limiting
metallurgy
(BLM)
Solder dam

Top surface
metallurgy Die
(TSM)

Substrate

FIGURE 3.18 Example of a flip-chip interconnect.

DK565X_003_r04.indd 106 11/11/2005 7:40:23 PM


Electronics Assembly 107

In solders. Solder bump diameters typically range from 4 to 10 mils. Typical solder
bump heights range from 2 to 8 mils. Minimum I/O pitches achieved using flip-chip
are in the neighborhood of 10 mils. The maximum number of I/Os bonded using
flip-chip interconnects are in the neighborhood of 700 or greater.
Substrate Mounting A substrate assembly can be mounted mechanically, with
epoxy, or with solder.
Mechanical Mounting Mechanical mounting requires the engineer to attach the
substrate assembly to a carrier plate that supports screws used to mount the substrate
to the case. The screw support is usually a flange that extends out from under the
substrate. Because ceramic cannot withstand the torque or stress a screw would
impose on it, the ceramic is first attached to a metal carrier plate, and then the carrier
is screwed into the case. The ceramic is formed or machined to provide either a
rounded notch out of a corner or a hole in the center for the screw. This method is not
commonly used for military hybrids, because a screw that penetrates the case floor
would result in a loss of hermeticity. Furthermore, a floor thick enough to support the
screws would counteract the weight and volume reduction offered by hybridization.
In commercial applications, substrates are mounted to carrier plates, protectively
overcoated, and then mechanically mounted within the systems because this method
provides for quick and easy replacement of the substrate assembly; the associated
cost is increased weight and volume, with decreased hermeticity.
Epoxy Mounting Epoxy mounting can be done with epoxy paste, but this requires
extensive process control of printing and curing. If the printed deposit is not smooth
and uniform, air pockets or voids can form when the substrate is mounted, while
curing epoxy outgasses the volatile solvents used to liquefy it. As outgassing occurs
under the larger area of a substrate, pockets of gas can accumulate, forming voids
that decrease adhesion and heat-transferring area. If a void is located directly under
a hot component, the thermal resistance is greatly increased, and thermal failure can
occur. Epoxy preforms, which resolve these problems, are therefore the norm. The
preform can be cut or punched to the desired size and shape and comes in a very
uniform thickness. The preform is placed in the case with the substrate on top of it,
and pressure is applied by placing a weight (a few hundred grams) on top of the
substrate, if component spacing allows, or by clamping down the ends of the substrate.
The unit is placed in a vacuum or nitrogen-purged oven to cure it; vacuum ovens are
recommended for their ability to extract outgassed material. The door to the oven
must not be opened at any time during the cure; this would greatly affect the profile,
or epoxy cure schedule. If the epoxy does not reach cure temperature in the right
amount of time, complete molecular interlocking cannot be obtained. If the ramp up
to the cure temperature is prolonged because heat escapes from the oven, the organics
that promote the interlocking can volatilize before the process is complete, reducing
the adhesive strength of the epoxy.
Solder Mounting Another way to mount substrates is with a solder preform, using a
low-temperature solder, such as a lead–tin composition. The case is typically gold-plated
to allow solder wetting. The back side of the substrate is also metallized. If the substrate

DK565X_003_r04.indd 107 11/11/2005 7:40:23 PM


108 Assembly Process: Finishing, Packaging, and Automation

is thick-film alumina, the metallization can be a thick-film paste of platinum–gold,


platinum–palladium–gold, or silver. Gold paste is not used because of the tendency of
gold to leach off the ceramic surface into the solder, forming intermetallics; this mass
transfer accelerates at elevated temperatures (upper operating conditions) until all of the
gold has diffused into the solder and none is left to maintain adhesion to the ceramic.
The addition of platinum or palladium to the gold alloy will prevent substantial leaching.
If the substrate is beryllium–oxygen, thin-film metallization must be used.
The actual process flow for solder mounting of a substrate varies for different
companies and applications. One approach is to reflow the solder on the floor of the
case with flux and a hot plate. Then the case is defluxed, typically in a degreaser. The
substrate backplane is burnished to roughen the surface and increase surface tension
and wetting. The substrate is then placed on the solder and scrubbed into place while
reflowing the solder on a hot plate under a dry nitrogen flow or in an environmen-
tal glove box, as shown in Figure 3.19. The substrate can also be reflowed in place
by clamping the preform and substrate into the case and sending the unit through a
forming gas furnace. This method cannot be used if thick-film resistors are on the
substrate being mounted; exposing these thick-film resistors to a reducing atmosphere
can cause changes of up to 600% in the resistor value because it reduces the oxides
comprising the resistor paste.

FIGURE 3.19 Substrate mounting in a glove box. (Courtesy of Westinghouse Electronics


Group.)

DK565X_003_r04.indd 108 11/11/2005 7:40:23 PM


Electronics Assembly 109

After the substrate is in place, the I/Os are wirebonded to the case leads. A final
internal visual is performed and full functional electrical tests are conducted. Figure 3.20
shows examples of delidded hybrid assemblies.
Biomedical Hybrids Biomedical hybrids are usually not mounted to any carrier
plate or case; they are implanted, obviating the need for a carrier plate. Mounting
within a metal case would increase the volume and produce package corners, which
could irritate the unit’s surroundings. Although metal cases provide a hermetic seal
against air and moisture, they would not protect the unit from the hostile environment
of body fluids (acids, enzymes, and other organics). Biomedical hybrids must be
protectively overcoated with a polymeric material that will be inert in its environment
yet withstand the chemical hostility of its surroundings.
Biomedical companies have found that by eliminating all contaminants prior to and
during the application of the protective overcoat and by carefully controlling the curing
process, complete and total adhesion can be obtained. The entire surface is coated; no
voids or pockets are formed along the substrate assembly surface. Although the coating
will absorb moisture from its surroundings, the moisture has no place to go; it penetrates
to a certain depth of the coating and stops since there is no void to which the moisture
can mass-transfer and condense. The moisture remains trapped in the overcoat material;
once the surface is saturated, it absorbs no more moisture (Troyk, 1987).

FIGURE 3.20 Hybrid assemblies. (Courtesy of Westinghouse Electronics Systems Group.)

DK565X_003_r04.indd 109 11/11/2005 7:40:24 PM


110 Assembly Process: Finishing, Packaging, and Automation

Thermal stress in a biomedical hybrid is not as great as in military and most


commercial applications; although the temperature is slightly elevated (98.6°F), it is
fairly constant. The maximum temperature gradient that a biomedical device would
experience is only a few degrees. Thus, thermal management is not a crucial issue
for these devices.

Package Sealing
Introduction to Package Sealing Once the hybrid passes precap inspection and
electrical testing, it is sealed. The most common sealing methods are welding
and solder sealing.
Welding In resistance welding, an electrode rolls over the edges of the lid while
current and pressure are applied. The current heats the lid and case because of the
resistance of the case material, so this method works only on cases and lids made of
materials with high resistivities, such as Kovar or stainless steel. If the case itself is
made of a copper alloy for thermal and grounding purposes, then a Kovar or stainless
steel seal ring can be brazed to the top of the case side walls to provide the necessary
resistance.
Another welding method uses a laser to hermetically seal the lid and case together.
Laser welding can be done on a wide variety of materials, including aluminum–silicon
alloys, ferrous alloys, and nickel alloys. Figure 3.21 shows the different lid configura-
tions used for soldering and welding.
Lid Deflection Lid deflection must be taken into account when designing an
electronic package, whether for a hybrid or for a module. The metal must be thin
enough to allow the lid and seal ring surface to reach welding temperature without the
entire case being heated. Usually, the lids used for resistance welding have a outer
perimeter thickness of 0.005 in., while the center metal is 0.01 to 0.015 in. thick, as
shown in Figure 3.21. This added center thickness gives the lid more rigidity. However,
it still allows lid deflection to occur during operation and testing conditions.
Example In military applications, hybrids must pass various environmental tests,
including leak testing, centrifuging, and vibrational testing. These tests impose
stresses that can cause lid deflection. For example, when a unit is leak tested, the
sealed package is exposed to a differential pressure of 2 atm (29.4 psi). If a standard
1 in.  2 in. hybrid is leak tested, the total deflection for a uniformly distributed load
can be expressed as

r 4
y  k1
Et 3

where y is the total deflection, E is Young’s modulus of the lid material, r is the width
of the lid, R is the length of the lid, t is the lid thickness, v is a uniformly distributed
load in psi, and k1 is a coefficient dependent on the ratio of R to r and the way the lid
is supported (Avallone and Baumeister, 1987).

DK565X_003_r04.indd 110 11/11/2005 7:40:24 PM


Electronics Assembly 111

0.015" ± .001"
0.005"
.003

* THE RADIUS IS TYPICALLY


0.000 - 0.030" BIGGER THEN
THE INSIDE RADIUS OF THE
A CORRESPONDING CASE

L = SAME LENGTH AS CASE


W = SAME WIDTH AS CASE
* A = CASE WALL THICKNESS 
0.010" TO 0.015"
W

L
ROLLER

CASE WALL

WEDGE JOINT

RESISTANCE WELDING
TYPICALLY USED ON HYBRIDS

(a)

FIGURE 3.21(a)

A standard 1 in.  2 in. hybrid would have a 1 in.  2 in.  0.015 in. Kovar lid.
Thus r  1 in., R  2 in., t  0.015 in., and E  20  106 psi (see Table 3.4). This
lid is welded or attached to the case, implying that the lid is fixed on all edges.
Using a fixed support model and given that R/r  2, k1  0.0277 (Avallone and
Baumeister, 1987), substituting the given information into the equation yields the
following:

29.4 14
y  0.0277  0.012 in.
20, 000, 000  0.0153

DK565X_003_r04.indd 111 11/11/2005 7:40:24 PM


112 Assembly Process: Finishing, Packaging, and Automation

tL = tC

0.010
(tL  .010)
0.000

RL = RC

0.010
TL = (TC  0.010)
WL 0.010

R 0.030 MIN .00


TVP LL = (LC  2A) .005
.000
WL = (WC  2A) .005

LL

LASER
BEAM
A
tC

TC

WELD JOINT

(b) LASER WELDING

FIGURE 3.21(b) Examples of (a) sealing lids and (b) solder sealing for hybrid packages.

DK565X_003_r04.indd 112 11/11/2005 7:40:24 PM


DK565X_003_r04.indd 113
TABLE 3.4 Typical Properties of Metallic and Ceramic Materials Commonly Used in Frenchtown Ceramic-Metal Brazed
Assemblies
Electronics Assembly

Metallic Materials
Stainless OFHC
Property Steel Steel 304 Kovar Alloy 42 Nickel Molybdenum (copper) Wcu10
— Fe 67 Fe 54 — — — — —

Composition by weight % Fe Cr 20 Co 17 Fe 58 Ni 99.0 MO CU Cu 10


99+ Ni 10 Ni 29 Fe 42 min. 99.9 99.96 W 90
Specific gravity (g/cm3) 8.7 7.8 8.4 8.2 8.8 10.2 8.9 17.2
Hardness Rockwell 50 90 70–85 60–80 50–80 55–60 10–45 30
(B, C, or 45N scale) B B B B B B B C
Tensile strength ( 103 psi) 55 85 70 60 60 170 30 125
Compressive strength ( 103 psi) — — — — — — — —
Young’s modulus of elasticity ( 106 psi) 28 28 20 20 30 50 17 37
Coefficient of linear thermal expansion 140 190 53 73 150 56 180 60
( 107/°C, 20–500ºC)
Thermal conductivity (cal cm/cm2 see ºC) 20°C 0.14 0.33 0.040 0.036 0.204 0.380 0.940 0.50
100ºC 0.16 0.39 0.043 0.041 0.200 0.330 0.940 0.49
Specific heat (cal/gºC) — 0.12 0.104 0.120 0.130 0.060 0.092 0.039
Electrical conductivity — — — — — — 100 40

(Continued)
113

11/11/2005 7:40:25 PM
114

DK565X_003_r04.indd 114
TABLE 3.4 Typical Properties of Metallic and Ceramic Materials Commonly Used in Frenchtown Ceramic-Metal Brazed
Assemblies (Continued)
Ceramic Materials
Property Alumina FA-94 Alumina 2082 Beryllia Thermal
Composition by weight % Al2O3 Al2O3 BeOa
94 94 99.5
Specific gravity (g/cm3) 3.63 3.60 2.80
Hardness Rockwell 80 80 60
(B, C, or 45N scale) 45N 45N 45N
Tensile strength ( 103 psi) 35 33 22
Compressive strength 410 410 225
( 10 3 psi)
Young’s modulus of elasticity 45 45 21
( 10 6 psi)
Coefficient of linear thermal 71 70 73
expansion ( 107/°C, 20–500ºC)
Thermal conductivity 0.05 0.05 0.55
(cal cm/cm2 see ºC) 20°C
100ºC 0.04 0.04 0.40
Specific heat (cal/gºC) 0.19 0.19 0.25
Electrical conductivity — — —

a These materials are most often used by Frenchtown Ceramics to produce ceramic-metal brazed assemblies. Other materials are available.
b Registered trademark, Brush Wellman.
Courtesy of Frenchtown Ceramics, Inc., Frenchtown, N.J.
Assembly Process: Finishing, Packaging, and Automation

11/11/2005 7:40:25 PM
Electronics Assembly 115

Checking this number against deflection measured during leak testing, the calcu-
lated value is much smaller than the measured value of 0.03 in. Further investigation
indicates that the case walls are not totally rigid; they can flex at the floor of the case
during leak testing. This suggests that the lid is not totally fixed.
Recalculating for a simply supported model changes k1 from 0.0277 to 0.1106
(Avallone and Baumeister, 1987), which in turn gives a deflection of 0.048 in. This
does indeed indicate that the actual lid deflection (0.03 in.) is between that of a fixed
lid (0.012 in.) and that of a simply supported lid (0.048 in.). For future calculations
of this hybrid’s deflections, a simply supported model should be used to ensure that
the calculated value always reflects the worst-case situation.
This example demonstrates that lid deflection must be accounted for when design-
ing a hybrid. To ensure that the lid does not come into contact with any components,
potentially creating shorts or damaging wirebonds, the bottom of the lid must be
0.048 in. above the highest component or wirebond in the hybrid.

Standoff Posts If the additional height needed to allow for lid deflection is not
available because of system height limitations, the design must be modified in
another way to compensate for lid deflection. One approach is to use a solder seal
or laser weld so the lid can be thicker and experience less deflection. Another
possibility is to reinforce the lid by adding ridges or dimples. The most common
solution, however, is to incorporate standoff posts, or spacers, in the assembly. This
can be done by mounting posts on the surface of the substrate assembly that support
the lid and limit its deflection. Care must be taken when designing these posts.
They must support the lid without deforming it (pointed posts could dent or even
puncture the lid during leak testing). They must also account for stack-up tolerances
to adequately limit deflection without exerting a force on the lid during sealing.
Another necessity in standoff post design is accounting for all the forces and
resulting deflections to which the case will be exposed. To continue the earlier
example, a 1 in.  2 in. hybrid tested per MIL-H-38534 will be exposed to a 2 atm
pressure differential during leak testing, 5000g during centrifuging, and vibration
at 200g rms during particle impact noise detection (PIND) testing. The previous
example indicated a worst-case deflection of 0.048 in. Similar calculations yield
expected worst-case deflections of 0.037 in. and 0.0015 in. for centrifuging and
PIND, respectively. These calculations yield a maximum expected deflection of
0.048 in. during leak testing. Because the lid will vibrate 0.0015 in. during PIND,
the posts must be designed to prevent the lid from contacting the components.
This design will involve a complex structural analysis to determine what number
of posts at what height and placement will accommodate the vibration potential.
The posts must be short enough to remain below the vibration range so that the lid
will not tap against the posts during PIND and cause a noise detection failure.

3.5.4 Hybrid Packaging


Introduction to Hybrid Packaging Once the substrate assembly, mounting, and
wirebonding are completed, the substrate is ready to be mounted in the package or

DK565X_003_r04.indd 115 11/11/2005 7:40:25 PM


116 Assembly Process: Finishing, Packaging, and Automation

case. Typical hybrid cases are metal with glass or ceramic feedthroughs, or ceramic
with brazed-on leads.

Metal Packaging
Introduction to Metal Packaging Metal cases are fabricated in three ways. The
first is two-piece construction with a Kovar plate brazed to a window frame of Kovar.
Both pieces are nickel and gold plated prior to brazing. The window frame is typically
machined out of a solid piece of metal. A second technique for fabricating a metal
package is to machine a bathtub out of a solid piece of metal. A third fabrication
method molds the metal in the shape of a bathtub. However the side walls of a gull-
winged leaded case are formed—by brazing, machining, or molding—the side walls
must have holes for feedthroughs drilled prior to plating. For insertion-mounted
hybrids, the holes are drilled in the bottom or floor of the case along the inside
perimeter.

Glass Seals To fabricate a matched seal, a donut of glass or a glass bead is inserted
into the unplated holes in the window frame or case; an unplated Kovar lead is placed
in the center of the donut. The case is sent, with glass beads and leads in place,
through a furnace profiled to allow the glass to melt. The glass reflows and fuses with
the oxide on the surface of the Kovar. A mild etch or cleaning method is used to
prepare the remaining Kovar surface for plating. The glass used in this process must
have a coefficient of thermal expansion (CTE) close, or equal, to that of the Kovar.
Figure 3.22 shows a side view of a matched glass seal.
In a compression glass seal, another common seal, the glass is reflowed to a Kovar
eyelet or ferrule. The lead and ferrule are plated as usual. Finished feedthroughs
are brazed into the side walls of the case. Figure 3.23 shows the machined case, the
feedthroughs, the solder preforms, the fixturing used to braze the feedthroughs into the
case, and a finished metal case with compression feedthroughs. The fixturing used for
the brazing is typically graphite, which can withstand the brazing temperatures without
losing its shape, and will not adhere to the brazing material. The case wall material can
be Kovar or, in power applications, a copper alloy. The ferrule and the glass are put in
compression because the expansion rate of the surrounding metal is greater than that of
the glass. Glasses typically used in feedthroughs have a design tensile strength of only
1,000 psi; tensile strengths of approximately 2,000 psi may be obtained with certain
glass compositions. The strength can be further increased to 3,000 psi by annealing
the glass or to 20,000 psi by tempering it (Stamps, 1990). Although these strengthen-
ing methods cannot be used for feedthroughs, the strength of the glass can be greatly
increased by placing it in compression. Actual compression strength measurements
are difficult to obtain; the glass usually exhibits tensile failure before the compres-
sion strength is gauged because of slight bending or torquing during the measurement.
Theoretically, the compression strength is in the 100 ksi range. While such a case con-
figuration involves more processing, the glass seals are much stronger.
Glass-to-metal cases have been used in the industry for decades. They are readily
available, use proven technology, are relatively inexpensive, and provide hermeticity.

DK565X_003_r04.indd 116 11/11/2005 7:40:25 PM


Electronics Assembly 117

CERAMIC FEEDTHROUGH

KOVAR OR
COPPER CORED
ALLOYS

COMPRESSION CERAMIC
FEEDTHROUGH

FIGURE 3.22 Various feedthrough configurations.

However, glass forms a thin, brittle, uncompressed meniscus during the reflow process
(see Figure 3.22), so the strength of the glass within the meniscus is low. Glass-to-
metal seals are plagued by meniscus cracking due to thermal and mechanical stresses
applied during temperature cycling and chip-out caused by lead forming or centrifugal
forces. When the glass chips out in the meniscus area, the lead base metal, typically
unplated Kovar, is exposed and becomes a corrosion site. If the chip-out occurs inside
a hermetically sealed package, corrosion forms much more slowly; however, on the
exterior of the case, a chip-out will eventually corrode. Cracks can lead to chip-outs
and, in the matched glass configuration, can propagate, theoretically producing a loss

DK565X_003_r04.indd 117 11/11/2005 7:40:25 PM


118 Assembly Process: Finishing, Packaging, and Automation

FIGURE 3.23 Compression feedthrough case components and fixturing. (Courtesy of


Westinghouse Electronics Systems Group.)

of hermeticity. The military has established visual inspection criteria specifying cracks
and chip-outs as rejectable.
Some methods for repairing glass chip-outs have been tried. The most common
is to cover the chipped-out area with an epoxy or overcoat. One argument against
this practice is that the epoxy or overcoat would not be sealed within a hermetic
atmosphere and would absorb moisture that would eventually migrate and reach the
Kovar, leading to corrosion or dendritic growth. A second argument is that the epoxy
would make it impossible to see the extent of the chip-outs, cracks, or other possible
problems, such as foreign material, embedded material in the glass, or glass overrun.
The military has thus rejected this repair process.
Another method sometimes attempted is to brush-plate the exposed base metal
with gold. Unfortunately, the gold by itself, without a nickel underplating, does not
provide a corrosion barrier; thus this method only postpones the development of
a corrosion site. Furthermore, the gold in direct contact with the current-carrying
Kovar lead sets up an electrolytic reaction that acts like a tiny battery and can result
in even more extensive corrosion.

Ceramic Feedthroughs In an effort to eliminate cracked-glass yield problems,


U.S. Department of Defense contractors incorporate ceramic feedthroughs. Ceramic
and metal packages with ceramic feedthroughs have been used in space applications

DK565X_003_r04.indd 118 11/11/2005 7:40:26 PM


Electronics Assembly 119

for many years. Ceramic is much stronger than glass; 95% alumina, for example, has
a tensile strength of 30,000 psi and a compression strength of 300,000 psi, making a
ceramic feedthrough an order of magnitude stronger than a glass feedthrough
(Stamps, 1990). Furthermore, it does not have a meniscus. Ceramic feedthroughs are
formed by metallizing the inside hole and outer ring of a ceramic donut, then brazing
in the lead. These feedthroughs can then be brazed into the case wall at lower
temperatures.
Figure 3.22 depicts the differences between glass and ceramic feedthroughs. Ceramic
feedthroughs offer greater strength and the absence of a meniscus, with its associated
cracking and chip-out problems. Because of the increased strength of the material, less
metal must surround the feedthrough to ensure the package’s mechanical integrity. With
glass seals, MIL-H-38534 requires that there be at least 0.04 in. of metal between the
glass and the seal surface to protect the glass from stress during resistance welding. No
standard for ceramic is given in MIL-H-38534; however, ceramic feedthrough packages
used for space applications typically measure under 0.040 in. from the ceramic to the
seal surface. Of course, any package used for military or space applications must be
qualified for very stringent environments and testing levels.
Although ceramic is stronger than glass and better able to withstand the stresses
imposed during processing, handling, operation, and testing, ceramic feedthrough
packages have disadvantages. They are more expensive than glass feedthrough pack-
ages, in which the feedthrough and lead are inserted into the case in one step, in a
furnace programmed to reflow the glass. Ceramic feedthroughs, in contrast, are much
more labor intensive. The ceramic must be formed into donuts by machining a sheet
of fired ceramic the same thickness as the desired donut height by laser drilling or
ultrasonic milling.
As the laser cuts through the ceramic, it melts the material, or slag, splattering it
onto the surrounding surface. Slag forms burrs with very smooth, glasslike surfaces
to which metallization will not adhere properly. To prepare the ceramic donut for
metallization, the surfaces must be cleaned of slag. One cleaning method is sand-
blasting, in which abrasive slurry is forced through the holes under high pressure.
An etching process is sometimes used, but this can lead to surface pitting. Another
method, developed by Laserage, a laser machining company, involves heat-treating
the laser-machined ceramic (Capp and Luther), which leaves the machined surfaces
with a finish similar to that of fired ceramic. Once the ceramic is prepared, it must
be metallized both inside the hole and on the outer perimeter. This can be done by
painting on a thick-film metal paste and firing it to the ceramic.
Another method of forming ceramic donuts is to mold them into shape while the
ceramic is still green, or nonfired. Before firing, a refractory metal is applied to the
donut hole and perimeter, then the unit is cofired. The metal must be nickel plated to
provide a corrosion barrier, and the nickel must in turn be plated with a solderable
or brazable metal, typically gold. Once the ceramic donut is formed and metallized,
leads or pins must be brazed into the donut holes, and the feedthrough must be brazed
or soldered into the metal package. In some cases, both brazing steps are performed
together, but this requires special tooling to hold the leads and donuts in place during
the process.

DK565X_003_r04.indd 119 11/11/2005 7:40:26 PM


120 Assembly Process: Finishing, Packaging, and Automation

Not only are ceramic feedthrough cases more labor intensive than using glass
cases, their manufacture also requires higher temperatures. Because ceramic firing
and brazing temperatures are much higher than glass reflow temperatures, more
sophisticated furnaces are required. The additional labor, higher material cost, and
elaborate equipment needed all make ceramic feedthrough cases more expensive
than their glass counterparts.
Ceramic also has different electrical properties than the standard glass used in
electronic packaging, which may affect the package design, depending on the electri-
cal requirements of the unit. For example, ceramic may enhance high-voltage pack-
ages because its porosity keeps voltage from traveling across the ceramic insulator
and shorting to the case. However, in analog or microwave applications, ceramic
feedthroughs may be undesirable because of their higher dielectric constant. The
dielectric constant of the insulating material will determine the impedance of the
feedthrough. The impedance of a coaxial cable or a feedthrough is

60 b
Z ln
 a

where Z is the characteristic impedance in ohms (),  is the dielectric constant of


the insulating material (i.e., the glass or ceramic), b is the radius of the insulating
material, and a is the radius of the lead (Leahy, 1989).
The dielectric constant of the glass used for feedthroughs is approximately 5;
dielectric constants of ceramics are higher [e.g., 96–99% alumina ceramics have
dielectric constants of approximately 9–10 (Leahy, 1989)]. If a glass feedthrough
were converted to ceramic with no change in dimensions, the impedance would
be decreased.
Ceramic Packaging Cofired ceramic is a frequently used alternative case material.
Although they are manufactured the same way as ceramic chip carriers, cofired cases
are larger and usually come with top-brazed leads on two sides (full-winged lead
configuration).
Ceramic cases can have an interior-mounted substrate, or the base of the case
can act as a multilayered substrate (an integral substrate package, an example of
which is shown in Figure 3.24). In medium- to high-volume applications (e.g.,
orders of approximately 1000 or more cases), integral substrate packaging can
greatly reduce manufacturing costs. Once the electrical design is fixed and the
vendor has tooled up to make the case, these cases can be produced less expen-
sively than metal cases.
In addition to reducing the cost, an integral substrate case eliminates the need to
fabricate a substrate, which lowers the hybrid’s material and labor costs. The integral
substrate also eliminates the need to mount the substrate in a case and wirebond it
to the package leads; interconnection is internal to the multiple layers of the case,
further reducing labor costs.
Other advantages of ceramic packaging are improved yields and reliability.
Ceramic cases eliminate yield problems associated with glass feedthrough cracking

DK565X_003_r04.indd 120 11/11/2005 7:40:26 PM


Electronics Assembly 121

FIGURE 3.24 Integral substrate ceramic hybrid case. (Courtesy of Kyocera America.)

and chip-outs. The ceramic package is appealing for space and military applications
because of its lower weight and noncorrosiveness. As with ceramic feedthroughs,
conversion from metal to all-ceramic cases is not always practical. Again, in analog
or microwave devices, the refractory metals used to fabricate cofired ceramic pack-
ages may not carry current fast enough, so the unit cannot operate at the neces-
sary frequencies. Furthermore, the dielectric constant of the ceramic package may
affect the electrical performance of a microwave device, causing an impedance mis-
match. Microwave packaging design is an example of the critical need for concurrent
mechanical, electrical, and material engineering.

Customization of Electronic Packaging Many designs dictate specific requirements


that cannot be met with standard metal or ceramic cases. High-power hybrids might

DK565X_003_r04.indd 121 11/11/2005 7:40:26 PM


122 Assembly Process: Finishing, Packaging, and Automation

utilize a case with a heat-spreading beryllium–oxygen bottom and a weldable Kovar


window frame; others might have a metal base, ceramic side walls, special grounding
requirements, and so on. Ceramic and metal combinations are frequently used. Just as
the hybrid itself has a customized electronic function, hybrid packaging often requires
customized cases to meet specifications.
The ceramics most commonly used in production today are 92–99% alumina
and 99% beryllia. Common packaging metals include Kovar, copper alloys, molyb-
denum, copper- or nickel-clad molybdenum, copper-clad Invar, copper tungsten,
and aluminum. These materials can be combined through special techniques and
advanced materials to meet packaging requirements. Table 3.4 lists the properties of
commonly used industrial ceramics and metals.
Special compliant adhesives can absorb the mismatch in thermal expansion
between the metals and the ceramic. Other methods of attaching metals to ceramic
have been developed, such as direct bonding methods.
Research is being conducted with an increasing number of materials that offer
even higher thermal conductivities, increased strength, better-matched coefficient of
thermal expansion (CTE), lighter weight, and so forth. Today’s leading-edge materi-
als include aluminum nitride, silicon carbide, A40 (silicon–aluminum material), and
other metal-matrix and composite materials.

3.6 PRINTED WIRING BOARD ASSEMBLIES


3.6.1 Introduction to Printed Wiring Board Assemblies
There are various kinds of printed wiring boards (PWBs) or printed circuit boards. The
oldest and most common are organic boards. Ceramic boards, offering better thermal
properties with closer CTE matching to the components, have been around for more
than 20 years but recently have been used increasingly in high-power applications, such
as standard electronic module (SEM) power supplies. The combination of polyamide
and integrated circuit technology has produced high-density surface-mount boards with
finer line definition of less than 0.001 in. Another board fabrication method directly
bonds copper foil to ceramic; directly bonded copper enables vast improvements in
power dissipation.
PWBs can use mixed technology; they can have both insertion-mounted compo-
nents, plugged into the PWB, and surface-mounted components on top of the board.
However, a single component must be either inserted or surface mounted.
All PWBs have similar assembly capabilities, but the type of board will limit
the type of assembly. An organic board with through-hole intralayer routing can be
assembled either by insertion or surface mounting, depending on the components
and the layout design. Ceramic boards are only surface mounted. However, ceramic
boards can easily be made double sided by metallizing both sides or by sandwiching
two boards around a carrier plate or heat sink.
Direct-bond copper offers strong, lightweight boards for high-power applications
but has limited layering and line definition. Another option, using thin-film polyam-
ide boards, is capable of supporting bare-die applications.

DK565X_003_r04.indd 122 11/11/2005 7:40:28 PM


Electronics Assembly 123

3.6.2 Organic PWB Assemblies


Introduction to Organic PWB Assemblies
Organic boards are composite structures; for example, in epoxy glass boards, glass
fibers are suspended in an epoxy resin. Polyamide boards are composed of glass
fabric, with polyamide as the dielectric material. These are the most common indus-
try combinations. The individual layers of organic boards are metallized with copper,
have patterns etched on them, and then are laminated together. Interconnection
between the layers is accomplished with through-holes, blind vias, or buried vias.
The method of intralayer connection and the design of the mounting pads determine
the type of assembly used. If only insertion-mounted components are used, then all
connections from the components to the board are through-holes; the mounting pads
are actually holes in the board. Routing signals from one component to another can
be accomplished by through-hole, blind via, or buried via. Because the assembled
board will have holes with leads extending through them, insertion boards typically
do not have components mounted on both sides. Any components mounted on the
bottom of the board are typically small, hermetic chip components, mounted with
epoxy to hold them in place as they travel through the solder wave.
Boards routed with buried vias have limited reworkability and lower yields, but
they offer increased surface-mounting area and fewer routing layers. These surface-
mount designs can also be routed to allow double-sided mounting. Surface-mount
boards must be laid out to include mounting pads of the right size, component toler-
ances, and proper solder fillets. These boards often have both leadless and leaded
components; the latter require surface-mount lead forming.
Electrical parameters must also be taken into account during board layout.
Components with a high voltage potential between them should be separated to pre-
vent arcing. When an arc occurs, the voltage travels across the board material and
can burn a trail, or carbon track, in the polymer on the board surface. If the arcing
itself does not short out the circuit or damage the components, the carbon track left
behind can cause shorts later on. Carbon tracking can be avoided by putting addi-
tional space between all tracks and components, but size limitations usually preclude
this solution. The more common prevention method is to give the board assemblies
a protective overcoat that reduces the available surface path for the voltage.
Boards can also be designed to accommodate both inserted and surface-mounted
components together. The layout of mixed-technology boards greatly influences the
assembly processing. If all inserted components are mounted on one side, low-profile,
hermetic, surface-mounted chip components can be tacked down, or epoxied, to the
bottom side, and reflow can be accomplished with wave solder. Therefore, mixed-
technology processing is similar to that for insertion mounting. If inserted and
surface-mounted components are placed on the same side, the processing is similar
to surface mounting.

Heat Sink Attachment For low-power commercial applications, components are


mounted directly on the PWB. For high-power or military applications, the first step
in the assembly process is to attach the PWB to a heat sink or carrier plate. With an

DK565X_003_r04.indd 123 11/11/2005 7:40:28 PM


124 Assembly Process: Finishing, Packaging, and Automation

insertion-mounted board, the carrier is either an edge support or a thermal ladder.


The edge support, which can be attached to the bottom or top of the PWB or wrapped
around its sides, braces the edge of the PWB, leaving the center free for device
placement. A thermal ladder is a heat sink attached to the top of the PWB with
windows or slots cut in it; the component (i.e., a plastic DIP, ceramic DIP, or cerpak)
straddles the ladder and the leads go through the slots into the PWB, as shown in
Figure 3.25. This design provides a thermal path directly from the ceramic component
to the heat sink, rather than through the leads.
Surface-mount boards typically use some edge supports, whether leadless or with
a configuration that permits both leads and package body to be mounted on the plane
of the board. Straddling a thermal ladder is difficult with these designs.
Heat sinks can also be attached to the bottom using a B-stage epoxy for organic
surface-mount PWBs. The PWB, epoxy film, and carrier plate are cured together
under controlled pressure and temperature. In some cases the PWB assembly is
double sided, sandwiching two boards over the heat sink; this cored board doubles the
surface-mountable area with a minimal increase in board assembly height or volume.
Although this design forces a thermal path through the PWB, the components are in
contact with the surface and assist with primary cooling, as depicted in Figure 3.25.
Thermal management of this design is enhanced by placing vias under the packages,
which are not only used for routing, but also act as thermal vias, to provide a more
efficient and direct path from component to heat sink.
Heat sink or core materials offer improved thermal conductivity over the PWB
material. These materials include aluminum, Kovar, copper-clad Invar, copper-clad
nickel, graphite composites, copper-clad molybdenum, and metal-matrix composites,
such as silicon–aluminum materials.
Aluminum has high thermal conductivity and is lightweight, less corrosive than
most metals, inexpensive, easy to machine, and readily available. Boards mounted to
aluminum heat sinks are often used for SEM modules. Copper has many of the same
properties as aluminum, but it is highly corrosive and heavy. Both aluminum and
copper have excessive CTEs compared with those of either organic or surface-mount
components. Copper-clad Invar offers the superior thermal and electrical conductiv-
ity of copper, while an Invar center gives the composite material a more desirable
CTE. Copper-clad Invar has a lower CTE than copper, but a higher one than Invar.
Bonding the board to the core material gives the assembly a CTE approaching that
of the leadless ceramic chip carriers (LCCCs). However, copper-clad Invar is many
times heavier and more expensive than aluminum, and is available only from Texas
Instruments or its sales representatives.
Graphite fiber composites offer high thermal conductivity, approaching that of
copper-clad material, yet are lightweight like aluminum. However, graphite is dif-
ficult to machine. Composites, like graphite and metal-matrix heat sinks, can have
thermal conductivity and CTE tailored to the assembly. Because the direction or
orientation of the fibers in a graphite composite directs the thermal path, resistance
along the path between the components and the heat sink can be reduced by orienting
the fibers in the same direction as the path. Likewise, controlling the composition of
the composite controls, to a degree, conductivity and CTE. For example, the thermal

DK565X_003_r04.indd 124 11/11/2005 7:40:28 PM


Electronics Assembly 125

THERMAL PATHS DIE

DIP
LEADS

THERMAL
LADDER

PWB

SOLDER
(a) JOINT

(b)

THERMAL VIA

LEADED CERAMIC
DIE CHIP CARRIER

PWB

HEAT SINK

(c)

FIGURE 3.25 Thermal paths. (a) Insertion-mount organic PWB with thermal ladder. Heat is
dissipated from the die through the dip package floor to the heat sink. (b) Insertion-mounted
organic PWB. Since the dip stands off the PWB, the heat must be transferred from the die, through
the dip package, through the leads, through the solder, to the PWB, out to an edge-support heat
sink. (c) Leaded surface-mounted organic PWB with cavity up chip carriers.

DK565X_003_r04.indd 125 11/11/2005 7:40:28 PM


126 Assembly Process: Finishing, Packaging, and Automation

conductivity of an aluminum–silicon composite can be increased by using more


aluminum—a better conductor—in the material, at the cost of an increase in CTE.
As the amount of aluminum is increased, both the conductivity and CTE move
toward those of aluminum. Thus, composite materials must be designed both metal-
lurgically and mechanically to optimize heat sink design. Research and develop-
ment of new composite materials is ongoing.
Most heat sinks can be attached with epoxy films, usually thermally conduc-
tive B-stage or prepreg sheets, or with other film adhesives. Material and adhesive
choices depend on the requirements of the system or PWB assembly. If cost and
weight are more critical than thermal management—that is, if the assembly will not
be subjected to severe or rapid temperature changes—aluminum attached with stan-
dard B stage is appropriate. If thermal management or power dissipation is critical
but there are no restrictions on weight or outgassing, then Invar copper attached with
a flexible silicon adhesive to absorb the CTE mismatch might be preferred. If thermal
management and weight are both critical, the answer might be a graphite heat sink
with silicone adhesive.
When military specifications are followed, the components of the PWB determine
heat sink and adhesive design. For example, a PWB assembly mounted inside a her-
metic module along with nonhermetic or bare-chip devices could not use a silicon adhe-
sive because of military outgassing limitations. The heat sink would not only have to
dissipate any heat exceeding the maximum junction temperature, it would also need a
CTE closely matching that of the PWB. In this case, the more expensive developmen-
tal composite materials would have to be considered, and a less flexible film adhesive
meeting outgassing limitations would have to be used. The components would have to
be designed or packaged to withstand their CTE mismatch with the PWB, either by
being small enough that the mismatch would have little effect, or by incorporating leads
formed to absorb the stresses caused by the CTE differential.
Insertion Mounting On an insertion-mounted board, all components and subassem-
blies must be leaded, and the leads must be configured or mechanically formed for
insertion into the mounting holes of the board. On pin-grid arrays, the leads are
already in the correct configuration for insertion, perpendicular to the bottom of the
package. Dual in-line packages (DIPs) are also designed for insertion, with the leads
along the side of the package, pointing down. The military standard configuration
specifies leads at a 6° angle from perpendicular.
Components such as resistors and capacitors are leaded round cylinders, with
leads formed into a right angle for insertion by pick-and-place equipment. This
machine sends a robotic arm to the bin, tape, or feeder to select the component,
picks it up by vacuum or tweezer grabbing, transports it to the lead former, where
it is pushed into a troughlike fixture that bends the leads as they are forced against
its side walls. Lead forming can also be done by two robotic arms that grab the
leads and turn to bend them appropriately. The holding arm then carries the lead-
formed part to the board, where it inserts or mounts it in a preprogrammed location,
as shown in Figure 3.26. Of course, there are many other ways of accomplishing
the lead forming. This is done within tight tolerances and within seconds. DIPs

DK565X_003_r04.indd 126 11/11/2005 7:40:29 PM


Electronics Assembly 127

FIGURE 3.26 Automatic component placement. (Courtesy of Westinghouse Electronics


Systems Group.)

are normally supplied in tubes or feeders for the placement equipment, with leads
already formed. Chip carriers and hybrids normally have preformed leads cre-
ated by arms that grab the leads and force them into the necessary L shape while
the package is tightly cradled. In high-volume applications, the feeders form the
leads. Figure 3.27 shows an example of lead forming. After the components have
been inserted, automatic lead trimmers cut and flatten the leads coming out of the
bottom of the board (see Figure 3.28).
To prepare the boards for component placement, the PWB is plated with a tin
flash that protects the copper metallization against corrosion, or a thick plate of
solder. Once all the components are inserted, the board is sent on a belt that pulls it
across a wave of flux foam and then a wave or waterfall of liquid solder. The bottom
of the board floats over these waves. The flux, activated by temperature, deposits
flux over the entire board, removes oxides on the surfaces to be soldered, and allows
solder to wet these surfaces and wick up the leads and into the holes, forming the
solder joints. Usually, components are mounted only on the top side of the board;
components mounted on the bottom must be small enough to ride in the waves with-
out compromising surface contact, and they must be attached to the bottom with an
epoxy so they will not fall off before the solder fillet interconnections are formed.
After soldering, the boards are cleaned and inspected prior to final electrical testing.
Figure 3.29 shows an example of an insertion-mounted PWB assembly.

DK565X_003_r04.indd 127 11/11/2005 7:40:29 PM


128 Assembly Process: Finishing, Packaging, and Automation

FIGURE 3.27 Lead forming. (Courtesy of Westinghouse Electronics Systems Group.)

Surface Mounting
Advantages Surface-mount components can be either leaded or leadless, such as
chip capacitors, chip resistors, or LCCs. Surface-mount organic boards can be routed
with any combination of through-holes, blind vias, or buried vias; PWBs can have
heat sinks attached directly to the back side of the board, because there is no need to
leave this area clear for lead protrusions.

DK565X_003_r04.indd 128 11/11/2005 7:40:29 PM


Electronics Assembly 129

(a)

PWB

(b) PLATED THRU HOLES

PWB

(c)

FIGURE 3.28 Insertion mounting and lead trimming. (a) Component and board;
(b) component inserted in a board with leads cut and clenched; (c) after wave solder.

The advantages of surface-mount technology over insertion mounting are reduc-


tions in routing layers, in the assembly’s profile, and in board area. Components
can be mounted on pads on the board’s surface, obviating the need for through-
holes for component mounting. In addition, board routing can be accomplished with
blind or buried vias, which also eliminate the need for through-holes and extra lay-
ering to allow for routing around the holes. Surface-mounted components can also
be reworked with an extractor, which heats only the component to be replaced and
avoids heating and reflowing the entire assembly. Figure 3.30 shows a component
being removed with an extractor.

DK565X_003_r04.indd 129 11/11/2005 7:40:30 PM


130 Assembly Process: Finishing, Packaging, and Automation

FIGURE 3.29 Insertion-mounted organic PWB assembly. (Courtesy of Westinghouse


Electronics Systems Group.)

FIGURE 3.30 Extractor rework equipment. (Courtesy of Westinghouse Electronics Systems


Group.)

DK565X_003_r04.indd 130 11/11/2005 7:40:31 PM


Electronics Assembly 131

Surface-mount components usually sit lower than their inserted counterparts and sit
directly on the board surface, not slightly above, as shown in Figure 3.25. They also offer
great reductions in board area compared with comparable DIPs or cerpaks, a decrease
further enhanced if leadless chip carriers are employed. For example, a 20-pin DIP has
0.1 in. center-to-center lead spacing or pitch, making it at least 1.1 in. long by approxi-
mately 0.25 in. wide; the area of the component itself is 0.275 in.2. The footprint for this
DIP would add an additional 0.15 in. to each side, so its required mounting area would
be 0.605 in.2. The body of the DIP would be approximately 0.25 in. thick and would
sit approximately 0.05 in. off the board, making its profile 0.3 in. overall. In contrast,
a 20-pin chip carrier has a body 0.3 in.2, with a height as low as 0.06 in. and I/Os along
the perimeter on all four sides. If the chip carrier is leaded, the footprint would extend
0.15 in. out from the body of the carrier, making the overall mounting area 0.62 or
0.36 in.2. The footprint of an LCC would extend only 0.05 in. out from its body, or
require a mounting area of only 0.16 in.2. Whether leaded or leadless, the chip carrier
would sit only 0.005 in. above the board. The leaded device would have an epoxy
preform under it, 0.003 to 0.005 in. in height, while the LCC would have a 0.005
in. solder joint standoff. The magnitude of the area reductions increases with pin
count.
With system trends toward increased complexity, decreased size, and cost effective-
ness, surface-mount designs offer great reductions in height, area, and system volume
and cost no more than inserted components. For example, a system requiring 30 PWB
assemblies uses ten large-scale integrated (LSI) circuits with I/O counts of 60 to 68 for
each assembly. Assuming that their footprints account for 60% of the needed board area
(the other 40% is required for spacing, small discrete components, and connector attach-
ment), an insertion PWB version would require a minimum area of 64.5 in.2, and each
board would be approximately 8 in.  8 in. They would also require center-to-center
spacings of 0.45 in. (0.3 in. component height, plus 0.04 in. board thickness, plus
0.01 in. clearance for bottom-side lead extension, plus 0.05 in. spacing between boards).
This same design, using surface-mount technology, would require boards of approxi-
mately 6 in.  4 in. or 5 in. square, with center-to-center spacings of 0.165 in. Thus,
an insertion-mount system of 30 such assemblies would require a minimum volume of
14 in.  8.5 in.  8.5 in., or 7.024 ft3, while the same system in a leaded surface-mount
design would require a minimum volume of only 5.1 in.  6.5 in. × 4.5 in., or 1.036 ft3.
This conversion would result in a 6.8:1 volume reduction. Figure 3.31 shows a leaded-
component, surface-mounted PWB assembly.
This reduction in volume at comparable cost can be further improved by convert-
ing to a leadless surface-mount design. Using the previous example, leadless board
assemblies would require the same spacing of 0.165 in. but would only have to be
4 in. 2. Thus, the system would need only 0.717 ft3, an overall reduction of 9.8:1. The
general rule of thumb to estimate the area reductions for conversion from insertion
mounting to leadless surface mounting is 8:1.
The price of leadless components is also comparable to or less than that of their
leaded counterparts. A leaded chip carrier is slightly more expensive than the leadless
version, since it requires brazing the leads onto the leadless version; leadless compo-
nents also allow larger placement tolerancing due to their ability to self-align.

DK565X_003_r04.indd 131 11/11/2005 7:40:33 PM


132 Assembly Process: Finishing, Packaging, and Automation

Surface-Mounted LCCCs on Organic PWBs The feasibility of surface mounting


leadless ceramic chip carriers on organic PWBs has been debated over the years.
While such designs can solve many sizing and spacing problems without requiring
leads to provide strain relief for the CTE mismatch between the PWBs and the
carriers, initial attempts at LCCC mounting to organic PWBs were discouraging.
Mounting small LCCCs, approximately 0.400 in.2 with 28 pins or less, to organic
PWBs had some limited success. However, as the LCCCs increase in size, the effect
of the CTE mismatch also increases; if the PWB assemblies are subjected to extreme
temperature changes, as required in military applications, the thermal stress imposed
by the mismatch can be great enough to cause catastrophic solder joint failures.
Consequently, only smaller LCCCs have been surface-mounted to organic PWBs in
military devices.
Cored Boards Attempts have been made to improve the CTE match between
LCCCs and PWBs by sandwiching the PWBs around a heat sink material, or core,
and laminating them with B-stage or film epoxy, using heat and pressure. The
resulting cored board has a composite CTE between those of the PWB material and
the core material; this composite CTE more closely matches the LCCC’s, but the
cored PWB dimensions are slightly changed. The core material, with a higher CTE
than the PWB, expands more than the PWB during lamination heating. As the cored
board cools, the core contracts more than the PWB normally would, causing some
shrinkage of the PWB. The artwork layer for the solder mask must be modified, as a

FIGURE 3.31 Leaded surface-mounted organic PWB. (Courtesy of Westinghouse


Electronics Systems Group.)

DK565X_003_r04.indd 132 11/11/2005 7:40:33 PM


Electronics Assembly 133

one-to-one image of the solder pads will be slightly larger than the pads on the board.
The adjustment will vary with the core material, the dimensions of the board, and the
CTEs of the board and the core material. An aluminum or copper-clad Invar-cored
board has to be adjusted to approximately 96–98% of its original size.
Components to Boards: Solder Interconnection Typically, eutectic Pb–Sn
solders with RMA (rosin mildly active) flux or 60 Pb–40 Sn alloys are used to
connect components to boards. These solders can withstand common commercial
and military operating temperature ranges yet reflow at substantially lower temperatures
(approximately 183°C) than most die-attach solder alloys. This eliminates the
potential for die-attach reflow during the surface-mount reflow process.
Solder Deposition Solder is applied in one of two ways. The first is to print solder
paste onto a tin or solder-flashed board (the copper-metallized boards have flashes or
coatings to prevent corrosion). Solder paste is normally wet printed 0.009 to 0.011 in.
thick, yielding reflowed solder joint heights of 0.005 to 0.006 in. that provide stress
relief during temperature cycling. Any additional print thickness would make
processing more difficult and cause solder balling and bridging (see Figure 3.32),
while only slightly increasing stress relief benefits.
Solder pastes can be screen or stencil printed. An 80-mesh screen with 9 mils of
emulsion will produce the desired thickness, but paste coverage on the mounting pads
will be nonuniform. Screen printing can leave too much solder on the screen, and the
wire mesh leaves peaks and valleys in the paste surface after printing. This uneven paste
distribution can result in solder bridging between two adjacent pads with excess paste
and in starved solder joints with too little paste. In extreme cases, LCCCs will not sit
parallel to the board after reflow or may be skewed against the footprint.
Self-alignment occurs during solder reflow, provided the LCCC has not been
bonded to the board with a thermally conductive adhesive for heat-transfer enhance-
ment. As the solder melts, its surface tension rises, increasing wetting (or adhesion)
to the PWB mounting pads and the LCCC castellations. These increases in surface
tension cause the solder joint to pull the LCCC into alignment over its footprint.

SOLDER
BALLS
STARVED
JOINT

BRIDGING
FIGURE 3.32 Solder joint defects.

DK565X_003_r04.indd 133 11/11/2005 7:40:33 PM


134 Assembly Process: Finishing, Packaging, and Automation

Self-alignment allows the process control placement window to open, improving


manufacturability. However, uneven paste distribution leans to nonuniform surface
tensions and inadequate self-alignment, leaving some larger and heavier LCCCs
skewed on the board.
Excessive solder paste can slump during prebake or reflow, spreading beyond the
footprint metallization. During reflow, only the solder wetting the surface is pulled
by surface tension onto the metallized pad; the surface tension within the remaining,
nonadhering solder causes it to form a ball. Solder balls, which come in various sizes,
can potentially roll around and lodge between two electrically uncommon tracks,
causing a short or initiating a dendritic growth site. Dendritic growth occurs when
moisture, condensing on the PWB, becomes polarized by the electromagnetic fields.
The polarized particles, or ions, will start to attract each other and form a chain,
which grows between conductive surfaces that attract the ions. If the ion chain grows
enough to bridge the conductive pads, it can carry enough low current to cause inter-
mittent or complete shorting.
Given these problems, screen-printed boards require substantial rework by a
trained operator who manually removes or adds solder to individual joints and takes
out any solder balls. This process is very labor intensive, so an alternative method is
recommended: replacing the wire mesh screen with a brass stencil. The stencil thick-
ness should be the desired wet print thickness—that is, a 0.01 in. stencil will yield
a uniform 0.01-inch-thick wet print. Because a more uniform wet print yields more
uniform solder joints, a stencil gives better print resolution, which makes alignment
easier and more accurate and decreases the formation of solder balls and bridges.
To further reduce solder balling and bridging, the solder paste is subjected to
a prebake prior to reflow. The prebake volatilizes some of the paste solvents that
provide a printable viscosity, making the paste tacky so it can set up. The prebake
must be controlled, and time limitations between printing, baking, and reflow must
be observed, to assure that only viscosity-lowering solvents are removed, not flux
and other reactive solvents needed for proper reflow and wetting. Best results are
obtained if all soldering steps occur within 4 h. Some solder pastes—for example,
Cermalloy—use organic wetting agents that cannot be exposed to elevated tempera-
ture prior to reflow. These need to be reflowed immediately following printing.
The second deposit method is to thick-plate solder onto the board. When plating
the solder to the desired thickness, all top-layer metal that is not to be plated is first
masked. This can be done in a variety of ways. To apply a permanent mask, a photo-
sensitive polymer is deposited over the entire surface. The coated surface is exposed
to a specific frequency of ultraviolet light through a photomask; the mask can expose
either the areas to be soldered or the areas to be masked, depending on whether a
positive or negative photoresist is used. The board is sent through a developer that
dissolves the unwanted polymer, leaving the desired areas masked. Another way to
apply a permanent solder mask is to cover the areas to be masked with a film cut into
the desired shapes and sizes. With either method, the solder mask must be capable of
withstanding the plating and soldering processes.
A temporary mask, another alternative, can be printed on with a stencil or screen.
The masking material is a dielectric that the solder will not wet, and it must be able to

DK565X_003_r04.indd 134 11/11/2005 7:40:34 PM


Electronics Assembly 135

withstand the soldering temperature and exposure to the flux. It must also be soluble
after reflow in a degreaser or other simple cleaning process. Temporary masking has
been accomplished by printing on dielectric or glass thick-film paste, which can be
printed on the board but left unfired. After the mask is printed and dried, the solder
paste can be printed. The mask can withstand soldering temperatures without curing.
Although these materials typically contain oxides that are attacked by the flux, only
the surface of the deposit is affected because they are thick-film printed. They can be
washed off with an isopropyl alcohol (IPA) rinse. One added advantage of a tempo-
rary mask is that as it is washed away, so is the flux left on its surface. Disadvantages
are that this method can be used only when the solder is printed on the boards as a
paste, and any unsoldered top metallization must be protected from corrosion. Thus,
the copper must be tin or nickel plated prior to masking, or the top conductor layer
must contain only solder pads, tracks that can be solder coated, such as digital tracks
or signal lines, or both. However, a solder-coated track may significantly slow the
signal, a problem at RFs and microwave frequencies, at which the speed of conductor-
surface signals is highly critical. Solder alloys typically have higher resistivities than
the copper metallization of the board, so even though the solder coating increases the
cross-sectional area, it does not necessarily improve signal speeds for RF designs.
The best design for thick-plate boards is to lay them out so that only the mount-
ing pads are on the top layer and all signal tracks are buried, as required for military
designs. This eliminates the need for any solder masking, but such a layout may
require additional layering.

Solder Reflow The most common reflow methods are vapor-phase and infrared
(IR) reflow. Belt reflow is sometimes used for prototype or small-volume products
but is not practical for production applications.
Both vapor-phase and IR furnace reflow require product-specific equipment pro-
gramming. The thermal mass and configuration of the product determine what program
will provide a temperature profile that gives the product consistent, correct ramp-up,
holds it at the proper peak temperature for the right interval, and cools it to allow
solder-joint formation without thermal overstress. If the product has a large thermal
mass (e.g., several large ceramic chip carriers), it will take more time and energy to heat
the components and their solder joints. Small units may need to enter the vapor furnace
faster, while larger devices may require a slower entry and longer vapor dwell time or
higher furnace temperature.
The first step, therefore, in optimizing a vapor-phase reflow process is to develop
a profile that provides steady temperature ramp-up rapidly enough to avoid volatiliz-
ing reactants before the melting point of the solder and complete reflow are achieved,
but slowly enough for the flux to work and self-alignment to occur. The profile must
also keep the part in the vapor at peak temperature long enough for complete reflow,
but not long enough to allow board degradation, overformation of intermetallics,
bake-on flux or discoloration, or delamination of polyimide or epoxy glass boards.
Some intermetallics must form to allow the metals in the solder to combine and
adhere to the metals on the board surface, but if all the surface metals diffuse into
intermetallics, adhesion is compromised; dwelling at the peak temperature too long

DK565X_003_r04.indd 135 11/11/2005 7:40:34 PM


136 Assembly Process: Finishing, Packaging, and Automation

can result in overformation of intermetallics. The profile also must enable rapid
cooling to provide proper grain growth without thermally overstressing the boards
by slowly lowering the parts into the vapor, letting them sit there for the correct
duration, then pulling them out of the chamber rapidly. After the boards cool to
approximately 90°C, they are immediately sent through a degreaser for cleaning and
defluxing before the flux has a chance to harden onto the board.
Adjusting the vapor-phase profile depends on the component with the largest ther-
mal mass. The vapor hits all surfaces simultaneously, and the heat is transferred from
the vapor to the component. Thus, the temperature at the surface of the component
and its solder joints is independent of surface area or location. The rate of heat transfer
into the solder joints depends on the rate of transfer into the component, which in turn
depends on the thermal mass. Therefore, if the vapor phase is programmed to allow for
complete reflow of the component with the largest thermal mass, it will reflow all other
components.
Even with an optimized printing process and vapor-phase profile, an unforeseen
problem may arise. As the vapor contacts the PWB, it transfers its latent heat and
condenses. The condensed liquid accumulates on the board surface until eventually it
runs off the sides, accumulating before the solder fully melts under large LCCCs that
reflow slowly; this keeps the liquid solder’s surface tension from getting high enough
to hold the LCCCs in place. In other words, the condensed vapor lifts the LCCCs,
washing away the solder paste and floating the LCCCs off the board. To remedy this
problem, the board is placed on a slight angle—about 10°—during the vapor phase;
this angle is just great enough to allow the condensation to run off without permitting
gravity to move any of the LCCCs during reflow.
The solvents used in vapor-phase reflow create additional problems. The Freon
used can be toxic if inhaled in large doses and will react chemically with the organics
(e.g., flux) in the solder paste. The flux and other chemical by-products accumu-
late over time and can alter the vapor temperature or introduce contaminants on the
PWBs; consequently, lot-to-lot process control can be difficult. Solvents, which are
very costly, must be replaced periodically and require special safety precautions,
along with costly waste removal. They also represent a danger to the ozone layer.
Reflow in multiple-zone furnaces, although successful for small assemblies such
as solder-sealing chip carriers or solder-chip mounting or sealing of hybrids, has
not been very successful for solder-mounting components onto PWBs. This is due
not to the limitations of the equipment but to the limitations of the modes of heat
transfer. In a standard multiple-zone furnace, heat is transferred by convection and
conduction and is dependent on surface area. The increase in temperature of a given
component and its board depends on its area, its mass, its specific heat, the area of
the overall assembly, and the component’s location on the assembly. The outer edges
of the unit are the first to heat up. As the heat continues to transfer into the unit,
a temperature gradient will form as the board under the component acts as a heat
sink and conductively pulls heat away from the component. Thus, components at the
center of the board will lag thermally behind those along the edges. In some cases,
the board layout can be designed to compensate for this phenomenon; the larger
components can be placed on the perimeter of the board, with increasingly smaller

DK565X_003_r04.indd 136 11/11/2005 7:40:34 PM


Electronics Assembly 137

devices placed closer to the center. However, this approach typically works only for
smaller PWB assemblies, because significant gradients still form. Furthermore, this
placement is usually not electrically conducive and causes unnecessary layering or
routing difficulties.
Furnace reflow, which uses radiation as the heat-transfer mode, has been made a
viable reflow process. In an IR furnace, in addition to convective and conductive heat
transfer, energy is transferred into the components and their solder joints by expos-
ing all surfaces simultaneously to IR radiation as the primary mode of heat transfer.
There are still thermal gradients, and heat transfer is still dependent on component
area and location, but to a much lesser degree. With proper equipment programming,
most PWB assemblies can be reflowed using an IR furnace. The advantages of IR
furnaces are improved long-term process control and operating costs and elimination
of the use of expensive and dangerous solvents.
In both vapor-phase and IR reflow, once an optimized profile window is estab-
lished for one product, it cannot be used for subsequent products. The program
controlling the descent speed, dwell time, and accent speed (vapor-phase) or the belt
speed, energy levels, or zone settings (furnace), must be adjusted for each product.
A product of higher thermal mass requires lower processing speeds and higher energy
settings to reach the same profile as a device with lower thermal mass.
Whether to use single or double reflow was an issue during the early stages of sur-
face-mount technology process development. Double reflow involves stenciling on the
solder paste, reflowing it, mounting the components, and reflowing the solder again;
both components and boards are thus pretinned prior to assembly. One theory held
that reflowing the solder twice reduced voiding—presumed to weaken solder joints or
to cause cracks to form and propagate—by forcing the gases trapped in the solder to
outgas. While double reflow heightened thermal stressing of the assemblies, increased
intermetallics that embrittled solder joints, and was more expensive, the gains of double
reflow were assumed to cancel out the risks. Still another theory posited that voiding
acted as a relief stress and that crack propagation would stop, not start, at a void.
Comparing the voiding after one reflow and after two, using x-rays and micro-
sectioning, determined that there is no correlation between the number of reflows
and the amount of voiding, nor between voiding and solder-joint failures. A second
reflow has mixed effects on voiding. Solder joints not located under LCCCs can
completely outgas small voids near the surface during the second reflow, which can
thus decrease voiding. However, multiple voids, especially if they are located under
LCCCs, can join together to form large single voids after the second reflow. In both
cases, the number of voids decreases; however, in the latter case, the percent of void-
ing remains the same. The larger voids formed after a second reflow are usually
directly under the LCCCs, where the solder is thinnest and the thermal stress due to
CTE mismatch is greatest. Voiding directly under a self-aligned LCCC, or an LCCC
with solder-mounted thermal pads under the dies, greatly hinders thermal power dis-
sipation from the dies inside the LCCCs through the solder, into the PWB, and finally
to the heat sink or core of the assembly. This thermal solder voiding issue can be
addressed by mounting the body of the LCCC with a thermally conductive adhesive
that then becomes the primary thermal path. It has also been determined that the

DK565X_003_r04.indd 137 11/11/2005 7:40:34 PM


138 Assembly Process: Finishing, Packaging, and Automation

amount and location of voids within solder joints is totally random and is not a func-
tion of the size or location of the LCCCs.
Studies in which both single- and double-reflowed boards were subjected to over
1000 temperature cycles of 55 to 125°C have shown no correlation between
solder joint failures—typically defined as specified percent increases in resistance—
and voiding within the joints. Failure analysis, through microsectioning of the board
assemblies and/or delidding of the LCCCs, showed that the decreases in resistance
were not due to solder joint voiding, cracks, or even wirebond failures; rather, the
stress concentration was located within the plated through-holes. In organic boards
cored to aluminum, copper-clad Invar, and graphite, the CTE mismatch between the
LCCCs and the PWB is overshadowed by the mismatch between the epoxy glass or
polyamide and the core material; consequently, the blind vias and plated through-
holes of the PWB fail before the solder joints. Redesigning cored PWBs to reduce
the thermal stress between the PWB and the heat sink can reduce thermal failures
in through-holes and vias. This can be accomplished by using a core material with a
CTE closer to that of the organic PWB and by laminating the PWB to the core with
a flexible adhesive capable of absorbing the thermal stress. However, this once again
compromises solder joint reliability. The solder responds to the applied strains and
resulting stresses with time-dependent plastic deformation, leading to solder-joint
failures due to accumulating fatigue damage (Engelmaier).
Special Mounting Techniques In production applications, components are typically
mounted with pick-and-place equipment. If solder paste is applied by printing onto the
boards, then placement follows immediately. The tackiness of the paste holds the parts
in place until the solder is reflowed and self-alignment or the surface tension of the
liquid solder takes over, just as inserting the leads in an insertion-mount board aligns
the components and holds them in place until the solder joints are formed. Single-sided
solder-paste PWBs thus do not need additional placement or mounting techniques. For
high heat dissipation and improved conductive cooling, thermally conductive adhesives
are used between the components and the boards. Solder paste can mount pretinned or
nonpretinned components; the flux in the paste allows it to wick up and wet the
component castellations or terminations, so pretinning is not absolutely necessary.
Pretinning is still recommended, however, to ensure complete and uniform wetting to
all I/Os. The paste’s initial tackiness and its later liquid surface tension do not overcome
gravity as well as a wave-soldered, inserted, and clenched component. In addition, the
solder paste must be processed within the time limitations of its solvents. Thus, a
solder-paste surface-mounted board does require extra process control and handling.
Parts on thick-solder-plated boards must be in place with an epoxy or a tacking
flux prior to reflow. Epoxy preforms can be manually or automatically placed under
larger chip carrier or hybrid components. Partial curing enables the epoxy to hold the
components in place while reflow forms the solder joints and finishes the curing of
the epoxy. The same technique can be used with manually applied or printed epoxy
pastes. Tacking involves printing a viscous flux over the thick-plated solder pads.
The flux serves two purposes: It provides a tacky surface that holds the components
in place, and it provides the cleaning and reducing agents needed to rid the solder
surfaces of any oxides that may have formed during storage.

DK565X_003_r04.indd 138 11/11/2005 7:40:34 PM


Electronics Assembly 139

Double-sided mounting also requires special processing. The parts mounted on


the bottom of the boards must be held in place both prior to and during reflow. The
tackiness of the solder paste or the printed flux in thick-solder plating may hold the
parts in place, but if they are too heavy, the surface tension of the liquid solder cannot
secure them until the solder joint is formed. Therefore, bottom-mounted components
must be epoxied in place.

3.6.3 Ceramic PWB Assembly


Introduction to Ceramic PWB Assembly
Ceramic PWBs can have cofired, thick-film, or thin-film metallization. All interlayer
connection or routing is accomplished with buried vias or with windows left out of
a printed dielectric layer or punched out of the green tape. All components are sur-
face mounted. Figure 3.33 shows a ceramic surface-mount assembly that may have
through-holes. In a cofired substrate, these are stacked, metallized vias that produce
a finished hole through the substrate similar to an organic PWB plated through-hole.
In a thick-film substrate, a hole is machined—punched out in the green stage, or laser
drilled or sonic milled out of the fired ceramic. This hole is then cleaned and metal-
lized by coating the walls with a thick-film conductor paste. The result is a plated
through-hole that can be used to route a signal from one side of the board to the other,
or to increase the thermal path at a specific location on the substrate.

FIGURE 3.33 Ceramic PWB assembly. (Courtesy of Westinghouse Electronics Systems


Group.)

DK565X_003_r04.indd 139 11/11/2005 7:40:34 PM


140 Assembly Process: Finishing, Packaging, and Automation

A patterned or metallized substrate, the starting point for any ceramic PWB
assembly, can be mounted to a heat sink or left unmounted. A ceramic PWB is a rigid
flat board that maintains its shape. The camber of such boards is typically 0.002 to
0.003 in./in., maintained approximately throughout processing and operation. These
boards will not warp during solder processing as organic boards will. By definition, the
ceramic board’s CTE matches that of the ceramic components mounted to it. However,
although there is no CTE mismatch, thermal stress is still generated by temperature
gradients formed because of the difference in the rate of heating due to the variant thermal
masses of the components and the board. Nonetheless, the need for ceramic chip carriers
(CCCs) to be leaded to allow for stress relief is greatly reduced. Leadless chip carriers
can be mounted to ceramic PWBs with minimal effect on the long-term reliability of
the assembly. Furthermore, the ceramic material from which the boards are fabricated
offers much better thermal conductivity than its organic PWB counterparts. Use of more
advanced ceramic materials, such as AlN, and incorporation of thermal through-holes in
the substrates, can further enhance the thermal performance of a ceramic PWB.
Ceramic PWBs are particularly valuable in military and space applications. They
offer reductions in thermal stress, reductions in volume (they can incorporate LCCCs
in their design), improved thermal management (due to their superior conductivity and
thermal via capability), and improved long-term reliability (due to a CTE matched with
that of their components, the reduction of thermal stress, and their rigidity). However, as
usual, these improved properties typically mean higher costs and longer design times.

Heat Sink Attachment Although ceramic boards are rigid and can stand on their own,
heat sinks are commonly attached for thermal reasons. The heat sink increases the ability
of the assembly to dissipate heat. It can also act as a carrier plate, providing a flanged
support that can be readily bolted or inserted into a chassis. In standard electronic
modules (SEMs), the module housing is the heat sink/carrier plate for ceramic PWBs.
The materials commonly used are Kovar, tungsten–copper alloys, composites such as
graphic fiber plates or silicon–aluminum alloys, and molybdenum. All of these materials
offer thermal conductivities at least as good as those of the ceramics and CTEs closely
similar to those of ceramics. The CTE match, thermal environmental requirements, out-
gassing restrictions, and linear size of the PWB dictate the attachment method.
Ceramic PWBs can be attached to their heat sinks with epoxy or solder. The
attachment methods used for ceramic PWBs are the same as those for ceramic hybrid
substrates. The mismatch in CTE is more critical for a ceramic PWB because it is
larger than a hybrid substrate. A flexible, stress-absorbing adhesive, such as silicon,
is typically used, provided that no bare dies are mounted on the ceramic PWB.
The back of the ceramic PWB is metallized depending on the attachment method.
For example, if the PWB is to be mounted to the heat sink with solder, its back is
metallized with a platinum–palladium–gold thick film to allow proper wetting while
preventing total leaching of the metallization into the solder. If the PWB is a thin-film
or cofired ceramic, the base metal is nickel and gold plated. The gold plating provides
a good wetting surface, while the nickel acts as a barrier metal providing corrosion
protection and preventing total leaching. Ceramic PWBs can also be bonded directly
to a heat sink during the conductive patterning step.

DK565X_003_r04.indd 140 11/11/2005 7:40:35 PM


Electronics Assembly 141

Surface Mounting
As discussed, ceramic PWBs are not insertion mounted; all components are surface
mounted. These components can have leads formed for surface mounting, but this is
not necessary and lessens the area and volume reduction benefits of using a surface-
mount design.
Leadless components can be used with high reliability because their CTEs are
very close to those of the ceramic PWBs. Leadless components include ceramic chip
carriers, ceramic chip capacitors, and ceramic thin-film and/or thick-film resistor
chips. Resistors can easily be incorporated into the layout of the board, making it a
printed circuit board or PCB, not just a PWB.
The most common mounting method for ceramic boards is to print solder paste
on them, mount the components, and reflow the boards in a vapor-phase or IR fur-
nace. Thick-solder plating of the mounting pads can be done to a thin-film or cofired
board; however, solder-paste printing is still commonly used.

3.6.4 Connector Attachment


In commercial applications, connectors need not be mounted to the PWB assembly.
Instead, the PWB assembly or card has I/O pads along one edge, allowing the card
to be inserted or plugged into an end-card connector, typically an integral part of the
backplane or motherboard of the system. Therefore, the PWB itself acts as the con-
nector for the PWB assembly.
In some commercial and most military applications, connectors are usually
mated pairs; half of the pair is attached to the PWB assembly, while the mating half
is mounted to the system backplane or motherboard. The connector can be attached
before, during, or after component mounting, though attachment before mounting is
not commonly done because it can interfere with solder-paste printing and compo-
nent placement by preventing the PWB from lying flat on the printer stage and the
placement equipment. Because assembling a PWB with connector in place requires
special fixturing and handling, the connector is usually attached at the same time or
after component mounting.
Connectors are typically placed along the edge of the PWB. Some can be mounted
and reflowed at the same time as the components, a common practice with thick-
plated organic PWB assemblies. The connector can be snapped onto the edge of the
assembly and aligned with rivets at each end. The assembly is run, with connector and
components in place, through the wave solder (insertion-mount boards) or through
a vapor-phase or IR furnace (surface-mount boards). To decrease the possibility of
dislodging components while snapping on the connector, the connector is attached
after component mounting and solder reflow. The I/O pads to which the connector
pins are attached are pretinned by thick plating them while plating or wave solder-
ing the component mounting pads or by printing solder paste on them and letting
the paste reflow during component mounting and reflow. After the components are
mounted and reflowed and the boards are cleaned, the connector is positioned with the
spring-loaded pins over the pretinned pads. Locally the connector pads are reflowed

DK565X_003_r04.indd 141 11/11/2005 7:40:35 PM


142 Assembly Process: Finishing, Packaging, and Automation

by exposing the area to a hot-air gun (shielding the rest of the board with a metal
plate), or by placing a hot bar over the connector pads.
Another method of connector attachment uses Raychem’s Solder Kwik product,
which works with either pretinned or plain connector pads. After component mounting
and reflow, the connector is positioned and the Solder Kwik preform is placed next to the
pads, aligning them with the capillaries through which the solder will flow. A hot bar is
placed over the reservoir of solder. The solder melts and flows through the capillaries of
the Solder Kwik preform onto the connector pads, where it wets the pads while wicking
up over the pins.
In addition to edge connectors, flex cables are sometimes mounted to the PWB
assembly in microwave or power modules (SEMs). Because the PWB is typically
mounted to the floor (or central web) of the module, it cannot have a connector
snapped over its edge, so a flex cable is surface mounted to the board. The other end
of the flex usually has a plug-in connector inserted into the module wall to provide
interconnection from the module to the system.

3.7 SYSTEM INTEGRATION


Components, such as bare dies and capacitors, are the building blocks of many elec-
tronic assemblies. Subassemblies, including these components, chip carriers, PGAs,
and hybrid assemblies, are the building blocks of PWB assemblies. In turn, PWBs
are the building blocks for final systems. The PWB assemblies can be components
inserted into the motherboard or backplane of the system (see Figure 3.34), or can be

FIGURE 3.34 Insertion-mount motherboard assembly with two connectors, jumper wires,
one ceramic daughterboard assembly, one organic surface-mount assembly, and discrete
insertion-mount components. (Courtesy of Westinghouse Electronics Systems Group.)

DK565X_003_r04.indd 142 11/11/2005 7:40:35 PM


Electronics Assembly 143

mounted inside SEMs, which in turn are plugged into the system chassis. How these
building blocks participate in the final system will dictate their design.
A typical packaging hierarchy of an electronic system is shown in Figure 3.35.
The lowest level of packaging, or the zeroth level, is generally considered to be the
semiconductor chip, although discrete passive devices such as resistors and capacitors
may also be included. The packaging of a chip or a set of chips in a functional and
protective chip carrier is referred to as the first level of packaging. Chip carriers can
range from single-chip (monolithic) carriers to very sophisticated multichip mod-
ules containing hundreds of chips and devices. The second level of packaging is
often referred to as the electrical circuit assembly (ECA). At this level, the individual

FIGURE 3.35 Packaging hierarchy of an electronic system.

DK565X_003_r04.indd 143 11/11/2005 7:40:35 PM


144 Assembly Process: Finishing, Packaging, and Automation

chip carriers are mounted on a common base, usually a printed wiring board. The
third level of packaging typically involves the interconnection of circuit boards and
power supplies to a physical interface, such as a chassis, control, or electromechani-
cal device or system. The third level of packaging may also involve the connection of
several boards within a supporting or protective structure such as a cabinet. Several
such cabinets are joined together to form the fourth level of packaging.

REFERENCES
Avallone, E. A., and Baumeister, T., III, Marks’ Standard Handbook for Mechanical Engineers,
9th ed., McGraw-Hill, New York, 1987, p. 5.54.
Capp, M., and Luther, R., Analysis of the effect of laser machining on 96% alumina ceramic
substrates and the advantages of new LaTITE finish.
Engelmaier, W., Surface mount solder joint long-term reliability: Design, testing, prediction,
Soldering Surface Mount Technology, 14–22.
Harper, C. A. (Ed.), Handbook of Materials and Processes for Electronics, McGraw-Hill,
New York, 1970.
Harper, C. A., Electronic Packaging and Interconnection Handbook, McGraw-Hill, New York,
1991.
Leahy, K., Microwave Hybrid Design Tutorial, presented at Sect. 1–3, Capital Chapter of
International Society for Hybrid Microelectronics, Twentieth Annual Symposium,
May 1989.
Manzione, L. T., Plastic Packaging of Microelectronic Devices, Van Nostrand Reinhold,
New York, 1990.
Pecht, M. (Ed.), Handbook of Electronic Package Design, Marcel Dekker, New York, 1991.
Schafft, H. A., Testing and fabrication of wire bonds electrical connections—A comprehen-
sive survey, Natl. Bur. Std. (U.S.) Tech. Note 726, pp. 80, 106–109, 1972.
Stamps, J., Sr., Mechanical Engineer, Structural Analysis Group, Westinghouse Electronics
Systems Group, personal interview, Baltimore, Md., February 27, 1990.
Troyk, P., Encapsulants as packaging for implanted electronics, presented at Capital Chapter of
International Society for Hybrid Microelectronics Symposium, May 1987.

SUGGESTED READINGS
Clark, R. H., Handbook of Printed Circuit Manufacturing, Van Nostrand Reinhold, New
York, 1985.
Coombs, C. F., Printed Circuits Handbook, McGraw-Hill, New York.
Dally, J. W., Electronic Packaging—A Mechanical Engineering Perspective, New York, 1989.
Dostal, C. A. (Sr. Ed.), Electronic Materials Handbook: Packaging, Vol. 1, ASM International,
Materials Park, Ohio, 1989.
Hunadi, R. et al., New ultra-high purity, electrically conductive epoxy die attach adhesive for
advanced microelectronic applications, Proceedings of the 1985 International Sympo-
sium on Microelectronics, Anaheim, Calif., November 1985.
Johnson, R. R., Multichip modules: Next-generation packages, IEEE Spectrum, 27, 1990.
Jowett, C. E., Reliable Electronic Assembly Production, Tab Books, Blue Ridge Summit,
Penn., 1971.
Kear, F. W., Printed Circuit Assembly Manufacturing, Marcel Dekker, New York, 1987.

DK565X_003_r04.indd 144 11/11/2005 7:40:39 PM


Electronics Assembly 145

Oscilowski, A., and Sorrells, D. L., Use of thermogravimetric analysis (TGA) in predicting
outgassing characteristics of electrically conductive adhesives, Proceedings of the
Technical Conference—IEPS, Fourth Annual International Electronics Packaging
Society, Baltimore, Md., October 1984.
Pandiri, S. M., Behavior of silver flakes in conductive epoxy adhesives, Adhesives Age, 30,
31–35, 1987.
Ruska, W. S., Microelectronic Processing—An Introduction to the Manufacture of Integrated
Circuits, McGraw-Hill, New York, 1987.
Sorrells, D. L. et al., Selection and cure optimization of conductive adhesives for use in AuSn
sealed microelectronic packages, Proceedings of the 1984 International Symposium
on Microelectronics, Dallas, Tex., September 1984.
Tummala, R. R., and Rymaszewski, E. J. (Eds.), Microelectronics Packaging Handbook, Van
Nostrand Reinhold, New York, 1989.

DK565X_003_r04.indd 145 11/11/2005 7:40:40 PM


DK565X_003_r04.indd 146 11/11/2005 7:40:40 PM
Index

A inspection, 42–43
justification, 44–47
Adhesives machine/operator relationships, 44, 54–57
component mounting, 100 material handling, 57–72
electronic package customization, 122 motion systems, 38–44, 39–42
organic PWB assemblies, 99 production design, 49–57, 50–53
printed wiring board assemblies, 126 robots, 37–38
Advantages and disadvantages, see also software interfaces, 48
Benefits standard machine bases, 33–37
automation, material handling, 58 storage equipment, 60–64
ceramic packaging, 120–121 testing, 42–43
continuous-motion machines, 36 transportation equipment, 66–72
dial indexing machines, 34 Assembly methods, 4
die assemblies, 74 Assembly process, 4, 8–9
floating work platform machines, 35 Automated assembly machines
hybrid microelectronics assemblies, 97 continuous-motion machines, 36–37, 37
in-line machines, 34 dial indexing machines, 34, 35
surface mounting, 129 floating work platform machines, 35–36
tape automated bonding, 104, 106 in-line machine, 34–35, 36
Aerospace industry, 31–32 robots, 37–38
Alignment, 54 standard machine bases, 33–37
Aluminum, 124 Automated factory, 32
AMRF, see Automated Manufacturing Automated guide vehicle system (AVGS),
Research Facilities (AMRF) 69, 70
Applications, see also specific type of Automated Manufacturing Research
application Facilities (AMRF), 48
die assemblies, 84–86 Automated retrieval systems, see
high-voltage, 89 Automated storage and retrieval systems
Approaches, material handling, 58–60 (AS/RS)
Arcing potentials, 89 Automated storage and retrieval systems
Arms, motion economy, 20–22 (AS/RS), 61, 61–63, see also Storage
AS/RS, see Automated storage and retrieval systems
systems (AS/RS) Automation concepts, 32–33
Assembly automation Avallone and Baumeister studies,
advantages and disadvantages, 58 110–111, 115
automation concepts, 32–33 AVGS, see Automated guide vehicle system
basics, 31–32 (AVGS)
benefits, 45–47, 47 Axes of motion, 38
factory, assembly machines, 32 Axis of symmetry, 54

147

DK565X_004_r04.indd 147 11/11/2005 5:39:12 PM


148 Index

B Cerpacks, 131, see also Ceramic


packaging
Ball bond, 85, 102–103 Chairs, 25
Balling and bridging, 133–134 Chip carrier assemblies
Ballistic motions, 22–23 basics, 91
Ball-limiting metallurgy (BLM), 106 ceramic chip carriers, 92–94, 93–95
Baumeister, Avallone and, studies, injection molding process, 92
110–111, 115 pin-grid array packaging, 94–95
Benefits, see also Advantages and plastic chip carriers, 91–92
disadvantages reaction injection molding process, 92
assembly automation, 45–47, 47 transfer molding process, 91
automated storage and retrieval CIM, see Computer-integrated
systems, 62 manufacturing (CIM)
power and free conveyors, 71 Cincinnati Milicron T3 material, 41
Bhagath studies, 73–144 Climbing/clinging parts, 51, 53
Biomedical hybrids, 109–110 CNC, see Computer numerical control
BLM, see Ball-limiting metallurgy (BLM) (CNC) system
Bloom, H., 48 Coast Guard radio receiver example, 6, 6
Board components, 133 Coefficients of thermal expansion (CTE)
Body positions, changes, 26 ceramic PWB assemblies, 140
Body strength, automation, 33 cored boards, 132–133
Boothroyd studies, 3 die assembly materials, 78
Bottlenecks, 15 electronic package customization, 122
Bowl driving forces, 33, 49 heat sink attachment, 124
Broken parts, 33 printed wiring board assemblies,
Bugs, see Debugging 122, 126
solder reflow, 138
surface-mounted LCCCs, 132
C Commercial applications
connector attachment, 141
Capacitance calculations, capacitor die assemblies, 85–86
banks, 87 heat sink attachment, 123
Capacitor banks, 86–88 thermal stress, biomedical hybrids, 110
Capec studies, 38 Component mounting, 99–101, see also
Capp and Luther studies, 119 Mounting
Carousel storage and retrieval system Component placement, 126, 127
(CS/RS), 63–64, 64–67 Computer-aided process planning
Cash flow, 46, 47, see also Costs; Financial (CAPP), 7
benefits Computer hardware requirements, 62
Cavity-up/cavity-down packages, 93 Computer-integrated manufacturing
Ceramic assemblies (CIM), 48
basics, 139, 139–140 Computer numerical control (CNC) system,
heat sink attachment, 140 39, 59
surface mounting, 141 Conductivity, component mounting,
Ceramic chip carriers, 92–94, 93–95 99–100
Ceramic feedthroughs, 117, 118–120 Connector attachment, PWBs, 141–142
Ceramic packaging, 74, 120–121, 121 Contingency plans, 62
Ceramic PWB assemblies, 139–141 Continuous-motion machines, 36–37, 37
Ceramic tabs, die assembly materials, 78 Controlled movements, 22

DK565X_004_r04.indd 148 11/11/2005 5:39:14 PM


Index 149

Controls, motion economy, 23 E


Conveyors, material handling, 69–72,
71–72 ECA, see Electrical circuit assembly (ECA)
Cored boards, 132–133 Efficiency, 33, 49
Costs, see also Financial benefits; Electrical circuit assembly (ECA), 144
Justification Electromagnetic interference (EMI), 89, 97
decisions, 1 Electronic packaging, 113–114, 121–122
justification, saving vs. making money, 31 Electronics assembly
work simplification, 15 basics, 73
Critical tolerances, 51 capacitor banks, 86–88
Cross-talk, 89 ceramic chip carrier package, 92–94,
CS/RS, see Carousel storage and retrieval 93–95
system (CS/RS) ceramic PWB assemblies, 139–141
chip carrier assemblies, 91–95
connector attachment, 141–142
D die assemblies, 74, 77, 78–86
elementary subassemblies, 74–89
Debugging, 56 hybrid microelectronics assemblies,
Decision process, 52–54 95–122
Default settings, 56 microwave frequency subassemblies,
Delay, flowchart symbols, 17 88–89, 90
Design organic PWB assemblies, 123–139
hybrid microelectronics assemblies, package architecture, 74, 76–77
96, 96–99 packaging, hybrid, 115–122
parts, 54 pin-grid array packaging, 94–95
production, 49–57, 50–53 plastic chip carrier package, 91–92
Dial indexing machines, 34, 35 printed wiring board assemblies,
Die assemblies 122–142
applications, 84–86 radio-frequency subassemblies,
basics, 74, 77, 78–86 88–89, 90
interconnection, 80, 83–84, 83–86 subassemblies, 74–89
materials, 78 system breakdown, 73–74, 75
processing, 78–80, 79, 81–83 system integration, 142–143, 142–144
Dimensioning, parts, 54 Electrostatic discharge (ESD), 97–98
DIPs, see Dual in-line packages (DIPs) Elementary subassemblies
Direct-bond copper, 122 applications, 84–86
Direction, changes in, 22 basics, 89
Distortion, 33, 49 capacitor banks, 86–88
“Do,” work simplification, 15–16 die assemblies, 74, 77, 78–86
Documentation, 7, 28, see also Work interconnection, 80, 83–84, 83–86
instructions microwave frequency subassemblies,
Double reflow, 137 88–89, 90
Double-sided mounting, 139 radio-frequency subassemblies,
Dragon rocket components, 31–33, 57 88–89, 90
Drawbacks, AS/RS, 62 EMI, see Electromagnetic interference
Drawings, responsibility, 7 (EMI)
Drop deliveries, 23 Encapsulated hybrid assembly, 96, 96
Dual in-line packages (DIPs), 91, Engelmaier studies, 138
126, 131 Engineering drawings, responsibility, 7

DK565X_004_r04.indd 149 11/11/2005 5:39:14 PM


150 Index

Engineering tree chart, 5 Glove box, substrate mounting, 108


Epoxy mounting, interconnection, 107 Gold, glass chip-outs, 118
Equipment Gold paste, 108
motion economy, 25–26 Gold wirebonds, 80, 83, 103
storage, 60–64 Graphite fiber composites, 124
transportation, 66–72 Gravity-feed bins and containers, 23
ESD, see Electrostatic discharge (ESD) Guide vehicle system, automated, 69, 70
Eutectic mounting, 79, 79–80 Gull-winged chip carriers, 95, 97
Extended-work-area robots, 38
Eye fixations, 23
H

F Hand motion classes, 22–25


Hands, motion economy, 20–22, 25
Factory, assembly machines, 32 Hardware requirements, 62
Factory Automation Systems Division, 48 Harper studies, 92, 102
Fatigue, 15, 44 Harris studies, 73–144
Feeder, vibrating action, 49 Heat sink attachment, 123–124, 125, 140
Feedthroughs, 117–118, 118–120 Height, workplace, 25
Field effect transistors (FET) High-density storage, 63
die assemblies, 74, 78, 85–86 High-voltage applications, 89
interconnection, 101–102 Horizontal carousels, 63, 64, 67
Financial benefits, 44–47, see also Costs; Hybrid microelectronics assemblies
Justification basics, 95
Finger motions, 22, 26 component mounting, 99–101, 102
Five-axis motion system, 38, 40 design, 96, 96–99
Flip-chip bonding, 106, 106 interconnection, 101–104, 104–110
Floating work platform machines, 35–36 metal packaging, 113–114, 116–122,
Flow diagram, 17, 19 117–118, 121
Flow process charts, 16–17, 18, 21 package sealing, 110–111, 111–114, 115
Fluxless solder reflow, 80 packaging, hybrid, 115–122
Ford, Henry, 3 processing, 99–115
Forearm, motion economy, 22 sealing packages, 110–111, 111–114,
Four-axis motion system, 38, 40 115
Free conveyors, see Power and free Hybrid packaging, 115–122
conveyors
Frenchtown ceramic-metal brazed
assemblies, 113–114 I
Furnace mounting, 80
Furnaces, reflow, 136–137 IBM corporation
software interfaces, 48
testing products, 43
G total system installation, 60
ICs, see Integrated circuits (ICs)
GE, see General Electric (GE) IMAs, see Integrated microwave assemblies
General Electric (GE), 46 (IMAs)
General Motors (GM), 60 IMDAS, see Integrated Manufacturing
Glass seals, metal packaging, 116–118, Database Administration System
117–118 (IMDAS)

DK565X_004_r04.indd 150 11/11/2005 5:39:14 PM


Index 151

Incompletely molded parts, 33 fabrication and design, 93–94, 94


Indented parts list, 5 surface mounting, 128, 131
Infrared (IR) reflow and furnaces, 135, Leahy studies, 120
137, 141 Lid deflection, 110–111, 111–114, 115
Injection molding process, 92 Line layouts, 11, 14
In-line indexing machine, 34–35, 36 Load-testing, 62
Insertion mounting, 126–127, 127–130 Locomotive Products Division (GE), 46
Inspection, 17, 42–43 LSI, see Large-scale integrated (LSI)
Installation, 59–60, 63 circuits
Integrated circuits (ICs), 97 Luther, Capp and, studies, 119
Integrated Manufacturing Database
Administration System (IMDAS), 49
Integrated microwave assemblies M
(IMAs), 95
Interconnection Machine/operator relationships, 44, 54–57
die assemblies, 80, 83–84, 83–86 Magnetism sensitivity, 33, 49
hybrid processing, 101–104 “Make ready,” work simplification, 15–16
Interfaces, 48, 97 Making vs. saving money, 31
Interpolation, simultaneous, 39 Manual assembly
Islands of automation, 59, 60 basics, 1–3, 2
Item placement planning, 62 equipment, 25–26
hand motion classes, 22–25
line layouts, 11, 14
J methods analysis, 12, 15–20
motion economy, 20–26
J-leaded chip carriers, 91, 97 operation sequences, 5–6, 5–11
Joint defects, solder, 133, 133 operator tasks, 20–22
Justification, see also Costs; Financial repairs, standard, 27–28
benefits routings, 7
automation, 3, 44–47 special instructions, 28
documentation, 5 standard processes, 26–28
injection molding cost, 92 tools, 25–26
visual aids, 10–11, 10–13
work instructions, 4–5, 7–99, 8–9
L workmanship standards, 26, 27
workplace environment, 23–25, 24–25
LAG, see Laser artwork generator (LAG) work simplification, 15–17, 18–19,
Large-scale integrated (LSI) circuits, 131 19–20, 21
Laser artwork generator (LAG), 98 workstation layouts, 11, 13–14
LCCC, see Leadless ceramic chip carriers Manzione studies, 91
(LCCCs) Material handling
Lead forming, 126, 128 advantages and disadvantages,
Leadless ceramic chip carriers (LCCCs) automation, 58
ceramic PWB assemblies, 140 approaches, 58–60
heat sink attachment, 124 automated storage systems, 61, 61–63
organic PWB assemblies, 132 basics, 57
soldering, 133–134, 136–138 carousel storage and retrieval system,
Leadless chip carriers (LCCs) 63–64, 64–67
ceramic PWB assemblies, 140 characteristics, 57–58

DK565X_004_r04.indd 151 11/11/2005 5:39:14 PM


152 Index

conveyors, 69–72, 71–72 Mounting, 78–80, 79, 81–83, see also


free conveyors, 69–72, 71–72 Component mounting; Surface
guide vehicle system, automated, mounting
69, 70 Multichip packages (MCPs), 95
islands of automation, 59, 60 Multicircuit hybrid packages (MHPs), 95
motion economy, 23 Multilayered ceramic chip carriers,
power and free conveyors, 69–72, 92–93, 93
71–72 Multiple zone furnaces, reflow, 136
retrieval systems, 61, 61–63
storage equipment and systems,
60–64 N
total system installation, 59–60
transportation equipment, 66–72 National Association of Electrical
transporters, 66–69, 68 Manufacturers (NEMA), 44
Materials, 54, 78 National Bureau of Standards (NBS), 48
McDonnell Douglas company National Institute of Standards and
assembly automation, 32 Technology (NIST), 41
automation justification, 45 NBS, see National Bureau of Standards
start-up problems, 57 (NBS)
testing products, 43 NEMA, see National Association of
MCPs, see Multichip packages (MCPs) Electrical Manufacturers (NEMA)
Mechanical mounting, 107 Nested parts, 51, 53
Messy materials, 54 NIST, see National Institute of Standards
Metal packaging, 74, 116–122 and Technology (NIST)
Methods analysis, manual assembly, Nonproductive time, 16
12, 15–20
MHPs, see Multicircuit hybrid packages
(MHPs) O
Microwave frequency subassemblies,
88–89, 90 One-person factory analogy, 1–3
MIL-H-38534 (standard), 119 Operations
Military applications flowchart symbols, 17
connector attachment, 141 instruction procedures, 26
die assemblies, 84 parts list, 10
heat sink attachment, 123 sequences, 5–6, 5–11
hybrid microelectronics assemblies, 99 Operator tasks, motion economy, 20–22
interconnections, 83 Organic PWB assemblies
package sealing, 110 adhesives, 99
printed wiring board assemblies, 126 basics, 123–124, 126–127
thermal stress, biomedical hybrids, 110 board components, 133
Moisture sensitivity, 33, 49 cored boards, 132–133
Momentum, motion economy, 22 deposition, 133, 133–135
Motion economy heat sink attachment, 123–124, 125
basics, 20 hybrid microelectronics assemblies, 97
hand motion classes, 22–23 insertion mounting, 126–127, 127–130
operator tasks, 20–22 interconnection, 133
tools and equipment, 25–26 leadless ceramic chip carriers, 132
workplace environment, 23–25 reflow, 135–138
Motion systems, 38–44, 39–42 soldering, 133, 133–138

DK565X_004_r04.indd 152 11/11/2005 5:39:14 PM


Index 153

special mounting techniques, 138–139 Prepositioning, tools and equipment, 25


surface mounting, 128–129, 130, Printed wiring board (PWB) assemblies
131–139, 132 basics, 122
Orientation, automated production design, ceramic assemblies, 139–141
49–50, 52, 53 connector attachment, 141–142
Overanalysis, 56 organic assemblies, 123–139
surface mounting, 128–129,
131–139, 141
P system integration, 142–143, 142–144
Processes, 4–7
Packaging Processing
architecture, electronics assembly, 74, die assemblies, 78–80, 79, 81–83
76–77 hybrid microelectronics assemblies,
hierarchy, electronic system, 143, 99–115
143–144 Production design, 49–57, 50–53
hybrid microelectronics assemblies, Progressive automation technique, 59
115–122 Psychological problems, 55–57
materials, 76 “Put away,” work simplification, 15–16
sealing, 110–111, 115 PWB, see Printed wiring board (PWB)
Part geometry, 33, 49 assemblies
Particle impact noise detection (PIND), 115
Parting line flash, 33, 49
Parts Q
assembly, easy, 51, 53
balancing, 54 Question-and-answer approach, 19–20
broken, 33
incompletely molded, 33
list, 9, 10, 62, 75 R
nested, shingle/climbing, clinging, 51, 53
problems, 53–54 Radio-frequency subassemblies,
replenishing supply, 55 88–89, 90
tangled, 50–51, 52 Raychem’s Solder Kwik product, 142
Payback/payoffs, financial, 44–47, see also Reaction injection molding process, 92
Costs; Justification Repairs, 27–28, 118
Pecht studies, 73–144 Residual magnetism sensitivity, 33, 49
Performance needs, AS/RS, 62 Restricted movements, motion
Personnel, number assigned, 15 economy, 22
PGA, see Pin-grid array (PGA) packaging Retrieval systems, material handling, 61,
PHPs, see Power hybrid packages (PHPs) 61–63
PIND, see Particle impact noise detection Revisions, responsibility, 7
(PIND) Rhythm, motion economy, 23
Pin-grid array (PGA) packaging, 94–95 Ribbon bonding, 89, 103–104
Planning, AS/RS, 62 Ripple effects, automation, 45, 47
Plastic chip carrier packages, 91–92 RMA, see Rosin mildly active (RMA)
Plastic packaging, 74 Robots, 38–44, 39–42
Polar properties, automated production Rosin mildly active (RMA), 133
design, 49 “Rossum’s Universal Robots,” 36
Power and free conveyors, 69–72, 71–72 Rotary indexing machine, 34, 35
Power hybrid packages (PHPs), 95 Routings, operation sequences, 7

DK565X_004_r04.indd 153 11/11/2005 5:39:15 PM


154 Index

S Statistics, manufacturing, 1, 2
Stitch bond, 102
Saturn (GM), 60 Storage systems, see also Automated
Saving vs. making money, 31 storage and retrieval systems (AS/RS)
Schafft studies, 103 automated, 61, 61–63
Schedules, work simplification, 15 component mounting, 99
Sealing packages, 110–111, 115 equipment, 60–64
Self-alignment, solder deposition, 134 flowchart symbols, 17
Self-inspection, 56 Subassemblies, 74–89
SEM, see Standard electronic Substrate mounting, interconnection, 107
module (SEM) Surface-mounted LCCCs, 132
Semiautomated operations, parts, 54 Surface mounting, see also Component
Shaped region, parts, 54 mounting; Mounting
Sheth, 31–72 ceramic PWB assemblies, 141
Shingle/climbing parts, 51, 53 organic PWB assemblies, 128–129, 130,
Shut-down instructions, 26 131–139, 132
Silver-filled epoxies, 99 Symmetry, axis, 54
Simultaneous interpolation, 39 System breakdown, electronics assembly,
Single-chip packages, 74 73–74, 75
Single reflow, 137 System installation, total, 59–60
Software, 48, 62 System integration, 142–143, 142–144
Soldering
alloys, 81–82
component mounting, 100–101, 102 T
deposition, 133, 133–135
interconnection, 133 TAB, see Tape automated bonding (TAB)
mounting, 107–109, 108 Tabs, die assembly materials, 78
paste, excessive, 134 Tangled parts, 50–51, 52
reflow, 135–138 Tanner studies, 1–28
Solder Kwik product, 142 Tape automated bonding (TAB), 104–106,
Sorting parts, 54 104–106
Space applications, 83–84 Taping iron-core transformers example,
Special instructions, manual assembly, 28 26, 27
Special mounting techniques, 138–139 Temporary mask, 134–135
Springs, 50–51 Testing, 42–43
Stamps studies, 116, 119 Thermal bonding methods, 83
Standard electronic module (SEM) Thermal interfaces, 97
ceramic PWB assemblies, 140 Thermal paths, 124, 125
connector attachment, 142 Thermal stress, biomedical hybrid, 110
heat sink attachment, 124 Thin-film polyamide boards, 122
printed wiring board assemblies, 122, Three-axis motion system, 38, 39
124, 126 Throughput, 62
system integration, 143 Tolerances, critical, 51
Standard machine bases, 33–37 Tools, motion economy, 23, 25–26
Standard processes, 26–28 Total system installation, 59–60
Standard repairs, 27–28 Transfer molding process, 91
Standoff posts, 115 Transportation, 17, 66–72
Start-up process, 26, 63 Transporters, material handling,
Static electricity sensitivity, 33, 49 66–69, 68

DK565X_004_r04.indd 154 11/11/2005 5:39:15 PM


Index 155

Troyk studies, 109 Wedge bond, 85, 102


Two-axis motion system, 38, 39 Welding, package sealing, 110, 111
Wirebonds
fusing currents, 83
U gold wirebonds, 80, 83, 103
interconnection, 101–104
Ultrasonic/thermocompression pull test, 86
wirebonding, 103 Wiring harness example, 8, 9
Ultrasonic wirebonding, 103 Work conditions and practices, 15
Underanalysis, 56 Work instructions, 4–5, 7–10, 8–9
Unsatisfactory work conditions and Workload, work simplification, 15
practices, 15 Workmanship standards, 26, 27
Workplace environment, 23–25,
24–25
V Work simplification, 15–17, 18–19,
19–20, 21
Vapor-phase reflow, 102, 135, 137, 141 Workstation layouts, 11, 13–14
Vertical carousels, 64, 66, 71
Vibrating action, feeder, 49
Visual aids, 4, 10–11, 10–13 X

X-Y-Z rotary model, 38


W

Walker studies, 1–28, 31–72 Y


Walking, work simplification, 15
Waste, excessive, 15 Young modulus, 110

DK565X_004_r04.indd 155 11/11/2005 5:39:15 PM


DK565X_004_r04.indd 156 11/11/2005 5:39:15 PM