CRC Electronic Components and Technology User Manual

Download

Electronic Components and Technology

THIRD EDITION

TUTORIAL GUIDES IN ELECTRONIC ENGINEERING

Series editors

Professor G. G. Bloodworth, University of York Professor A. P. Dorey, University of Lancaster Professor J. K. Fidler, University of Northumbria

This series is aimed at rst- and second-year undergraduate courses. Each text is complete in itself, although linked with others in the series. Where possible, the trend toward a “systems” approach is acknowledged, but classical fundamental areas of study have not been excluded. Worked examples feature prominently and indicate, where appropriate, a number of approaches to the same problem.

A format providing marginal notes has been adopted to allow the authors to include ideas and material to support the main text. These notes include references to standard mainstream texts and commentary on the applicability of solution methods, aimed particularly at covering points normally found difcult. Graded problems are provided at the end of each chapter, with answers at the end of the book.

Electronic Components and Technology

THIR D ED ITIO N

Stephen Sangwine

CRC Press is an imprint of the Taylor & Francis Group, an informa business

Boca Raton London New York

CRC Press Taylor & Francis Group 6000 Broken Sound Parkway NW, Suite 300 Boca Raton, FL 33487-2742

No claim to original U.S. Government works Version Date: 20140113

International Standard Book Number-13: 978-1-4200-0768-8 (eBook - PDF)

This book contains information obtained from authentic and highly regarded sources. Reasonable efforts have been made to publish reliable data and information, but the author and publisher cannot assume responsibility for the validity of all materials or the consequences of their use. The authors and publishers have attempted to trace the copyright holders of all material reproduced in this publication and apologize to copyright holders if permission to publish in this form has not been obtained. If any copyright material has not been acknowledged please write and let us know so we may rectify in any future reprint.

Except as permitted under U.S. Copyright Law, 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.

Visit the Taylor & Francis Web site at http://www.taylorandfrancis.com

and the CRC Press Web site at http://www.crcpress.com

Preface to the Second Edition vii

Preface

Acknowledgments

Author

1 Introduction 1

2 Interconnection technology 5

Jointing Discrete wiring Cables Connectors Printed circuits Printed circuit assembly Rework and repair Case study: A temperature controller Summary 29 Problems 30

3 Integrated circuits 31

Review of semiconductor theory Integrated-circuit fabrication Semiconductor packaging Handling of semiconductor devices Custom integrated circuits Summary 50 Problems 51

xiii

11 12 15

18 24 25 26

33 35 43 45 46

4 Power sources and power supplies 53

Energy sources Batteries Power supplies Summary 74 Problems 74

5 Passive electronic components 77

Passive component characteristics Resistors Capacitors Inductors Summary 92 Problems 93

54 55 60

77 83 87

6 Instruments and measurement 95

Quantities to be measured Voltage and current measurement Frequency and time measurement Waveforms — The oscilloscope Summary 109 Problems 110

7 Heat management 111

Heat transfer Thermal resistance Heat sinking Forced cooling Advanced heat-removal techniques Summary 121 Problems 122

8 Parasitic electrical and electromagnetic effects 123

Parasitic circuit elements Distributed-parameter circuits Electromagnetic interference Applications studies Summary 149 Problems 150

96 98

101

104

112 113 114 119

120

123 128 133 142

9 Reliability and maintainability 153

Failure The “bathtub” curve Measures of reliability and maintainability High-reliability systems Maintenance Summary 168 Problems 168

10 Environmental factors and testing 171

Environmental factors Type testing Electronic production testing Summary 184 Problems 185

11 Safety 187

Electric shock Other safety hazards Design for safety Summary 198

References 201

Answers to problems

154 155 156 162 165

171 178

180

188 194 195

203

Index

205

Preface to the Second Edition

This book is intended to support Engineering Applications studies in electronic engineering and related subjects such as computer engineering and communications engineering at rst- and second-year undergraduate level. Engineering Applications, abbreviated as EA, is a term rst used in the report of the Finniston inquiry into the future of engineering in the United Kingdom. Finniston used the terms EA1 and EA2 to refer to the rst and second elements of a four-stage training in Engineering Applications, to be taken as part of a rst-degree course in engineering. Later, the Engineering Council, established as a result of the Finniston report, expressed the concept of EA1 as

An introduction to good engineering practice and the properties, behaviour, fabrication and use of relevant materials, systems and components.

and EA2 as

Application of scientic and engineering principles to the solution of practical problems of engineering systems and processes.

Although EA studies should be integrated into the fabric of a degree course, there is a need to draw out elements of practice to provide emphasis. It is intended that this book should be used as a source, complementing other texts, for such studies.

In the context of electronics, product design is an activity that begins,

by and large, with components rather than materials. This is not to say that a study of materials is not relevant as a part of EA1, but as there are many existing texts covering the subject, materials has been excluded from this book in favour of more coverage of components.

The book begins with an introduction to electronic interconnection technology including wiring, connectors, soldering and other jointing techniques, and printed circuits. Chapter 3 is devoted to the very important technology of integrated circuits, concentrating on their fabrication, packaging, and handling. Components is taken to include power supplies, as in many applications a power supply unit is bought-in as a subsystem. The main characteristics of power supplies and batteries are covered in Chapter 4. Passive electronic components are introduced in Chapter 5, and with them the book begins to include a major theme developed in Chapters 7 and 8: the parasitic effect. This includes the nonideal properties of passive components introduced in Chapter 5, heat and its management in Chapter 7, and parasitic electromagnetic effects in Chapter 8. EA1 is essentially practically oriented and will include

These quotations are taken

These quotations are taken from Standards and routes to

from Standards and routes to

registration (second edition),

otherwise known as SARTOR,

otherwise known as SARTOR, published by the Engineering

published by the Engineering Council in January 1990. They

Council in January 1990. They are reproduced here with the

are reproduced here with the permission of the Engineering

permission of the Engineering Council, United Kingdom.

Council, United Kingdom.

vii

laboratory-based work, including the use of tools and instruments. A new chapter has been added to the second edition to add to the utility of the book in supporting EA1 studies and laboratory activities. Thus, Chapter 6 introduces the instruments and measurements used in electronics and related subjects. Chapter 9 reviews good engineering practice in relation to reliability and maintainability, two important aspects of design which, unfortunately, are often overlooked by electronic circuit designers. Chapter 10 introduces environmental inuences on electronic products and the subject of testing both for environmental endurance and in production. The nal chapter in the book introduces safety.

The book assumes that the reader has taken the rst one or two terms of a degree course, although some of the earlier material could be studied sooner. Extensive cross-references to more specialized texts have been given in the marginal notes and the bibliography, including, where appropriate, references to other texts in the Tutorial Guides Series. These have been updated for the second edition to include later books in the series where appropriate, the latest editions of technical standards, new editions of books previously listed, and some additional books published since 1986. Other major revisions in the second edition include updated information in Chapter 3 to take into account changes in IC technology since 1986, changes to the nal chapter to take into account new legislation, and some new illustrations.

Part of the aim of this book is to inform the reader about components, technology, and applications, but it is also intended to create an awareness of the problems of electronic engineering in practice. I hope that the readers of this book will be encouraged to tackle these problems and go on to become competent and professional electronics engineers.

Safety Note

The material in Chapter 11 is of course at an introductory level only, and readers are cautioned that professional competence in safe electrical design cannot be achieved merely by studying the contents of this chapter.

viii

Preface

Since the publication of the second edition of this book in 1994, some very signicant changes have occurred in technology, particularly the further miniaturization of electronic products, and the steady and quite dramatic increases in the speed of computers. In electronics, surfacemount technology has become almost universal and vastly better batteries have been developed for laptop computers and portable phones. Technically, however, electronics has not changed in revolutionary ways. When revising the book for this third edition, it was a surprise to discover that components and technologies featured in the second edition were still commercially available. Nevertheless, the revisions needed after 12 years were extensive, but they did not require the text to be restructured. Examples of the areas that needed updating are: the introduction of lead-free solders and digital oscilloscopes, and new types of batteries. The bibliography has been brought up to date, and all references to technical standards, European Union directives, and the like have been checked and, where necessary, updated.

The book has now been in print for 19 years. The previous editions were published in the United Kingdom and largely written for a British audience. In revising the book for this third edition, the opportunity has been taken to make the book more usable elsewhere in the Englishspeaking world, by small changes in terminology and vocabulary, and by reference to international standards, rather than British Standards, where applicable.

This new edition was prepared electronically, which should make it easier to update at reprinting if the publishers wish to do so. Therefore, please contact me with any corrections or suggested amendments at S.Sangwine@IEEE.org.

Stephen J. Sangwine

Colchester, United Kingdom

Acknowledgments

The rst edition of this book developed from a lecture course that I rst presented in 1985 as a part of new EA material introduced into engineering courses at the University of Reading. I would like to thank my former colleague Peter Atkinson for his early suggestion that a book could be written and for his help and encouragement while I wrote the book and subsequently. I would also like to thank S. C. Dunn, former chief scientist at British Aerospace, who made many helpful suggestions at an early stage, including the theme of parasitic effects. The following people contributed advice, criticism, or technical background during the preparation of the rst edition, and I wish to acknowledge their help: Susan Partridge of General Electric Company (GEC) Hurst Research Centre for reading the rst draft of Chapter 3; Alistair Sharp of Eurotherm Ltd. for help with the case study in Chapter 2; John Barron of Tectonic Products, Wokingham, for help with the subject of printed circuits; John Terry of the Health and Safety Executive and Ken Clark, deputy director at Baseefa, for both help and criticism of Chapter 10 (now Chapter 11); Dr. George Bandurek at Mars Electronics at Winnersh for commenting on Chapter 8 (now Chapter 9); and Martin Thurlow of the Electromagnetic Engineering Group, British Aerospace, and David Hunter of the Army Weapons Division, British Aerospace, for assisting with illustrations and background. I am also grateful to Carole Hankins for typing the manuscript and to my wife, Elizabeth Shirley, for her support over many months of writing and revising. Professor A. P. Dorey was consultant editor for this book and made many helpful comments on drafts of the manuscript.

During revision for the second edition, the following people contributed advice and I wish to acknowledge their help: Trevor Clarkson of King’s College for suggesting the addition of measurement; John Whitehouse for suggestions in Chapter 8; and John Terry (again) for reading the former Chapter 10 and suggesting numerous updates, which I have incorporated into Chapter 11. I would also like to thank Charles Preston for helping with the illustrations, particularly for Chapter 6. Professor A. P. Dorey was, once again, consultant editor and made helpful comments on drafts of the new material.

During revision for the third edition, the following people assisted with illustrations and I wish to acknowledge their help: Mark Johnson of Megger Limited; Dr. Ursula Kattner of National Institute of Standards and Technology (NIST) for help with lead-free solders; Stephen Head and Dave Bremner of Eurotherm Ltd.; Marcus Brain of Technisher Überwachungsverein (TÜV) Product Services; Steve Smith of Schaffner Limited; R. J. Sullivan of Aavid Thermalloy; Jeff Weir, Naomi Mitchell, and Cole Reif of National Semiconductor; Paul Bennett of Bulgin Components; Greg Macdonald of Amphenol Canada and Gilles Dupre

of Amphenol-Socapex, France; Natasha Moore of BEAB–ASTA (the British Electrotechnical Approvals Board and Association of Short-Circuit Test Authorities); Yuko Takahashi of Fujitsu Limited; and Gary Silcott of Texas Instruments.

Several companies and organizations have supplied illustrations or granted permission for me to use their copyright material, and they are acknowledged in the text. Permission to reproduce extracts from BS-EN 61340-5-1: 2001 is granted by the British Standards Institution (BSI). British Standards can be obtained from BSI Customer Service, 389 Chiswick High Road, London W4 4AL, United Kingdom; Telephone: +44 (0)20 8996 9001. The author thanks the International Electrotechnical Commission (IEC) for permission to reproduce information from its International Technical Specication IEC 60479-1, Fourth Edition (2005) and from its International Standard IEC 61340-5-1, First Edition (1998). All such extracts are copyright of IEC, Geneva, Switzerland. All rights reserved. Further information on the IEC is available from www. iec.ch. IEC has no responsibility for the placement and context in which the extracts and contents are reproduced by the author, and IEC is not in any way responsible for the other content or accuracy therein.

xii

Author

Stephen J. Sangwine was born in London, United Kingdom in 1956.

He has a B.Sc. degree in electronic engineering from the University of Southampton, United Kingdom (1979), and a Ph.D. degree from the University of Reading, United Kingdom (1991). He is currently a senior lecturer with the Department of Electronic Systems Engineering at the University of Essex, Colchester, United Kingdom. From 1985 to 2000, he was a lecturer with the Department of Engineering at the University of Reading, where he wrote the rst and second editions of Electronic Components and Technology. From 1979 to 1984, he worked in the civilian nuclear power industry at the United Kingdom Atomic Energy Authority’s Harwell Laboratory, designing radiological monitoring instruments, including one of the very earliest applications of complementary metal-oxide semiconductor (CMOS) microprocessors to a pocket-sized instrument.

As well as authoring Electronic Components and Technology, Dr. Sangwine coedited The Colour Image Processing Handbook (Chapman & Hall, 1998), and he has also authored or coauthored more than 70 papers, the majority in the eld of image processing.

His principal research interest is in linear vector ltering and transforms of vector signals and images, especially using hypercomplex algebras, on which he collaborates with researchers in the United States and France. In 2005, he was a chercheur invité (visiting researcher), Centre National de la Recherche Scientique (CNRS), at the Laboratoire des Images et des Signaux, Grenoble, France, for 7 months, with nancial support from the Royal Academy of Engineering, United Kingdom.

Dr. Sangwine has been a senior member of the Institute of Electrical and Electronics Engineers (IEEE) since 1990.

xiii

Introduction

Modern electronic engineering products are found in a wide range of applications environments from the oor of the deep ocean (submarine cable repeaters) to geostationary orbit (microwave transceivers on board communications satellites), from the factory oor (industrial process controllers and numerically controlled machine tools) to the ofce (computers and printers). They can be found in the home (audio and video systems and microwave ovens), in schools (computers and pocket calculators), in hospitals (computerized tomography [CT] and magnetic resonance imaging [MRI] scanners, bedside monitors), inside the human body (heart pacemakers), and inside road vehicles (electronic ignition and engine management, antilock braking). Electronic products can also be found in the pocket (portable phones, personal audio, and video players). These products may be mass produced by the million, or they may be one-off special systems. They may be intended to last for decades, or they may be designed deliberately for a fairly short life. They should all be t for their intended purpose and be of signicant use to their users.

All electronic products depend on the physical and electrical properties of insulating, conducting, and especially semiconducting materi-

als, but by and large, the designer of an electronic product works with components and technologies, such as integrated circuit (IC) tech-

nology, rather than with basic materials. A critical aspect of product design is the interconnection of components, and for this reason this book starts with a chapter covering the technology of interconnection. The technology of interconnecting electronic components, circuits, and subsystems was, until the publication of this book, often neglected in electronic engineering texts at degree level. It is true that the detailed layout of a printed circuit board (PCB) is not a task likely to be undertaken by a graduate engineer unless the PCB is to carry high-frequency or high-speed circuitry. Nevertheless, a PCB has electrical properties and its design, together with the choice of components to go on it, can have a signicant effect on the performance, the cost of production, the production yield, and the reliability and maintainability of the assembled board, and quite likely the product of which it is a part. Jointing techniques, especially soldering, are of tremendous importance in electronic engineering, and solder is an engineering material that should be specied as carefully as a mechanical engineer species structural steel: what type of solder is best suited to a particular application? In many cases, just “solder” will not do.

The third chapter deals with IC technology. Only a few engineers are involved in high-volume IC design, but a more signicant number design or use semicustom ICs. Consequently, the treatment in this book is not for the IC specialist: it is aimed at the much larger group of electronics engineers who will be using ICs or designing a gate-array or standardcell IC of their own.

Competent electronics engineers need a good understanding of the components and subsystems from which their designs will be constructed and the instruments needed to test and characterize prototypes. They must be aware of not only the ideal behaviour of components, subsystems, and instruments, but also their performance limitations. The next three chapters, therefore, cover power sources and power supplies (an important class of electronic subsystem), passive electronic components, and instruments and measurement. To understand the performance limitations of components, an engineer must appreciate how the components are fabricated. To understand the performance limitations of power supplies and instruments, an engineer must appreciate the principles on which they operate. Chapters 5 and 6 cover these topics as well as provide factual information for reference.

The study of electronic components introduces the third major theme of this book: the parasitic effect. Real electronic components and circuits, as opposed to ideal ones, possess parasitic properties that are incidental to their intended properties. A wire-wound resistor, for example, is also inductive and has an impedance that varies with frequency. Heat is produced in signicant quantity in some electronic systems, and positive design measures often have to be taken to remove it. Electromagnetic energy can radiate from electronic circuits and couple into other circuits, causing faulty operation. A chapter has been devoted to this and other parasitic electromagnetic effects. This book does not attempt to cover all possible parasitic effects: to do so would be impossible even in a much larger book and would serve little useful purpose. Electronics engineers must learn to expect parasitic effects and try to take them into account when designing electronic products.

So far this introduction has dealt with matters that affect design and performance in ways that are important at the beginning of the life of a product. Without an understanding of components, technology, and parasitic effects, the design engineer will not be able to design good products that meet the required level of performance at the required cost. Many electronic products, however, will have a life that lasts far longer than the designer’s interest in the design. It is during the operating life of a product that long-term effects become important. Components and materials age: they deteriorate physically and chemically, and ultimately they fail. The study of these problems and of the prediction of product life is known as reliability. Not surprisingly, the reliability of a product can be inuenced by its design, for better or for worse, and if a product is capable of being repaired, the ease and expense with which it can be restored to working order can also be affected by decisions taken at the design stage. Reliability can also be inuenced by a product’s operating environment. Did the designers consider the effects of temperature, humidity, corrosion, and dust? Is there some unknown environmental factor that will doom their product to early failure? As with parasitic effects, after introducing some of the many environmental hazards to electronic equipment, this book leaves the readers to consider what the problems of their products’ environment might be.

Lastly, this introduction has dealt with the electronic product itself: will it work and continue to work for long enough? Will it succumb to environmental stress? Engineers must also look at their designs from

another viewpoint: will they do anyone, or the environment, any harm? All design engineers, including those working in electronic engineering, have a professional duty to consider safety when designing products, and in many countries a statutory (that is, legal) duty also. The nal chapter introduces the subject of safety in electronic engineering.

Interconnection technology

Objectives

To emphasize the importance of interconnection in electronic

□

product design. To discuss jointing technology, especially soldering and solderless

□

wire-wrapping. To outline the main types of discrete wiring and cabling.

□

To describe the technology of printed circuits.

□

To give an introduction to rework techniques.

□

To present a short case study illustrating the importance of

□

interconnection in industrial product design.

All except the smallest of electronic systems are built up from subsystems or subassemblies that are in turn built from electronic components such as resistors, capacitors, transistors, integrated circuits (ICs), displays, and switches. A desktop personal computer, for example, is likely to be built from a power supply subsystem, a main circuit board, and a number of peripheral subsystems such as a CD/DVD drive and plugin memory modules. Small self-contained electronic products such as pocket calculators and portable phones are often built directly from components with no identiable subsystems.

From the lowest level of component up to the system level, the constituent parts of an electronic system have to be interconnected electrically. The lowest level in the hierarchy of interconnection is the electrical joint. From the very earliest days of electronics, long before the invention of the transistor and integrated circuit, soldering has been an important technique for making electrical joints. Hand soldering is still used in prototype work, repair work, and, to a much lesser extent, production. Not all electrical joints in an electronic system need to be soldered: the technology of solderless wire-wrapping is well established in digital electronics, for both prototype and production wiring, and joints can also be made by insulation displacement, welding, or crimping. Components and subsystems are interconnected by wiring that can be in the form of either discrete wires and cables or printed circuits. A short case study at the end of this chapter illustrates the trend in electronic engineering over the last 20 years towards printed circuit interconnection wherever possible, avoiding discrete wiring because of the high cost of assembling and inspecting individual wires.

The printed circuit board, or PCB, is tremendously important in almost all applications areas of modern electronics. Not only does it provide a cheaply mass-produced means of interconnecting hundreds or thousands of individual components, but it also provides a mechanical mounting for the components.

In the very early days of

In the very early days of electronics when thermionic

electronics when thermionic (vacuum tube) valves were used,

(vacuum tube) valves were used, interconnection with discrete

interconnection with discrete wiring was normal — but this

wiring was normal — but this was soon superceded with the

was soon superceded with the advent of transistors by the

advent of transistors by the introduction of printed circuit

introduction of printed circuit boards.

boards.

Copper Copper

Before soldering

Air gap

Solder

After soldering

Figure 2.1 Diagrammatic representation of a soldered joint (not to scale).

Jointing

Several techniques are used in electronic engineering for making electrical joints. The most important of these is soldering, which is used mainly for attaching and jointing components to PCBs, but it is also widely used for jointing in cable connectors. Another important technology is solderless wire-wrapping, which nds application in prototype wiring for logic circuits and in production wiring of backplanes interconnecting PCBs. Welding is used in some specialized electronic applications and is a very important jointing technique in integrated circuit manufacture. Finally, in applications where soldering or welding cannot be used, mechanical crimping can make sound electrical joints.

Soldering

Solder is a low-melting-point alloy of tin and other metals, used for making electrical and mechanical joints between metals. Soldering does not melt the surfaces of the metals being joined, but adheres by dissolving into the solid surface. Figure 2.1 shows an idealized cross-section of a soldered joint illustrating this point. Soldered joints can be made by hand using an electrically heated soldering iron or by a mass-soldering process in which all the joints on a PCB are made in one automated operation. Both techniques are important, and they are described in detail below and in a later section of this chapter.

In the twentieth century, solders used in electronics were almost always alloys of tin and lead. From July 2006, the European Union (EU) has required lead (and many other toxic substances) to be eliminated from electronic and other products, and therefore lead-free solders are now used in place of tin–lead solders. Because of the historical importance of tin–lead solder in electronics, we start with its properties before progressing to the characteristics of lead-free solders.

Commercial tin–lead solders were available with several different proportions of tin to lead and with traces of other metals to enhance their properties. Tin and lead are soft metals with melting points of 232°C and 327°C respectively. Alloys of these metals generally start to melt at a temperature of 183°C, which is lower than the melting temperature

100

0:100 20:80 40:60

Tin:lead ratio

(by weight)

Liquid

Plastic

19%

tin

63:37 97%

tin

Solid

60:40 80:20 100:0

232

200

300

327

400

183

°C

Figure 2.2 Phase diagram for tin–lead solder alloys (simplied).

of either pure metal. Figure 2.2 shows a simplied phase diagram for tin–lead alloys. The ratio (by weight) of tin to lead is plotted horizontally with 100% lead on the left and 100% tin on the right. The vertical scale represents temperature. In the top region of the diagram, above the line extending from the melting point of lead at 327°C across to the melting point of tin at 232°C via the point at a tin:lead ratio of 63:37 and a temperature of 183°C, the alloys are liquid. The bottom region of the diagram represents the ranges of temperature and tin:lead ratio over which the alloys are solid. The two triangular regions represent temperatures and compositions where the alloys are in a plastic state consisting partly of solid and partly of liquid. The alloy with a tin:lead ratio of 63:37 is the only one that changes sharply from solid to liquid at a single temperature. This alloy is known as a eutectic alloy. It is fully liquid at the lowest possible temperature for a tin–lead alloy.

For general electronic jointing, a 60:40 solder was used, which is fully molten at about 188°C. 40:60 solder, which is fully molten at about 234°C, was also readily available for applications where a higher melting point was needed.

The most common lead-free solder for use in electronics is an alloy of tin (Sn), silver (Ag), and copper (Cu) in proportions of about 95 to 96% tin, 3 to 4% silver, and 0.5 to 1% copper. Alloys of this composition have a melting point of around 215 to 218°C, which is somewhat higher than the melting point of a 60:40 tin:lead solder. Figure 2.3 shows part of the phase diagram for the tin–silver–copper alloys (the full phase diagram takes the form of an equilateral triangle, but since a practical alloy for electronic soldering has 95 to 96% tin, only one corner of the full diagram is shown). Temperature is shown by contour lines, since to plot them would require a third dimension, out of the page. The higher melting point of lead-free solders means that more care has to be taken to control soldering processes, since the higher temperatures could more easily cause thermal damage to electronic components.

Four requirements must be met if a good soldered joint is to be made, and an understanding of these is essential in order to develop skill at hand soldering. First, the surfaces to be joined must be solderable. Not

Phase diagrams in general and

Phase diagrams in general and the tin–lead phase diagram

the tin–lead phase diagram in particular are discussed by

in particular are discussed by Anderson et al. (1990).

Anderson et al. (1990).

The proportion of tin (by weight)

The proportion of tin (by weight) is conventionally stated rst.

is conventionally stated rst.

PCBs sometimes incorporate

PCBs sometimes incorporate a solderability test pad in an

a solderability test pad in an unused corner of the board,

unused corner of the board, so that the solderability of the

so that the solderability of the board or a batch of boards can

board or a batch of boards can be checked before component

be checked before component assembly.

assembly.

Mercury is toxic. Do not try this

Metal

terminal

Globular solder with no “wetting”

of surfaces

Wire

Correctly “wetted” solder joint

1 mm (typ.)

Solder wire with

integral ﬂux cores

Metal

terminal

Globular solder with no “wetting”

of surfaces

Wire

Correctly “wetted” solder joint

1 mm (typ.)

Solder wire with

integral ﬂux cores

0.5 1.0 1.5 2.0 2.5 3.0

226

222

320

(Sn)

Ag3Sn

Mass % Ag

Mass % Cu

Mercury is toxic. Do not try this experiment.

experiment.

Figure 2.3 Partial phase diagram for tin–silver–copper lead-free solder alloys. (Courtesy of Dr. Ursula Kattner, National Institute of Standards and Technology [NIST], United States.)

all metals are readily soldered without special techniques. Aluminium, for example, is an extremely reactive metal that rapidly oxidizes on exposure to air to form a passivating oxide layer that prevents solder from alloying with the underlying metal. It is possible to demonstrate the reactive nature of aluminium by removing the oxide layer with a little mercury rubbed onto the aluminium surface. The freshly exposed surface reacts rapidly with moisture in the air, and the metal becomes hot. Gold is very easily soldered because the metal does not oxidize. Copper and brass oxidize readily, but the oxide layer can be removed easily, so that these metals are solderable. The second requirement for a good soldered joint is cleanliness: the surfaces to be joined must be free of grease, dust, corrosion products, and excessively thick layers of oxide. In most electronics applications, the surfaces to be soldered will have been plated with gold or a solder alloy during manufacture in order to provide a readily solderable surface. Coating a metal with solder is known as tinning, and protects the metal from oxidation. Cleaning is not usually needed, therefore, unless a component or PCB has been stored for a long time or becomes contaminated with grease or dirt. The third requirement is that any layer of oxide on the surfaces must be removed during soldering and prevented from regrowing until the molten solder has alloyed with or “wetted” the surface. This is achieved by a ux that chemically removes the oxide layer and reduces surface tension, allowing molten solder to ow easily over the surfaces to be joined.

For hand soldering, solder wire with integral cores of ux is used. The uxes commonly used for electronic work are rosin based and chemically mild, leaving a noncorrosive residue. More powerful uxes based on acids may be needed on less solderable metals, but must be thoroughly cleaned off afterwards because they leave a corrosive residue. The nal requirement for a good soldered joint is heat. Both surfaces

to be joined must be heated above the solidication temperature of the

Wire and terminal

cut into each other

at pin corners

One or two turns

of insulation

provides strain

relief

About seven turns of bare

wire evenly and

closely spaced

solder, otherwise the solder will chill on contact and will fail to ow evenly and alloy with the surfaces. When soldering by hand, heat transfer from the soldering iron to the joint is improved if there is a little molten solder on the tip of the soldering iron. The joint should be heated with the iron, and the solder wire applied to the joint (not the iron). The iron should not be removed until the solder has owed into the joint. It is very important that the joint is not disturbed until the solder has fully solidied, otherwise a high-resistance (dry) joint will result from mechanical discontinuities in the solder.

Wire-wrap jointing

Soldering is a good technique for mass jointing on PCBs, but it has several disadvantages for discrete wiring joints. An alternative jointing technique exists for logic circuits and low-frequency applications known as solderless wire-wrap. Special wire and terminal pins are used for wire-wrap jointing, and special hand-operated or electrically powered tools are required. Various types of IC sockets and connectors, including PCB edge connectors, are made with wire-wrap terminal pins. Wire-wrap interconnection is used for prototype and production wiring of logic boards and for production wiring of backplanes for interconnecting a rack of PCB subunits. It is a faster technique than soldering, requires no heat, produces no fumes, and can easily interconnect terminals spaced as little as 2.5 mm apart. Figure 2.4 illustrates a typical joint. The terminal pin is typically 0.6 to 0.7 mm square and 15 to 20 mm long. The wire is solid and about 0.25 mm in diameter. The joint consists of about seven turns of bare wire and one to two turns of insulated wire, wound tightly around the terminal pin. At each corner of the terminal pin, the wire and the pin cut into each other, making a metalto-metal connection, which improves with age through diffusion of the two metals. The wire is under tension and slightly twists the square pin. The insulated turns of wire act as a strain relief at what would otherwise be a weak point of the joint.

Wire-wrapping was rst

Wire-wrapping was rst developed at Bell Telephone

developed at Bell Telephone Laboratories, New Jersey, United

Laboratories, New Jersey, United States, circa 1950.

States, circa 1950.

Figure 2.4 A wire-wrap joint.

Oﬀset hole for wire

Centre hole to ﬁt

terminal pin

End view of a wire-wrap bit

Oﬀset hole for wire

Centre hole to ﬁt

terminal pin

End view of a wire-wrap bit

Tracy Kidder’s classic book

Wrong – “Daisy-chained” connections

Right

Tracy Kidder’s classic book

The Soul of a New Machine,

rst published in 1981, tells the

rst published in 1981, tells the story of a group of computer

story of a group of computer engineers working on a 1970s

engineers working on a 1970s minicomputer and testing a

minicomputer and testing a prototype machine wired with

prototype machine wired with wire-wrap technology.

wire-wrap technology.

Figure 2.5 Right and wrong ways of wire-wrapping a group of terminals.

The tools required for making wire-wrap joints are more expensive than those needed for making soldered joints, but the time saved in production work soon covers the cost of the tools. The wire is stripped using a tool that ensures that the correct length of insulation is removed. The stripped end of the wire is then inserted into the offset hole of a wrapping bit, which may be part of a hand-operated or powered tool. The centre hole of the wrapping bit is then slipped over the terminal pin, and the tool is rotated evenly to wrap the wire around the pin. Most wire-wrap terminal pins have enough length for up to three joints.

One of the very signicant advantages of wire-wrapping for prototype wiring is the ease and speed with which joints can be unwrapped to allow circuit modications. A special tool is required to unpick a joint, and any joints further out along the pin have to be unpicked rst. For this reason, when a group of pins are to be connected, each wire should be at the same level on both pins connected. Figure 2.5 shows the right and wrong ways of connecting a series of pins. In the “daisy-chained” arrangement, a change to one wire can often require many other wires to be unpicked and remade. Because of the cutting action of the wire on the corners of the terminal pins, there is a limit to the number of times that a pin can be rewired. The small number of modications likely to be needed to build a prototype are not likely to cause problems with poor joints, but continual reuse of wire-wrap IC sockets could be more trouble than the sockets are worth. The wire removed from an unpicked joint is not reusable, so in nearly all cases, both ends of a wire have to be unpicked. The unpicked wire should be carefully removed from the circuit and thrown away, taking care that fragments of stripped wire do not fall back into the circuit. An intermittent short circuit caused by a piece of loose wire can take a long time to diagnose.

One nal point about wire-wrapped connections is that there is little point in trying to arrange the wires into tidy bundles: direct point-topoint wiring can be easier to inspect and check, and minimizes problems with crosstalk.

Insulation displacement

Insulation displacement is a mechanical jointing technique in which an unstripped insulated wire is pushed between the sharp edges of a forked

terminal. The wire is thus held mechanically and the insulation is displaced by the cutting edges of the terminal, making an electrical joint. The two sides of the terminal are pushed slightly apart by the insertion of the wire and thus exert spring pressure on the conducting core of the wire. The technique is widely used for connecting discrete wires in the cabling of domestic and commercial telephone sockets, because it is fast and does not require soldering or power tools.

Welding

Welding is a jointing technique where two metal surfaces are placed in intimate contact and then fused together by melting both surfaces. Heat can be applied by a ame, thermal conduction, or electric heating. There are limited applications of welding in electronics, except in the bonding of ICs to their packages. External connections from bonding pads on an IC are made by attaching ne gold wires using either ultrasonic welding or thermocompression bonding, as described in the next chapter.

Crimping

Insulation displacement jointing

Insulation displacement jointing is also used in some types of

is also used in some types of cable connector, as discussed

cable connector, as discussed later in this chapter.

later in this chapter.

A fth jointing technique used in electronics (and much more extensively in electrical engineering for heavier currents) is crimping. A crimped connection is made by crushing a special terminal onto a wire of the correct size using a purpose-made tool. The wire is gripped mechanically by the crushed terminal. Electrical contact depends on the mechanical integrity of the joint. Unlike a wire-wrapped or soldered joint, the electrical connection is not gas tight and can therefore be prone to corrosion. Crimping is an especially useful technique for jointing unsolderable wires such as aluminium, and for rapid wiring assembly in production.

Discrete wiring

Modern electronic product designers tend to avoid using discrete wiring in favour of printed-circuit interconnection because of the high cost of hand jointing and the likelihood of errors. Some wiring is nearly always needed, however, especially in larger systems.

A wire is a single or multistranded conductor with or without insu-

lation, whereas a cable is a collection of wires or conductors bound together, possibly with some overall insulation or other protection. Equipment wire used in electronics is normally made of copper or tinned copper with either polyvinyl chloride (PVC) or polytetrauoroethylene (PTFE) insulation. PVC-covered wire can be used at up to 70°C. Above this temperature, PTFE-covered wire must be used. The size of a wire can be stated either as a cross-sectional area (in mm2) or as the number of strands followed by the diameter of each strand. Thus “7/0.2” represents seven strands of 0.2 mm diameter. IEC 60228 species standard wire cross-sections in mm2.

For many electronic purposes, the voltage and current ratings of the lightest equipment wire are far greater than the requirements of the circuit (Table 2.1). A 7/0.2 PVC-covered wire, for example, is rated at

1.4 A (this is fairly small wire). This rating is conservative (the wire

Crimping is very widely used

Crimping is very widely used in the automotive industry for

in the automotive industry for rapid assembly of vehicle wiring

rapid assembly of vehicle wiring harnesses.

harnesses.

You may still nd wire sizes

You may still nd wire sizes quoted in British Standard Wire

quoted in British Standard Wire Gauge (SWG) or American Wire

Gauge (SWG) or American Wire Gauge (AWG). You will need a

Gauge (AWG). You will need a table of wire gauges to nd out

table of wire gauges to nd out the cross-sectional area.

the cross-sectional area.

Table 2.1 Current ratings of copper equipment wire

Construction Cross-section area (mm2) Current rating (A)

(a) PVC insulated

7/0.2 0.22 1.4 16/0.2 0.5 3.0 24/0.2 0.75 4.5 32/0.2 1.0 6.0

(b) PTFE insulated

7/0.15 0.12 3.5 7/0.2 0.22 6.0 19/0.16 0.38 9.0

could carry a greater current without overheating) to allow for the possibility of several wires being bundled together in a conned space. A 7/0.2 PTFE-covered wire is rated at 6A because the insulation can withstand higher temperatures caused by self-heating. As the following worked example shows, however, there will be a considerable voltage drop along wire of this cross-section carrying 6A.

Worked Example 2.1

Calculate the voltage drop along 500 mm of 7/0.2 PTFE-insulated copper equipment wire carrying a current of 6A. The resistivity of copper is 1.7 × 10–8 Ωm.

Solution

The cross-sectional area, A, of the wire is 7π × (0.1)2 mm2 or

0.22 mm2. The resistance of a length, l, resistivity, ρ is

Hence the resistance of the 500 mm length is

1.7 × 10

–8

× (0.5)/0.22 × 10–6, or about 40 mΩ

From Ohm’s law, the voltage drop is about 240 mV. If this wire is used to connect a power supply to a load only half a metre away, the voltage at the load will be 0.48 V less than the voltage at the terminals of the power supply (there will be a drop of 0.24 V in each conductor) if the full-rated current of the wire is drawn.

This solution has ignored any change in resistance due to self-heating

of the wire that would increase the voltage drop.

Cables

Cables are used mainly for signal and data transmission and for interconnecting subsystems within an electronic system. Table 2.2 summa rizes the main types of cable used in electronic engineering, and these

Table 2.2 Common cable types

Type Construction Applications Screened One or more wires with an

overall metal braid or helical screen and insulation.

Twisted pair Two wires insulated and

twisted together, covered with overall sheath, and possibly screened.

Coaxial One solid or stranded

conductor surrounded by dielectric, metal braid, and outer insulation.

Twin feeder Two conductors laid

parallel about 10 mm apart, insulated, and separated by a web of plastic.

Ribbon 10 to 50 stranded

conductors laid parallel and coplanar, covered and separated by insulation.

Low-power signal

transmission at up to audio frequencies.

Signal transmission at up

to 100 MHz.

Signal transmission at up

to 1 GHz.

Radio receiver antenna

downleads.

Parallel logic

interconnection in microprocessor and computer systems.

are illustrated in Figure 2.6. Multicore cables can combine these types within one cable for special applications. Each type of cable has its own range of uses and its own characteristic parameters.

There are only a few applications in electronic engineering for cables of the type known as ex. These consist of several insulated wires with overall insulation, such as three-core mains ex used to connect mainspowered equipment to a mains outlet (this is one of the few applications). The reason for this is that cables are designed to carry electromagnetic signals and the cable must either exclude unwanted signals present in the surroundings (interference) or else prevent energy escaping from the cable (and causing interference elsewhere). A screened cable is intended to prevent pickup of unwanted signals. The wire or wires within the cable carry the signal (typically from a microphone or other transducer) and are surrounded by a metal screen wound helically or woven from bare wires in the form of a braid. The screening is effective only against electric elds and high-impedance electromagnetic elds (with a strong electric component). Magnetic elds and low-impedance electromagnetic elds (with a strong magnetic component) cannot easily be screened against. The effect of a magnetic eld on a cable can be reduced, however, by twisting a pair of wires together. This means that currents induced in the wires by a changing magnetic eld tend to cancel because each twist of the wires reverses the polarity of the wires relative to the eld. A twisted-pair cable can be used to transmit

Chapter 8 discusses

Chapter 8 discusses electromagnetic effects in

electromagnetic effects in greater detail. Carter (1992)

greater detail. Carter (1992) discussed electrostatic and

discussed electrostatic and magnetic screening.

magnetic screening.

Outer insulation

(a)

Stranded screen

(b)

(c)

(d)

(e)

Braided outer

conductor

Inner conductor

Conductors

Dielectric

The theory of transmission lines

The theory of transmission lines is covered by Carter (1992).

is covered by Carter (1992).

Figure 2.6 Common cable types: (a) screened, (b) twisted pair, (c) coaxial, (d) twin feeder, and (e) ribbon (insulation displacement).

frequencies of up to 100 MHz, but energy will be radiated at megahertz frequencies and a coaxial cable should properly be used for the higher frequencies. Signicant applications of twisted-pair cables for high-frequency signals include 100 megabit/s (Mbit/s) Ethernet, and digital sub scriber line (DSL)/asymmetrical digital subscriber line (ADSL) (which is carried over twisted-pair telephone cables). Coaxial cables will carry signals down to zero frequency, but their main use is for transmission of radio frequency (r.f.) signals at up to 1 GHz.

Any cable that is longer than the wavelength of the signals being carried must be regarded as a transmission line. Cables designed to carry signals of frequency higher than audio frequencies therefore have characteristics which include transmission-line parameters. The two most important characteristics are the characteristic impedance, Z0, which is typically 50 to 150 Ω and the attenuation, α, which is usually stated in dB per metre, 100 m, or km (the frequency must also be given, since α depends on frequency). The characteristic impedance is independent of the length of a cable. Three other parameters commonly stated for a cable are the capacitance per metre (typically < 100 pF), the maximum working voltage, and the operating temperature range.

Ribbon cables are widely used in microprocessor and computer systems to carry parallel logic signals over distances as long as 5 to 10 m, but typically much shorter. Their main advantage over conventional multicore cables is that they can be mass terminated: a connector can

Table 2.3 Common types of cable connectors

Type Construction Features/applications IEC 60320 C14 Three-pole male body

with female contacts, female sockets with male contacts.

BNC 50/75 Ω

Coaxial bayonet. Instruments, general

10A rating mains

connector, sockets available with integral lters.

screened and coaxial connector.

“D” type 9- to 50-way connectors

with contacts in two or three rows available for soldering, wirewrapping, ribbon cabling, and PCB mounting.

RJ11/RJ45 Plastic shell with 4 or

8 sprung-nger contacts and latch.

Widely used multipole

connectors, 25-way version used for “RS-232” data transmission, 9-way for serial ports on computers.

Telephones, local area

networks (LANs, or Ethernet).

be tted to the cable in one operation taking less than a minute. If a ribbon cable is carrying high-speed digital signals, each conductor should operate as a transmission line. For this reason it is common practice to earth alternate conductors. If an even number of signals, n, is to be carried, as is normally the case, the cable should have 2n + 1 conductors so that all signal conductors have an earth on both sides. This ensures that the characteristic impedances of all the signal conductors are equal.

Connectors

Connectors, or plugs and sockets, are used in electronic products to make electrical connections that can be easily disconnected and reconnected. They are used to connect external cables to equipment and are also tted internally to allow subassemblies (such as PCBs) to be disconnected and removed easily for repair. One can therefore classify connector types into two main categories: cable connectors, suitable for making external connections to equipment; and wiring connectors, designed for use internally.

Table 2.3 lists a few common types of cable connectors and is limited to widely used, standardized designs. Some of these are illustrated in Figure 2.7. There are many types of proprietary connector, especially for use with multicore cable. Most cable connector types are keyed or polarized mechanically in some way so that they can be connected in one position only. Connectors designed to be attached to the end of a cable are called free connectors and incorporate some form of strain relief to grip the cable mechanically so that tension in the cable is not transmitted to the electrical joints inside the connector. Fixed connectors are designed for mounting on a panel or PCB.

(a) (b)

(e)

Figure 2.7 Five common types of cable connectors: (a) IEC 60320 type C14 3-pole 10 A mains connector, (b) inlet socket, (c) BNC 50 light-emitting diodes [LEDs]). (Courtesy of (a,b) Bulgin Components; (c) Jonas Bergsten; (d) Amphenol-Socapex, France; and (e) Amphenol, Canada.

Ω coaxial, (d) 9-way male “D”-type, and (e) RJ45 Ethernet jack socket (with integral

The contacts within a connector are referred to as male or female. When two mated connectors are separated, the live side of each circuit, if any, should be on the female contacts. The male contacts (which are accessible) should be electrically dead. In nearly all cases, this is a positive safety requirement and is one of the consequences of the widely accepted safety rule that live parts shall not be accessible.

Some of the factors to be considered when choosing a connector are: the electrical ratings and characteristics such as maximum working voltages and currents, contact resistances, insulation resistance, and transmission line parameters; the temperature ratings and intended operating environment; the reliability and life of the connector; cost; and tooling needs for making the electrical connections. The transmission line parameters are applicable only to r.f. connectors and will include the characteristic impedance and the voltage standing wave ratio (VSWR).

The environmental conditions under which a connector will be working are most important. Connector types capable of use out of doors (splashproof or waterproof) are much more expensive than types intended for indoor use, because of the complexity of the waterproof seals and the need to protect the connector from atmospheric corrosion. The life of a connector is inuenced by the number of mate–unmate operations: a heavy-duty connector is designed to withstand the wear and tear of frequent use, while other types are designed for occasional mating and unmating only. Many cable and connector types require special tooling to prepare the end of a cable and to make the electrical connections to the connector. Additionally, some training and skill are needed if a good-quality termination is to be made. A possible solution to this problem is to buy-in terminated cables from a specialist cabling contractor. Some common cable and connector congurations are available commercially as ready-made cable assemblies. This is especially true for mains voltage equipment leads tted with IEC 60320 connectors at one end, and national mains plugs at the other end, and also for telephone and local area network cables with RJ-11 and RJ-45 jacks.

Wiring connectors are somewhat less standardized than cable connectors, and different manufacturers’ products are often not interchangeable. They are used for making connections to the edges of PCBs (edge connectors), for attaching ying leads to PCBs, and for connecting ribbon cables. (Some ribbon cable connectors are suitable for making external connections, but in general their use is conned to interior interconnections.) Many wiring connectors are attached to wires by crimping rather than soldering, because crimping is a much quicker operation in production.

Ribbon cable connectors make electrical contact with the cable conductors by insulation displacement jointing, as discussed earlier in this chapter. A typical proprietary contact design is illustrated in Figure 2.8a. All the connections within the connector (up to 50) are made simultaneously in a single pressing operation, making this type of cable and connector an economical and virtually error-free method of interconnecting logic PCBs. Ribbon cable connectors are also available for connection to dual-in-line (DIL) IC sockets, as shown in Figure 2.8b.

Where individual wires have to be connected, push-on terminals (of the type commonly used in automobile wiring) or screw terminals are

When an electromagnetic wave

When an electromagnetic wave travelling along a cable meets an

travelling along a cable meets an electrical discontinuity such as

electrical discontinuity such as a connector, some of the wave

a connector, some of the wave energy is reected and sets up

energy is reected and sets up a standing wave. The VSWR

a standing wave. The VSWR is a measure of the amount of

is a measure of the amount of reection from the discontinuity.

reection from the discontinuity. Chapter 7 of Carter (1992)

Chapter 7 of Carter (1992) discusses VSWR and other

discusses VSWR and other transmission-line parameters,

transmission-line parameters, such as the characteristic

such as the characteristic impedance.

impedance.

Sharp edges for

piercing and

stripping insulation

Four contact

points on wire

Stranded conductor

Assembly pressure

(a)

(b)

Rear slot

wire grip

Hardened

beryllium copper

General references for this

General references for this section are Scarlett (1984) and

section are Scarlett (1984) and Edwards (1991).

Edwards (1991).

A prototype PCB can be

A prototype PCB can be manufactured by computer-

manufactured by computercontrolled milling, in which

controlled milling, in which copper is removed by a cutting

copper is removed by a cutting tool. This avoids the use of

tool. This avoids the use of chemicals.

chemicals.

Figure 2.8 Insulation displacement connectors: (a) contact arrangement and (b) DIL plug connector assembled to cable. (Courtesy of Thomas & Betts Ltd. Design covered by U.S. Patent 3.964.816.)

often used. Some screw terminals are specically designed to accept bared wires (tinned with solder if multistranded); others are designed to take a spade or eyelet terminal crimped to the end of a wire.

Printed circuits

Printed circuits are used in almost all application areas of electronic engineering. Rigid printed circuit boards account for the majority of applications, providing both mechanical mounting and electrical interconnection for components. Flexible printed circuits are also popular as a substitute for discrete wiring.

Not all PCBs have electronic components mounted on them, and a signicant use for PCBs is to provide interconnections among other PCBs, particularly in card cages or card racks where a number of boards slide into a frame and connect with a backplane that provides electrical interconnection among the boards. Another possibility is to use a PCB as an electrical and mechanical base on which to mount subboards containing functional blocks of circuitry of standardized design. This is known as the motherboard–daughterboard technique, and it is used both to permit modular design and to simplify manufacture by allowing easy assembly of multiple PCBs into a small space.

Printed circuits are usually manufactured by chemical etching and electroplating processes. The patterns of conductors, or tracks, on a

printed circuit are dened photographically from a master photographic

Laminate

Copper

Plated-through hole (PTH)

PTH with connection to upper

internal layer

PTH with no connection

to internal layers

(a)

(b)

(c)

lm. The lm itself is usually made by photoplotting from computeraided design (CAD) software. For some very specialized applications, the lm may be made by photographing manually prepared artwork. Before the days of cheap computers, this was the normal method for preparing PCB lms.

Rigid printed circuit boards

There are three types of rigid printed circuit boards shown diagrammatically in Figure 2.9. Single-sided boards with a conductor pattern on one side only are the cheapest type. They are mainly used for very low-cost, low-component-density, consumer applications such as portable radios. Double-sided boards can carry a greater density of circuit interconnections and are much more common than single-sided boards. Double-sided boards normally have plated-through holes (PTHs), so that interconnections between one side of the board and the other are made during board manufacture. Through-hole plating is a process by which copper is deposited on the inside walls of holes drilled through the board. A thin layer of copper is rst deposited by electroless plating onto the nonconducting board material. The copper thickness is then

Artwork is a term used in the

printing and publishing trades for

printing and publishing trades for image material to be reproduced

image material to be reproduced photographically.

photographically.

Figure 2.9 PCB construction: (a) single-sided, (b) doubled-sided platedthrough, and (c) multilayer.

Multilayer boards can be

Multilayer boards can be fabricated with more than 20

fabricated with more than 20 conductor layers, although fewer

conductor layers, although fewer than 10 is a more usual gure.

than 10 is a more usual gure.

built up by electroplating. Double-sided boards without plated-through holes were common in the 1980s because they were cheaper to manufacture, but they are now too expensive to assemble because link wires or pins have to be soldered through some of the holes to make connections between one side of the board and the other, and through-hole plating is now a much cheaper alternative. Multilayer boards have internal layers of conductors as well as the conductors on the outer faces of the board. Connection to the internal layers is by PTHs. The manufacturing process for multilayer boards is more elaborate (and therefore more expensive) but makes possible higher-density boards than could be achieved otherwise. Multilayer boards also offer better electrical performance, for reasons that are covered in Chapter 8. They are used almost universally in computers.

Flexible printed circuits

Flexible circuits can be made in the same congurations as rigid boards, including through-hole plating. Multilayer exible circuits, however, tend to have only a few layers because exing of the board puts stress on the copper layers. Flexible circuits may be used instead of discrete wiring or to achieve a compact assembly of circuitry by folding a circuit into a small volume, which can be especially important in some military applications such as missiles. A further signicant use for exible circuits is in applications where mechanical movement must be allowed and, in these applications, the exible circuit will be subjected to repeated exing throughout its life. A common example is in computer hard disk drives, where the moving head is connected via a exible circuit to the main (rigid) circuit board. Components may be mounted on a exible circuit, but the use of larger components may require a exirigid board fabricated in one piece with some sections exible and some rigid. Flexible circuits are usually based on a polyimide plastic lm with copper layers. Unlike rigid boards, the outer faces of the circuit are of polymer, not copper, to prevent delamination of the copper when the circuit is bent.

Component mounting

There are two methods of mounting components on a PCB, as illustrated in Figure 2.10. Through-hole mounting, in which component leads pass through holes in the board, is a technique that has been used from the advent of printed wiring, but has now been superceded for many applications by surface mount technology in which component leads or pads are soldered to the surface of a PCB without passing through the board. Surface mounting was tried during the 1960s, using welding to attach IC leads to copper, but it became feasible using solder jointing in the 1980s and 1990s. The main advantage of surface mounting is the smaller size of surface mount components, which allows greater component density on PCBs. Holes are still needed, of course, to make connections between one side of the board and the other, or to connect to internal layers, but these holes can be much smaller than those needed for component leads. Surface mount technology was driven by the need to make IC packages smaller (because of the large numbers of pins) but

Dual-in-line IC

7.6 mm

4 mm

PCB

Solder

(a)

(b)

Solder

Surface-mounting IC

Adhesive

Figure 2.10 Component mounting methods: (a) through-hole, and (b) surface mount.

ended up contributing to the miniaturization of electronic products as smaller components were developed for surface mounting.

Rigid board materials

Most professional printed circuits are made from breglass-reinforced epoxy resin laminate. The breglass reinforcement is usually in the form of woven cloth. For very low-cost applications (such as pocket radios or cheap electronic toys), synthetic-resin-bonded paper (s.r.b.p.) may be used. Fibreglass boards are more heat resistant and more stable dimensionally than s.r.b.p., and also absorb less moisture from the air, resulting in better long-term insulation resistance. Flame-retardant grades of breglass board are available for applications where there is a risk of severe overheating. One advantage of s.r.b.p. is that holes may be punched rather than drilled, which reduces the manufacturing cost provided the expense of preparing the punch tool can be recovered over a large number of boards produced.

Manufacturing processes

The starting point in the manufacturing process for single- and doublesided rigid boards without PTHs is a photographic lm, actual size, dening the pattern of conductors required on the board, and a piece of board material coated on one or both sides as appropriate with a layer of copper. The copper is then coated with photoresist, which is sensitive to ultraviolet light. The photographic lm is then laid in contact with the resist, and the whole assembly exposed to ultraviolet light. Exposed regions of the photo-resist undergo a photochemical change, leaving an imprint of the conductor pattern on the resist. The exposed board is then chemically developed in a tank of developer or in a conveyor machine

The thickness of the copper

The thickness of the copper layer is stated in micrometres

layer is stated in micrometres (µm) or as a weight per unit area

(µm) or as a weight per unit area (often ounces per square foot).

(often ounces per square foot). A common thickness is 35 µm

A common thickness is 35 µm (1 ounce-foot–2).

(1 ounce-foot–2).

Scarlett (1984) described many

Scarlett (1984) described many variants on the basic techniques

variants on the basic techniques discussed here.

discussed here.

in which developer is sprayed onto the board. This removes the resist, except in areas where copper will be left on the nal board. These areas remain covered by resist — so called because its job is to protect these areas of copper from chemical etching. After development, the pattern of conductors on the nal board is visible as the pattern of resist.

The developed board is then etched, usually in a warm solution of ferric chloride, to dissolve the unwanted copper. After etching, the remaining resist is removed with a solvent, and the board is ready for drilling.

Boards produced in quantity are drilled automatically on a numerically controlled (NC) drilling machine. Drilling coordinates for an NC drill can be generated automatically from CAD software. Drilling may also be done by hand in a vertical drilling machine, by eye (sight drilling) for a prototype board, or by using a jig for small-batch production work.

The manufacturing process for double-sided PTH boards starts with drilling of the blank board. The drilled blank has copper all over both faces to conduct electroplating current to the holes in the board. The inside walls of the holes are therefore plated with copper before the track pattern is etched onto the board. During etching, the etch resist protects the insides of the holes from the etchant. There are many variations on PCB manufacturing techniques that cannot be discussed here for lack of space.

Multilayer board manufacturing is a more elaborate process than double-sided PTH board manufacturing but has some steps in common. The internal copper layers of a multilayer board are etched individually, as described above for conventional boards, except that each layer of board material is thinner than conventional board laminate. Intermediate layers without copper, known as pre-preg, are not fully cured: the epoxy resin has only been partially heat-treated and is still capable of plastic ow under heat and pressure. When all the internal copper layers have been etched, the sheets of laminate and pre-preg are assembled together and bonded under heat and pressure in a press. Normally, a stack of boards is pressed at the same time, separated from each other by sheets of steel and PTFE-based plastic lm.

After bonding, the board is externally similar to a double-sided PTH board: the outer faces are still covered completely with copper. The board is drilled, and the through holes are plated in the same way as in a double-sided PTH board. With multilayer boards, of course, some of the plated holes make electrical contact with internal copper layers, so the quality of the drilled holes has to be good. Finally, after through-hole plating, the outer faces of the multilayer board are etched and nished in the same way as in double-sided PTH boards.

Solder resists and legends

Most printed circuit boards are coated with an epoxy resin material called solder resist, usually dark green in colour. The purpose of this coating is to prevent solder sticking to unwanted areas of the board during mass soldering, perhaps causing short circuits, or solder bridges, between adjacent tracks. It also serves two secondary purposes: less solder adheres to the board during soldering (solder is expensive), and the coating reduces moisture absorption during the board’s life, thus

improving reliability. The solder resist is applied as a liquid by screen printing through a ne mesh screen.

The nal manufacturing process, before assembly of components onto the board, is printing with a legend. This identies the component positions and reference numbers on the board, the board type number, and perhaps the date of manufacture or the version number. All of this information is useful if the board has to be repaired. The legend is screen printed onto the board on top of the solder resist with an epoxybased ink.

Printed circuit CAD

Most PCBs are designed using CAD software, and the lms required for board manufacture are generated by photoplotting. In the early days of CAD software, manual design and artwork preparation were still used for high-frequency and microwave boards, but today CAD software is available even for microwave boards.

The rst decisions to be made in designing a PCB are the size and shape of the board, if not already determined by the application, and the method of construction to be used. Unless there are electrical reasons for choosing multilayer construction, there is often a choice between multilayer and double-sided plated-through boards. The time taken to design a board can run into weeks, so if a double-sided design has to be abandoned partway through because of difculty in accommodating all the connections, a considerable amount of money and time will have been wasted. On the other hand, the choice of multilayer construction adds additional material and manufacturing costs to every board produced. Of course, if a product is a development, or variant, of an existing design, the choice made for the existing design will probably be kept.

When the board dimensions and outline have been decided, the second stage in the design process is to decide how and where to position components. The component placements will be inuenced by the logical structure of the circuit, by electrical requirements such as power distribution, and by the need to minimize interactions among different parts of the circuit.

The nal step in the design process is to nd a path for every connection in the circuit. In practice, this stage of design interacts with component placement, as some components may have to be rearranged. Finding a path, or route, for each connection is not easy, as tracks cannot cross each other on the same side or layer. A common design for double-sided boards is to have all the tracks on one side of the board running horizontally and all those on the other side running vertically. Where a track changes from one side of the board to the other, a via hole is normally used, not a component lead hole. The process of nding paths for the tracks is called routing and is often performed automatically by CAD software, although manual routing is also possible and may produce a layout with better electrical performance, as discussed in Chapter 8. Once the routing is complete, the photographic lms required for manufacture are generated in actual size by a photoplotter that exposes sheets of lm with the patterns for each side or layer of the

Sensitive amplier inputs, for

Sensitive amplier inputs, for example, would normally be

example, would normally be positioned well away from power

positioned well away from power supply rails and output signals.

supply rails and output signals.

The photographic lm used is

Solder

Solder bath

Weir Flux Preheat

Conveyor

PCBs with

components

Pump

The photographic lm used is lithographic lm: it reproduces

lithographic lm: it reproduces only clear or full black shades.

only clear or full black shades.

board. These lms are then developed by a wet chemical process in the same way as any other photographic lm.

Printed circuit assembly

Mass production of assembled and soldered PCBs is highly automated. Components are inserted into a board by high-speed automatic component insertion machines or, in the case of surface-mounted components, by machines that x the components to the board with a drop of adhesive. Some components may have to be inserted by hand, but this is now uncommon as nearly all components are designed for automatic handling. Once all the components are in place, the whole board is masssoldered, so that all soldered joints are made in one fast, cheap process. There are two main mass-soldering techniques in use: wave soldering and reow soldering.

Wave soldering

Mass soldering of conventional through-hole printed circuit boards (as opposed to surface mount boards) is normally done in a wave-soldering machine. Figure 2.11 illustrates the principles of the process. PCBs loaded with components pass along a conveyor over a wave of molten solder maintained by a pump from a solder bath. The underside of the board is preheated and uxed before the board reaches the wave. As the board passes across the wave, the underside of the board is washed with molten solder. If conditions are correctly adjusted, just sufcient solder stays on the board to make good joints, with no globules or icicles of solder on the component leads that are later cut off with rotating cutters.

Reow soldering

Surface mount PCBs are normally mass-soldered by a reow process. A solder–ux paste is applied to the joints, and components are stuck to the board with adhesive. The board is then heated to melt or reow the solder by one of two methods. In vapour-phase reow, the board is passed through a tank in which an inert uorocarbon liquid is boiling, lling the tank with hot vapour. The vapour condenses on the board, giving up its latent heat of vapourization to the board and thus heating the solder joints. Precise temperature control is achieved because the board is heated to no more than the boiling point of the liquid. In infrared reow, the solder paste is heated by infrared radiation from electric elements.

Figure 2.11 Principle of wave soldering.

Design considerations

Manufacture of a board can be made easier and cheaper by good PCB design. If a board is to be wave-soldered, for example, the direction of ow of the solder should be checked before the board is designed, and tracks on the solder side of the board laid out in the direction of ow. This reduces the chance of solder forming bridges between tracks. For the same reason, ICs should be laid out with their rows of pins across the direction of ow. If a board is to be assembled with automatic component insertion equipment, setting up the machine may be easier if all axial components are laid out on a common pitch and all integrated circuits are oriented in the same direction.

Rework and repair

The ease with which wire-wrapped joints can be remade is a distinctive feature of wire-wrap technology. Soldered connections on PCBs and elsewhere may also need to be altered or remade for several reasons.

Minor faults can occur in PCB manufacturing and assembly that require some manual attention to the board. The process of correcting production faults is called rework. If a board needs attention to correct a fault that has occurred in service, the work is called repair. Similar manual techniques are applicable in either case, although the types of fault encountered may be different. Rework may involve removal of excess solder from a board, removal and replacement of an incorrect component, or addition of a component omitted during manufacture. Repair may involve reconnection of a broken wire or PCB track, replacement of failed components, and restoration of a mechanically damaged or burnt area of PCB. Some types of repair and rework on PCBs are delicate jobs for skilled craftsmen, but some limited skill at component replacement is needed by most electronics engineers doing development work on hardware.

One of the most common rework and repair operations is removal of solder to release a component from a PCB. There are two main ways of doing this, one using capillary action and one using partial vacuum. Copper braid impregnated with ux can be used to draw the solder out of a joint or hole by applying a soldering iron to the braid while in contact with the solder to be removed. Once the solder is molten, capillary action draws it into the braid and out of the joint. The end of the braid is then cut off and discarded. An alternative, less messy technique is to use a solder-sucker syringe or suction soldering iron. Here, the iron is used to melt the solder, and a sharp suck from the spring-loaded syringe or suction iron removes the molten solder from the joint. Both techniques require skill if the joint is not to be overheated, causing damage to nearby components or delamination of a PCB track. Component removal on plated-through boards can be especially difcult: when removing an IC, it is better to cut all the IC leads to release the body and then unsolder the leads one by one. It is fairly easy to pull out the barrel of a plated-through hole while removing a component lead. On a double-sided board, this may not be a problem as the new component can be soldered on both sides of the board, but if the hole

Recommendations on PCB

Recommendations on PCB design are given in IEC 60326-3

design are given in IEC 60326-3 (1991).

(1991).

Further details of repair and

Further details of repair and rework techniques can be

rework techniques can be found in BS6221-21 (2001) and

found in BS6221-21 (2001) and BS6221-25 (2000).

BS6221-25 (2000).

connects to an internal layer on a multilayer board, further repair may be impossible.

Damaged tracks on a PCB can sometimes be repaired by soldering discrete wires to the board to bridge across a damaged area, or by using self-adhesive copper strip that can be stuck to the board and soldered to the undamaged parts of the track. This is less likely to be possible on boards with very ne or closely spaced tracks. If alterations are needed to a PCB, the same methods can be used: tracks can be disconnected by cutting across in two places with a sharp knife and then carefully lifting the piece in between with the end of the blade.

Damaged areas on a PCB can be restored using epoxy resin after removal of any loose fragments or burnt areas (caused, for example, by a component being burnt out by a fault).

Case study: A temperature controller

There is no automatic method

There is no automatic method of assembling and soldering

of assembling and soldering discrete wires. Wire-wrapped

discrete wires. Wire-wrapped wiring can, however, be

wiring can, however, be connected by machine in

connected by machine in applications such as card cage

applications such as card cage backplanes.

backplanes.

This chapter has introduced several interconnection technologies including discrete wiring, printed circuits, and connectors. Modern electronic products are designed without discrete wiring as far as is possible because of the high cost of assembling and soldering wires. There are several different applications for discrete wiring in electronic products, summarized in Table 2.4, alongside alternative interconnec tion techniques that can be used for the same applications but avoid the need to solder wires by hand.

The product illustrated in Figures 2.12 and 2.13 is a temperature con troller, manufactured in several versions with different types of input and output. Figure 2.12 shows an early version of the controller designed in the early 1980s, while Figure 2.13 shows a later version. Modern designs such as the controller shown in Figure 2.14 use digital displays and pushbutton controls for setting the temperature and are also designed from the start to have no discrete wiring and to be easily assembled.

Table 2.4 Alternatives to discrete wiring

Application Alternative techniques Wiring to front-panel

controls, connectors, and displays.

Direct PCB mounting switches,

potentiometers, and connectors. PCB tted behind panel with ribbon cable interconnect to main PCB or right-angle mounting components tted direct to main PCB.

Connection to

transformers and power supplies.

PCB mounting components. Push-on

crimped wiring connectors with PCB terminals soldered into board during mass soldering.

Inter-PCB wiring

where a product consists of more than one PCB.

PCB motherboard. Board-to-board

connectors. Ribbon cables connecting to PCB sockets. Flexi-rigid boards made in one piece and folded to t into the product.

Figure 2.12 Mark 1 temperature controller.

(Product illustrated courtesy of Eurotherm Ltd.)

Figure 2.13 Mark 2 temperature controller.

(Product illustrated courtesy of Eurotherm Ltd.)

The Mark 1 version of the controller, shown in Figure 2.12, had two PCBs. The larger board included the input and output terminals at the rear edge, a transformer and power supply, temperature control circuits, and the main output circuit. The smaller board contained circuitry that varied from one version of the product to another to accommodate different types of optional second output circuit. The two PCBs were

Figure 2.14 Modern temperature controller with digital display and push-button setting. (Illustration courtesy of Eurotherm Ltd.)

connected by a wiring loom of eight colour-coded wires, making 16 joints to be soldered by hand. The wiring loom was assembled separately before soldering to the PCBs. A typical input to a controller of this type is a voltage from a thermocouple sensing the temperature to be controlled, while the output may be a thyristor circuit for controlling the power delivered to an electric furnace element. The desired temperature was set on the large wheel, and viewed through a window in the front panel. The controller is shown without its outer cover, and was typically mounted into an industrial control panel.

The Mark 2 version, shown in Figure 2.13, had no discrete wiring. It was still in production in the early 1990s, although by then it had been superceded by newer models. The smaller PCB was fabricated with edge contacts that located onto square pins staked into the main PCB. These pins were mass-soldered, with all other connections on the main board made by wave soldering. Four connections at the rear of the smaller board were made with a wiring connector pushed onto pins in the main PCB. There was thus no hand soldering needed, and the wiring loom had been eliminated. The connections formerly made by the wiring loom were then made at low cost by wave soldering with no possibility of error. This meant that there was no need for the connections to be checked: a visual inspection of the solder joint quality was sufcient. The change in wiring design together with other design changes meant that the Mark 2 version could be assembled by only ve people compared to seven for the Mark 1. Those ve people assembled

800 units per week. The modern design shown in Figure 2.14 has eliminated the mechanical temperature-setting wheel, and uses daughter boards to permit a range of options to be assembled simply by sliding in the appropriate daughter boards.

The objective of modern interconnection techniques in electronic product design is to minimize total manufacturing costs, and this involves a balance among material and parts costs, assembly costs, and capital costs for assembly equipment. Other factors such as parts inventories and the amount of work in progress on a production line are also important. Many of these aspects of electronic production depend on product design, and design for production is an important objective in modern electronic engineering design.

Summary

Interconnection technology is of fundamental importance in the design and manufacture of electronic products. Electrical joints in electronic systems are most often made by soldering, although for some purposes wire-wrapping or crimping is used. Soldered joints were made in the past with a low-melting-point tin–lead alloy that dissolves into the surfaces of the metals being joined. From 2006 onwards, tin–lead solder has been replaced in most applications by lead-free solders to avoid the use of lead, which is toxic. Cleanliness, ux, and sufcient heat are essential requirements for a good soldered joint. Different solders are available for different applications: the type of solder to be used should be selected with care. Wire-wrap jointing is an alternative to soldering for some applications and has advantages in ease of alteration for prototype work. The integrity of a wire-wrap joint depends on good metalto-metal contact brought about by the pressure of the wire on the sharp corners of the terminal pin.

Discrete wiring and cabling are best avoided where possible, but are still essential in many applications. Wiring is used mainly for making internal connections, and cabling for external connections. The selection of wires and cables for a particular application requires care and attention to the characteristics of the wire or cable. Connectors are used where a connection must be easily disconnected and reconnected, and for connecting cables to equipment. The selection of a connector requires the same care as the selection of any other electronic component.

The principal interconnection technology described in this chapter has been the printed circuit, manufactured by chemical etching using photographic techniques to transfer a conductor pattern from a lm to a board. Single-sided PCBs are the simplest and cheapest type. Double-sided and multilayer boards with connections between sides and to internal layers by plated-through holes are manufactured by more elaborate processes and are therefore more expensive. Printed circuit designs are usually prepared using CAD software. PCBs can be masssoldered using a wave-soldering machine or by vapour-phase reow.

Alterations and repairs to PCBs require skilled techniques and some special tools for removal of solder from joints.

Problems

1 2

4.95 V

5 V

50 mm 50 mm 50 mm 50 mm

Power supply connections this end

2.1 Calculate the minimum cross-sectional area of copper wire required to carry a current of 10 A with a voltage drop of less than 50 mV per metre of conductor. The resistivity of copper is 1.7 × 10–8 Ωm. Ignore any heating effects.

2.2 A printed-circuit backplane is to be designed for an industrial con trol system. Four circuit boards are to be plugged into the backplane, each drawing 0.5 A at 5 V and spaced 50 mm apart. The 5 V and 0 V power-supply connections are at one end of the back plane 50 mm from the rst board. What width of track is needed in 35 µm copper if the voltage at the far end of the backplane is to be no less than 4.95 V? The resistivity of copper is 1.7 × 10–8 Ωm. Ignore any heating effects.

Integrated circuits

Objectives

To emphasize the importance of integrated-circuit technology in

□

electronic engineering. To describe the manufacturing processes used in integrated-circuit

□

production from preparation of raw material to packaging and testing. To introduce handling precautions for semiconductor devices.

□

To outline the range of custom integrated-circuit design methods

□

from full-custom through gate arrays to programmable logic. To discuss briey the importance and pitfalls of second sourcing.

□

The widespread application of modern electronic products is made possible, above all else, by the technology of the monolithic integrated circuit (IC), rst developed during the 1960s. Integrated circuits have had a profound effect on electronic product design and utility in at least four ways. First, IC technology is inherently a mass-production technology, producing low-cost devices. Second, ICs are miniaturized circuits typically about 10 mm across or less. This makes it possible to manufacture small electronic products such as wristwatches, pocket calculators, portable phones, and MP3 players, which combine a high level of functionality with small size. Third, ICs are reliable: good-quality pocket calculators, for example, are more likely to fail through wear and tear on their keys than through the failure of their ICs. Compare this with the earliest computers built from thermionic valves, which required several failed valves to be replaced per day. Finally, the advantages just described created pressure on IC designers to reduce the power consumption of their circuits, resulting in electronic products that can be powered by small primary batteries.

The same technology that produces the IC is also used to manufacture modern discrete semiconductor devices including transistors and thyristors. Externally, these components are simple, with three or four terminals and comparatively straightforward function compared to an IC. Internally, their detailed structure can be quite elaborate, particularly for high-power devices.

From the mid-1960s onwards, the complexity of ICs as measured, for example, by the number of transistors on one chip grew exponentially. Figure 3.1 illustrates this by plotting the complexity of one man ufacturer’s microprocessor chips on a logarithmic scale against year of introduction. As can be seen, the number of transistors per chip doubled roughly every 18 months. Towards the mid-1980s, however, this trend began to slow down as IC manufacturers encountered the practical limits of then current IC fabrication technology. There were two contributions to the rapid doubling of IC complexity. One was a progressive reduction in the size of individual elements on the chip, and the other

Monolithic means fabricated

from one piece of (literally)

from one piece of (literally) stone. The development of

stone. The development of monolithic IC technology was

monolithic IC technology was driven by the demands of the

driven by the demands of the USA’s Apollo space programme

USA’s Apollo space programme to land astronauts on the Moon

to land astronauts on the Moon in the 1960s.

in the 1960s.

More detailed treatments of

More detailed treatments of some of the material in this

some of the material in this chapter are given by Morant

chapter are given by Morant (1990) and Sparkes (1994).

(1990) and Sparkes (1994).

The relationship illustrated by

The relationship illustrated by Figure 3.1 is known as Moore’s

Figure 3.1 is known as Moore’s law after Gordon Moore, a

law after Gordon Moore, a

founder of Intel Corporation, who

founder of Intel Corporation, who rst predicted exponential growth

rst predicted exponential growth in integrated circuit complexity in

in integrated circuit complexity in a 1965 article.

a 1965 article.

1970 1975 1980 1985 1990 1995 2000 2005

Number of transistors

Year of introduction

4004

8008

8080

8086

80286

80386

80486

Pentium

Pentium II

Pentium III

Pentium 4

Itanium

Itanium II

Core Duo

Figure 3.1 Growth of IC complexity: number of transistors on microprocessor chips plotted on a logarithmic scale against the year of introduction. (Source: Intel Corporation.)

was a gradual increase in the area of chips as improvements in process quality reduced the probability of defects on a chip. Innovation in circuit design also had some inuence on chip complexity during the earlier years. The reduction in size of IC elements, such as transistors, caused several problems that have tended to slow down the doubling of IC complexity. One was that small circuit elements required narrower lines to be dened on the chip during fabrication but the optical photolithographic methods and materials used are limited to linewidths of no less than about 1 µm. Modern processes use deep ultraviolet light (of shorter wavelength) to dene linewidths of around 50 nm. The elec-

Devices may be scaled down

Devices may be scaled down to a few µm using simple

to a few µm using simple calculations, but below this size

calculations, but below this size other effects become signicant

other effects become signicant and new circuit techniques are

and new circuit techniques are needed.

needed.

tronic behaviour of circuit elements or devices is also affected by size reduction or scaling, so that changes in circuit design are needed as devices become smaller. The power density within a chip also increases as more and more devices are packed onto a chip and removal of heat becomes a problem.

The earliest ICs contained only a few tens of transistors or fewer

than ten logic gates, and are now known as small-scale integration or

LSI and VLSI ICs tend to be

LSI and VLSI ICs tend to be digital or logic circuits rather than

digital or logic circuits rather than analogue circuits because, with

analogue circuits because, with only a few exceptions, analogue

only a few exceptions, analogue circuits of high complexity are

circuits of high complexity are not needed. Because of the

not needed. Because of the complexity of modern ICs, they

complexity of modern ICs, they can only be designed with the

can only be designed with the aid of computer software, usually

aid of computer software, usually known as computer-aided design

known as computer-aided design (CAD).

(CAD).

SSI circuits. Later circuits, before the development of microprocessors, were known as medium-scale integration or MSI circuits and contained up to 100 logic gates or several hundred transistors. Larger chips such as 8-bit microprocessors are known as large-scale integration or LSI circuits. Chips with more than 10,000 gates, such as 16-bit and 32-bit microprocessors, were referred to as very large-scale integration or VLSI circuits. The largest chips in production in 2006 have about 200 million transistors, and the term VLSI now seems rather archaic. The possibility of fabricating a monolithic circuit covering a whole wafer, from which chips are normally separated, has been studied and given the name WSI, for wafer-scale integration, but it has never been realized commercially.

Review of semiconductor theory

Pure (intrinsic) silicon has an electrical resistivity of 2.3 × 10–3 Ωm at room temperature, which is about 11 orders of magnitude greater than that of a good conductor such as copper at 1.7 × 10–8 Ωm and at least 12 orders of magnitude less than that of a good insulator such as polystyrene at about 1015 to 1019 Ωm. Silicon is a tetravalent element in Group IV of the Periodic Table and has four electrons available for chemical bonding. Crystalline silicon is covalently bonded and has a tetrahedral lattice structure like that of diamond. The properties of silicon devices and integrated circuits depend on the ability of a silicon lattice to incorporate trivalent and pentavalent dopant atoms from Groups III and V of the Periodic Table. Group III dopants, such as boron, can contribute only three bonding electrons to the silicon lattice, leaving a vacant covalent bond or hole, which behaves as if it were a positive mobile charger carrier. Group III dopants are known as acceptors because the vacant bond accepts an electron from a neighbouring atom in the lattice. Silicon with an excess of acceptors is known as p-type silicon. Group V dopants, such as phosphorus and arsenic, have ve bonding electrons, of which only four can contribute to bonding with the surrounding silicon atoms. The fth electron is thus relatively free to move around the lattice and contribute to conduction. Group V dopants are known as donors because of this addition of a mobile electron to the lat- tice. Silicon with an excess of donors is known as

n-type silicon. Doped silicon is known as an extrinsic semiconductor because its electrical properties are dominated by the effect of the dopant atoms.

A typical IC chip is about 5 to 15 mm across by 0.75 to 1 mm thick, and is cut from a wafer of semiconductor material containing hundreds of chips (Figure 3.2). Components within the integrated circuit such as transistors, diodes, and, to a limited extent, resistors and capacitors are built from regions of extrinsic semiconductor formed in the top 10 to 20 µm of the chip by incorporation of dopant atoms into the silicon crystal. Figure 3.3 shows how a p–n junction diode and an n–p–n junction tran sistor can be formed. The bulk of the chip shown has been doped to form a p-type substrate. An electrical connection to the substrate is made at the base of the chip for reasons that will become clear in a moment. Regions of n-type semiconductor have been formed by incorporating sufcient dopant atoms to create an excess of mobile electrons over the holes created by the p-type dopants. Within the n-type regions, further p-type and n-type regions have been formed by adding further dopants. The diode and transistor structures can be seen from the diagram to correspond with the conceptual structures shown. There are also some other p–n junctions present in the integrated circuit, formed by the ntype collector region of the transistor and the p-type substrate, and by the n-type region around the diode and the substrate. To maintain electrical isolation between the diode and the transistor, the substrate must be tied to the most negative potential in the external circuit so that the collectorto-substrate junction is reverse-biased. The surface of the chip is covered with silicon dioxide, which is an electrical insulator, and connections to the terminals of the diode and transistor are formed from a metallization layer, or layers, which is deposited over the silicon dioxide and makes contact with the underlying devices through holes, or windows, in the

The discussion in this chapter

The discussion in this chapter is mainly in terms of silicon

is mainly in terms of silicon because this element accounts

because this element accounts for the majority of ICs produced

for the majority of ICs produced today. A detailed discussion of

today. A detailed discussion of semiconductor physics and the

semiconductor physics and the theory of semiconductor devices

theory of semiconductor devices is outside the scope of this book.

is outside the scope of this book. Till and Luxon (1982) have given

Till and Luxon (1982) have given a more detailed treatment of

a more detailed treatment of most topics in this chapter.

most topics in this chapter.

Two atoms bonded covalently

Two atoms bonded covalently each contribute one electron to

each contribute one electron to the bond. Bonding is discussed

the bond. Bonding is discussed by Anderson et al. (1990).

by Anderson et al. (1990).

The majority of ICs fabricated

The majority of ICs fabricated today contain complementary

today contain complementary metal-oxide semiconductor

metal-oxide semiconductor (CMOS) logic, because CMOS

(CMOS) logic, because CMOS offers low power and high circuit

offers low power and high circuit density. Despite the low power

density. Despite the low power dissipation of individual gates,

dissipation of individual gates, a chip with millions of gates can

a chip with millions of gates can dissipate signicant power when

dissipate signicant power when clocked at a high frequency.

clocked at a high frequency. CMOS is a logic technology

CMOS is a logic technology using n-channel and p-channel

using n-channel and p-channel metal-oxide-semiconductor eld-

metal-oxide-semiconductor eldeffect transistors (MOSFETs)

effect transistors (MOSFETs)

on the same chip. Morant

on the same chip. Morant (1990) has given structures for

(1990) has given structures for MOSFETs.

MOSFETs.

Resistors and capacitors occupy

Resistors and capacitors occupy more chip area than transistors.

more chip area than transistors. IC designers, therefore, tend

IC designers, therefore, tend to use circuit techniques that

to use circuit techniques that avoid the need for passive

avoid the need for passive components. Ritchie (1998)

components. Ritchie (1998) has given examples of these

has given examples of these techniques.

techniques.

In practice, diodes are often

In practice, diodes are often formed from a bipolar transistor

formed from a bipolar transistor with the base and collector

with the base and collector connected to form a diode-

connected to form a diodeconnected transistor, which

connected transistor, which requires less chip area by

requires less chip area by a factor of β than a diode of

a factor of β than a diode of equivalent current rating. Ritchie

equivalent current rating. Ritchie (1998) has given examples of IC

(1998) has given examples of IC designs using this technique.

designs using this technique.

Figure 3.2 Monolithic integrated circuits: a typical modern wafer (300 mm)

Anode

a c

p n

Substrate contact

b e

Silicon dioxide

Aluminium

EmitterCathode Collector

Base

(a)

(b)

(c)

−1 mm

<20 µm

n n np

containing hundreds of chips. (Courtesy of Texas Instruments.)

Figure 3.3 Monolithic integrated circuit structure showing possible structures for a p–n junction diode and a n–p–n junction transistor: (a) circuit symbols, (b) conceptual structures, and (c) integrated-circuit realization.

insulator. Traditionally, aluminium is used for metallization, but copper is now used for some high-performance chips. External connections to the chip are made by ne gold wires connected to bonding pads around the edge of the chip, using techniques described later.

There are other methods of isolating devices on a chip from each other and many possible geometrical arrangements of n-type and p-type regions, silicon dioxide layers, and metal layers. The geometry of IC devices and the sequence of fabrication steps required to create them are known collectively as a process. Development of a process can be a lengthy and expensive undertaking, and precise details of proprietary processes are not released by manufacturers. There are, however, only a few major steps in a fabrication process, which are combined and repeated in various sequences (up to 300 steps) to build up the desired IC structure. These process steps consist of doping processes to incorporate dopant atoms into the silicon lattice to form regions of n-type and p-type semiconductors of dened depth, lateral geometry, and dopant concentration; crystal growth to build up new layers of silicon; oxide growth to form silicon dioxide layers; lithography to transfer images to the silicon; and etching to remove regions of silicon or silicon dioxide. The entire volume of an IC chip must be a single crystal of silicon, so the dopant atoms must either be incorporated into an existing crystal lattice or be included as the lattice grows.

Integrated-circuit fabrication

Morant (1990) has given

Morant (1990) has given examples of the sequence of

examples of the sequence of steps required to fabricate ICs

steps required to fabricate ICs using a typical CMOS process.

using a typical CMOS process.

IC fabrication is a mass-production activity. Dozens of wafers are processed simultaneously through many of the fabrication processes, and each wafer contains hundreds of chips. Much of the processing is carried out at high temperatures of up to 1200°C, although there is a trend towards lower-temperature processing for the fabrication of ICs with small-geometry devices because of undesired dopant diffusion. The dimensions of individual device features on an IC can be as little as 50 nm in current production ICs, and smaller in research designs. Devices are approximately 0.5 µm square in production chips.

Worked Example 3.1

Estimate the size of a storage cell on a type 27128 128 k bit electronically programmable read-only memory (EPROM) chip that measures 4 mm by 5 mm (1 k = 210 = 1024).

Solution

The chip area is 20 mm2. Dividing this by 128 × 1024, the area of a storage cell is 153 µm2. Assuming the cells to be square, they are about 12 µm across.

Very high standards of cleanliness are required in an IC fabrication facility (Figure 3.4), because dust particles are much larger than the dimensions of device features and the image of a dust particle reproduced photolithographically on an IC may obliterate several devices, rendering the chip useless. A very high standard of air ltration and operator cleanliness is therefore essential. Operators must wear special

Clean room air quality is dened

Clean room air quality is dened by ISO 14644 in terms of the

by ISO 14644 in terms of the number of particles of diameters

number of particles of diameters greater than 0.1, 0.2, 0.3, 0.5,

greater than 0.1, 0.2, 0.3, 0.5, 1, and 5 µm per cubic metre of

1, and 5 µm per cubic metre of air. The highest grade of clean

air. The highest grade of clean room (Class 1) must have no

room (Class 1) must have no more than 10 particles greater

more than 10 particles greater than 0.1 µm in diameter, and no

than 0.1 µm in diameter, and no more than 2 greater than 0.2 µm,

more than 2 greater than 0.2 µm, per cubic metre of air. Humans

per cubic metre of air. Humans typically shed 250,000 0.5 µm

typically shed 250,000 0.5 µm particles per minute even at rest

particles per minute even at rest (over the whole body) and over

(over the whole body) and over 1,000,000 when moving about.

1,000,000 when moving about.

Trichlorosilane is a colourless

Single-crystal

silicon

75–400 mm

Molten silicon

Alumina crucible

e Czochralski process

(the apparatus is enclosed

in an argon atmosphere).

Single-crystal

silicon

75–400 mm

Molten silicon

Alumina crucible

e Czochralski process

(the apparatus is enclosed

in an argon atmosphere).

Trichlorosilane is a colourless liquid with a boiling point of

liquid with a boiling point of 33°C at a pressure of 758 mm

33°C at a pressure of 758 mm of mercury. The distillation is

of mercury. The distillation is carried out at reduced pressure

carried out at reduced pressure because trichlorosilane is

because trichlorosilane is unstable and cannot be distilled

unstable and cannot be distilled at atmospheric pressure.

at atmospheric pressure.

Figure 3.4 Part of an integrated circuit fabrication facility.

(Courtesy of Texas Instruments.)

hooded overalls to reduce contamination with skin particles. Smoking is not permitted, even during breaks outside the clean room, because of the quantity of particles exhaled for some time after smoking. Cosmetics are also banned because of the danger of particulate contamination. There is a trend towards automation in modern IC fabrication facilities to reduce the number of operators required and therefore improve cleanliness.

IC fabrication facilities are very expensive to set up and are expensive to operate, so they must produce large quantities of ICs, or ICs of high value, to be economically worthwhile.

Preparation of silicon wafers

Silicon occurs naturally as the second most abundant element after oxygen in the Earth’s crust, making up about 25% by weight of the crust in the form of silicates and silica (SiO2). Silica is found in a variety of forms including int and quartz. Naturally occurring high-purity silica sand is the starting material for the preparation of device-grade silicon for IC manufacture. Silica is reduced to silicon by reaction with carbon in an electric furnace giving silicon and carbon monoxide:

SiO

+ 2C → Si + 2CO

The impure silicon is then converted to trichlorosilane (SiHCl3) by reaction with hydrochloric acid:

Si + 3HCl

→ SiHCl3 + H

and the trichlorosilane is then puried by distillation and converted back to polycrystalline silicon by reaction with hydrogen. For devicegrade silicon, typical residual impurity concentrations for Group III and V elements are less than one part in 109. Single-crystal silicon is required for IC fabrication. There are two methods of fabricating single-crystal ingots of silicon: the oat-zone process and the Czochralski process. In the oat-zone process, a polycrystalline ingot is converted to

single-crystal form by heating a small zone using radio frequency heat-

4 8 10

. ×

4 8 10

. ×

Buried layer

Epitaxial

layer

Original substrate thickness

Buried layer

Epitaxial

layer

Original substrate thickness

ing. The molten zone is passed up the ingot, and single-crystal silicon forms behind the molten zone. The oat-zone process can also be used as a renement or purication technique because impurities tend to remain in the molten zone. Most silicon for IC manufacture is made by the Czochralski process, in which a single crystal ingot is formed by slowly withdrawing a rotating seed crystal from a crucible of molten puried silicon. The crystal orientation of the silicon is determined by the orientation of the seed crystal and is carefully chosen and indicated by grinding a at along one side of the ingot. The crystal orientation of the wafers cut from the ingot is important when the wafers are scribed and broken into chips, as cleavage occurs more easily along some crystal planes.

Ingots of up to 400 mm diameter can be grown. They are sawn into thin discs or wafers with a diamond saw. Small wafers of 75 to 100 mm diameter are about 0.5 to 1 mm thick. Larger sizes have to be thicker to prevent warping during processing. The surface of a sawn wafer is rough and damaged by the sawing process. The damaged layer is removed by lapping, and the wafer is then chemically etched to leave an optically smooth mirror nish. The processed wafers are inspected for atness because later processing depends on the projection of images onto the wafer surface, and any signicant deviation from atness will cause loss of denition in the image transferred to the wafer.

Epitaxial growth

Some semiconductor fabrication processes require the addition of extra layers of silicon on the surface of the wafer. Added silicon is known as an epitaxial layer, and it must have the same crystal structure and orientation as the wafer itself and be grown as an extension of the wafer so that no crystal boundary exists between the original surface and the new layer. An epitaxial layer might be grown after creation of a doped region in the original substrate to form a buried layer, or to form regions of different doping type or concentration to the substrate. Epitaxial layers are typically less than 20 µm thick.

Till and Luxon (1982) discussed

Till and Luxon (1982) discussed the importance of crystal

the importance of crystal orientation. For a general

orientation. For a general introduction to crystalline

introduction to crystalline materials and crystal structure,

materials and crystal structure, see Anderson et al. (1990).

see Anderson et al. (1990).

Lapping is a surface-nishing

Lapping is a surface-nishing process using a ne abrasive

process using a ne abrasive paste.

paste.

Epitaxial silicon can be grown

Epitaxial silicon can be grown on an insulating substrate with

on an insulating substrate with a suitable crystal structure. An

a suitable crystal structure. An important example is sapphire

important example is sapphire (Al2O3) used in the silicon-on-

(Al2O3) used in the silicon-onsapphire (SOS) process for

sapphire (SOS) process for CMOS circuits.

CMOS circuits.

Worked Example 3.2

How many silicon atoms make up a 20 µm layer?

Solution

A crude approximation to within a factor of 2 or 3 can be calculated from the density and atomic weight of silicon, and the proton– neutron mass. (The mass of electrons in an atom is negligible.) The density of silicon is 2300 kg m–3, the atomic weight is close to 28, and the masses of the proton and neutron are about 1.7 × 10 con atom has a mass, therefore, of (28) 1.7 × 10

–27

or 4.8 × 10

The number of atoms in a cubic metre of silicon is 2300/4.8 × 10

4.8 × 1028. Assuming the atoms to be packed cubically (which they are not), there would be

atoms along each edge of the cube. The number of atoms across the thickness of a 20 µm epitaxial layer is thus compared to the enormous numbers of atoms in bulk material.

× 20 µm or about 70,000, a surprisingly small number

-27

kg. A sili-

–26

kg.

–26

Silicon atoms are added to the substrate surface in a reactor at a tem-

Regions to be doped

Silicon dioxide masking layer

Silicon

Regions to be doped

Silicon dioxide masking layer

Silicon

perature of 800 to 1100°C. The required silicon atoms can be produced by the pyrolytic decomposition of silane (SiH4) at around 1000°C.

This is the overall reaction. The

This is the overall reaction. The reaction

reaction

2SiCl2 → Si + SiCl

takes place on the silicon

takes place on the silicon substrate after production of

substrate after production of SiCl2 in the gas stream by the

SiCl2 in the gas stream by the reaction

reaction

SiCl4 + H2 ↔ SiCl2 + 2HCl

The thicknesses required for

The thicknesses required for these two purposes are very

these two purposes are very different: typically 20 nm

different: typically 20 nm (0.02 µm) for an IGFET gate and

(0.02 µm) for an IGFET gate and up to 1 µm for an isolation layer.

up to 1 µm for an isolation layer.

Silicon nitride (SiN3) is another

Silicon nitride (SiN3) is another possible dielectric used in IC

possible dielectric used in IC fabrication.

fabrication.

Lithography means, literally,

“writing on stone.” The

“writing on stone.” The techniques discussed here are

techniques discussed here are in principle the same as those

in principle the same as those used for PCB manufacture and

used for PCB manufacture and described in Chapter 2. The

described in Chapter 2. The scale of lithographic images

scale of lithographic images in IC fabrication is, however,

in IC fabrication is, however, about 1000 times smaller than

about 1000 times smaller than those used in PCB manufacture.

those used in PCB manufacture. (Typical linewidths are 0.5 µm

(Typical linewidths are 0.5 µm and 0.5 mm respectively.)

and 0.5 mm respectively.)

SiH

→ Si + 2H

or by reduction of silicon tetrachloride (SiCl4) by hydrogen

2SiCl

+ 2H2 → Si + SiCl4 + 4HCl.

Dopants can be included as an integral part of an epitaxial layer by adding traces of gases such as diborane (B2H6), phosphine (PH3), and arsine (AsH3) to the gas ow through the reactor. Careful control of the gas concentrations is essential if the number of crystal defects in the epitaxial layer is to be minimized.

Oxide growth

One of the most frequent steps in many IC fabrication processes is the formation of a layer of silicon dioxide (SiO2) on the wafer surface. Silicon dioxide is an excellent dielectric, and it can be used as an insulating layer within a device such as an insulated-gate eld-effect transistor (IGFET, or MOSFET) or to isolate an epitaxially grown region of semiconductor from the substrate. A nal layer of silicon dioxide (apart from the metallization layer, which is described later) can passivate the surface of an IC to protect it from atmospheric contaminants. The most signicant application of silicon dioxide in IC fabrication is as a masking layer that is etched to dene regions to be doped.

A silicon dioxide layer can be produced by a chemical reaction between the wafer surface and either oxygen or steam, or by a deposition process similar to epitaxial growth. Thermal oxidation requires a temperature of 800–1250°C controlled to within ±0.5°C or better together with careful control of the oxygen concentration (by using, for example, oxygen–nitrogen mixtures) and the time of processing to within 5 to 20 seconds. Because oxidation is a reaction between oxygen and the silicon surface, the reaction rate reduces as the thickness of oxide increases. Growth of a 1 µm isolation layer can take over an hour, whereas the formation of an IGFET gate layer may take only a few minutes. The oxide layer forms at the expense of the underlying silicon because silicon atoms from the wafer react to form the oxide. The thickness of silicon converted to oxide is about 30% of the nal thickness of the oxide layer.

Thermal oxidation is carried out in quartz furnace tubes slightly larger than the wafer diameter and up to 2 m long. The wafers are held vertically in quartz boats or carriers and are processed in large batches of perhaps 100 wafers at a time.

Lithography

The various regions of extrinsic semiconductor, oxide, and metallization making up the devices and interconnections on a chip have to be dened in the form of an image on the wafer surface. Techniques to do this are known as lithography.

Wafer

Photoresist

Glass mask

Chromium

UV light

Exposed regions of

resist

Wafer

After development

Figure 3.5 Principle of photolithography.

The earliest technique used in IC fabrication, and still very important today, is photolithography. Figure 3.5 illustrates the principle. The wafer is coated with a layer of photoresist a few micrometres thick. The resist is applied to the wafer as a drop of liquid while the wafer is spinning at high speed, ensuring that the resist is evenly distributed. The wafer is then gently baked to drive off the resist solvent. Resists are polymeric materials sensitive to ultraviolet (UV) light. Exposure to UV radiation can either cause a polymerization reaction or depolymerize the resist depending on whether a negative or positive image is required. The earliest resists for IC fabrication were of the negative image type, but for modern IC fabrication, positive resists are used because of their superior denition. After exposure, the resist is developed in a chemical solution that dissolves the unpolymerized regions of resist, leaving selected regions of the wafer coated with a tough polymer to resist chemical etchants or ion beams.

The image pattern to be transferred to the resist is dened by a photomask (or just “mask”), or reticle. These are thin quartz plates from 1.5 to 3 mm thick and originally as large as the wafer. For modern wafers of 200 to 400 mm diameter, the reticle is much smaller, and it is moved in steps (by a stepper machine) across the wafer to expose the whole wafer in a series of exposures. The image is dened on the mask or reticle by a chromium layer, which is generated lithographically from the IC design, either by photographic reduction from artwork or by the electron-beam lithography technique (described below). The earliest ICs were fabricated by contact printing, in which the chromium side of the mask touched the resist-coated wafer. The obvious problem with contact printing was damage to the mask, and to overcome this, projection printing systems were developed so that the mask could be kept away from the wafer surface. Projection printing is essential when using steppers, and a reduction lens is used so that the pattern on the chip is several times smaller than that on the mask. Accurate alignment or

A plasma is a fully ionized gas

Wafer

Undercutting

Resist

Wafer

Undercutting

Resist

A plasma is a fully ionized gas consisting of electrons and

consisting of electrons and atomic nuclei. The term is used

atomic nuclei. The term is used here to refer to a partly ionized

here to refer to a partly ionized gas consisting of electrons, ions,

gas consisting of electrons, ions, and neutral atoms.

and neutral atoms.

registration is essential, and both mask and wafer must have alignment marks to facilitate this. All the masks in a set for fabricating a particular IC design must of course be accurately made so that registration of the image dened by one mask with all others in the set is achieved.

Photolithography using ultraviolet light is limited to linewidths of no less than about 1 µm because of diffraction effects at line edges. Shortwavelength (deep) UV is used for photolithography of most modern ICs with linewidths down to 50 nm or so. An important technique that also overcomes the resolution problem is electron-beam lithography. An electron beam of around 0.2 µm diameter or less is directed onto a resist-coated surface (either a mask or a wafer) on a high-precision x–y coordinate table. The resist must be an electron-beam resist, not an optical resist, and the x–y table must be accurate to within fractions of a micrometre. The process is slow: full exposure of even a small 75 mm wafer can take more than an hour. Electron-beam lithography has the very signicant advantage for mask fabrication of direct transfer of design information from a CAD system to the mask with no optical reduction process. It also nds application for fabrication of prototype chip designs without the expense of making a mask set, using directwrite-on-wafer imaging.

Etching

Etching is the removal of unwanted regions of material from a wafer.

A typical example is cutting of holes or windows in a silicon dioxide layer prior to dopant diffusion or implantation into the regions under the windows. The regions to be etched are dened by the pattern in a resist created by a lithographic process as described in the previous section. There are two important etching methods used in IC fabrication: wet etching and plasma etching.

Wet etching was the earliest technique and is still used in production. Wafers to be etched are immersed in an acid bath and agitated to ensure even etching. Silicon is etched with a nitric acid–hydrouoric acid mixture, and silicon dioxide with a hydrouoric acid–ammonium uoride solution. The amount of material removed is dependent on temperature and immersion time.

Plasma etching is a dry technique in which reactive gaseous atoms react with the exposed regions of the wafer to form gaseous reaction products that are removed by a vacuum pump. The reactive atoms are generated by breakdown of molecules in a gas heated by radio frequency electromagnetic energy.

Both wet and dry etching have the disadvantage that the etchant undercuts the resist as sketched in the margin. This effect has to be allowed for in the design of an IC layout, and it limits the packing density that can be achieved.

Diffusion and ion implantation

In order to produce regions of n-type and p-type semiconductor within a wafer, dopant atoms must be introduced into the crystal structure. Dopants can be incorporated into epitaxial layers during deposition as described earlier. If dopants are to be incorporated into an existing crystal, they

can be introduced by solid diffusion or ion implantation. Both techniques

Selector

magnet

Accelerating

voltage

Focusing coil

Rotating support

Wafer

−

Ion

source

Selector

magnet

Accelerating

voltage

Focusing coil

Rotating support

Wafer

−

Ion

source

require an oxide masking layer to dene the regions to be doped.

Diffusion was the earliest process used for doping a wafer and takes place in two stages: predeposition and the diffusion process itself. The dopant material can be deposited on the wafer by spin coating with a liquid or by deposition from a gas in a diffusion furnace. After deposition the dopant atoms are still concentrated near the wafer surface. The second stage of the diffusion process distributes the dopant atoms to the required depth in the wafer by heating the wafer to around 1000°C in a diffusion furnace, of identical construction to the furnace described earlier for oxide growth. Separate furnaces are essential for diffusion and oxide growth, otherwise the oxide furnace will become contaminated with dopants. During diffusion, the windows in the diffusion-masking oxide layer must be sealed to prevent the dopants from diffusing out of the wafer. This can be done by regrowing oxide over the windows by using an oxidizing atmosphere in the diffusion furnace. Some predeposition processes produce a surface glassy layer that serves the same purpose. Careful control of diffusion time and temperature ensures that the dopant atoms diffuse to the desired depth, which is typically between

0.3 and several micrometres. Diffusion also occurs laterally, of course, and this must be allowed for in the process design rules.

Ion implantation is a more recent method of doping in which the dopant is red at the wafer as an ion beam inside a vacuum chamber. The ion energy is typically between 25 and 200 keV, and can be precisely controlled so that the ions penetrate the wafer surface to a controlled depth. The total quantity of dopant introduced into the wafer can also be controlled to within ±10% or better. Typical ion doses are between 1015 and 1020 ions m–2. The ion beam is broad compared to the line dimensions on the wafer, but is narrow compared to the wafer dimensions, so that the beam has to be scanned and the wafer rotated to ensure even exposure to the beam.

Implanted ions are not bonded into the silicon lattice: they occupy interstitial sites among the lattice atoms. The silicon lattice itself is also disrupted by the implanted ions as they lose energy by collision with the lattice atoms. After ion implantation, therefore, wafers must be heattreated to repair the crystalline structure and activate the implanted dopant by incorporating the dopant atoms into the silicon lattice. Dopant atoms that are bonded into the lattice in place of a silicon atom are said to occupy substitutional sites. The heat-treatment process is known as annealing and is carried out at similar temperatures and for similar times as thermal oxidation and diffusion. During annealing, of course, diffusion occurs and the implanted dopants migrate through the lattice to some extent.

Metallization

The nal stage of wafer processing is deposition of a metal interconnect to connect the individual devices on the chips and to form bonding pads for connection to the external circuit. Four or more layers of metal can be used, separated by dielectric with different conductor patterns on each layer. Windows are etched in the dielectric layers to make contact

Design rules specify the

Design rules specify the limitations of a fabrication

limitations of a fabrication process, including device sizes

process, including device sizes and minimum separations

and minimum separations between doped regions. Modern

between doped regions. Modern IC CAD software includes

IC CAD software includes programs called design-rule

programs called design-rule checkers that verify that all

checkers that verify that all aspects of a design comply with

aspects of a design comply with the process design rules.

the process design rules.

Some interconnections may be

Some interconnections may be made by diffused regions or by

made by diffused regions or by deposited polycrystalline silicon

deposited polycrystalline silicon layers.

layers.

A metal–semiconductor junction

A metal–semiconductor junction is known as a Schottky junction.

is known as a Schottky junction. Schottky transistors are used

Schottky transistors are used in Schottky and low-power

in Schottky and low-power Schottky transistor–transistor

Schottky transistor–transistor logic (LS TTL) circuits.

logic (LS TTL) circuits.

with the separate devices. A layer of metal, usually aluminium, sometimes with a small amount of silicon, is then deposited on the wafer, usually by vacuum deposition from vapour in a vacuum chamber. After deposition, the wafer is sintered at around 400°C. This improves the electrical connection between the deposited metal and the silicon. Modern high-speed ICs often have copper interconnects rather than aluminium. The better conductivity of copper allows faster charging of on-chip capacitances and therefore shorter rise and fall times on logic signals, which permits faster clocking.

In some devices, such as Schottky diodes and transistors, the metallization forms part of the device as well as being an interconnect. This is also true in metal-gate MOS (metal-oxide semiconductor) devices, where the gate electrode is fabricated as part of the metallization layer.

Testing, dicing, and bonding

Testing large digital circuits is

Testing large digital circuits is a nontrivial operation. Wilkins

a nontrivial operation. Wilkins (1990) has given an introduction

(1990) has given an introduction to the problem and discusses

to the problem and discusses some approaches to testable

some approaches to testable design.

design.

After metallization, all the individual ICs on the processed wafer have to be tested. Testing machines have needle probes that are pressed onto the bonding pads around the edge of the chip. There may be additional pads for test purposes that will not be connected externally. Not all the ICs on a wafer will work: those that fail the test are noted by the testing machine for later rejection. (This was once done with a drop of ink to mark the rejected chips, but modern machines simply note the coordinates of the rejected chips and remove them mechanically after dicing.) The percentage of chips that pass the test is known as the production yield. In the early production batches of a new chip or process, the yield may be quite low — many of the chips on a wafer are rejects.

The wafer is now scribed with a diamond scriber to separate the wafer into chips, each of which will be mounted into an IC package. An electrical connection to the back of the chip is necessary for most types of IC, and a common technique is eutectic bonding to a gold-plated header. Gold and silicon form a eutectic alloy at 370°C. The bonding pads on the chip now have to be connected to the external leads or pads of the package. This was once done by hand by an operator using micromanipulators working through an optical microscope, but is now carried out by automatic machinery with automated optical inspection. Gold wire is connected by ultrasonic welding or thermocompression bonding between the bonding pads on the IC and the package connections. The number of connections varies from eight for a 741-type operational amplier to 400 or more for a modern microprocessor. Once all the connections are in place, the package can be sealed, moulded, or otherwise completed. The packaged ICs are then tested electrically again and may also be subjected to heat, vibration, pressure, or vacuum to detect weak circuits that would otherwise fail early in life.

Chips may also be used unpackaged in hybrid microcircuits or in multichip modules (MCMs). Hybrid microcircuits consist of a combination of bare monolithic ICs with printed resistors and capacitors mounted on a ceramic substrate. Till and Luxon (1982) discussed the technology of hybrid microcircuits. Multichip modules can consist of bare monolithic ICs attached to a silicon substrate with interconnects fabricated on the substrate.

Semiconductor packaging

Figure 3.6 illustrates a selection of packaging styles used for ICs, while Table 3.1 summarizes the main characteristics of the most important package types. Hermetically sealed (airtight) metal cans were the earliest form of IC package and are still available for a few linear circuits. Power voltage regulator ICs are often packaged in metal cans with a thick base similar to a power transistor package. The dual-in-line (DIL) package has been used since the 1960s for packaging both logic and linear circuits, but it was largely superceded in the 1990s by surface mount packages, although many ICs are still available in DIL packages. The DIL package is designed for soldering into through-holes on a printed circuit board (PCB) or for insertion in a socket. Plastic DIL packages are moulded onto the metal lead frame after the chip has been bonded to the leads. Ceramic DIL packages may be more reliable than plastic packages, but they are also more expensive. They can be of either the frit-seal type, consisting of two ceramic slabs cemented together with a glassy ceramic onto a lead frame, or the side-brazed type with leads brazed onto connection pads along the sides of the package. The IC chip is mounted in a cavity in the ceramic and covered with a hermetically sealed metal lid in the case of a side-brazed package, or with the top slab of ceramic in the case of the frit-seal package. The most common

See Figure 7.1b for an illustration

See Figure 7.1b for an illustration of a power transistor package.

of a power transistor package.

Plastic packages might, for

Plastic packages might, for example, allow ingress of

example, allow ingress of moisture along the leads, leading

moisture along the leads, leading to corrosion of the leads or the

to corrosion of the leads or the IC bonding pads.

IC bonding pads.

Table 3.1 Characteristics of the principal types of integrated circuit package

(a) Through-hole mounting types, 2.54 mm lead spacing

Maximum number

Type

of leads

Features

Hermetic metal can 12 Leads on pitch circle,

almost obsolete

Plastic dual-in-line

48 Low cost

(DIL) Ceramic dual-in-line 64 High reliability Ceramic pin-grid array

(PGA)

> 400 Low board area for

number of pins

Ball-grid array (BGA) > 400 Designed for reow

soldering to boards

(b) Surface-mounting types, 1.27 mm lead or pad spacing

Maximum number

Type Plastic

of leads or pads Features

28 Low cost

small outline (SO) Plastic-leaded

124 Compact, low cost

chip carrier (PLCC) Leadless ceramic

124 High reliability

chip carrier (LCCC)

2.54 mm is equal to 0.1 inch.

2.54 mm is equal to 0.1 inch. Early packages were designed

Early packages were designed to Imperial (inch) dimensions

to Imperial (inch) dimensions except in the former USSR

except in the former USSR and Eastern Europe, where

and Eastern Europe, where metric packages were used.

metric packages were used. Some newer package types are

Some newer package types are designed to metric dimensions.

designed to metric dimensions.

(a) (b)

(e) (f)

Figure 3.6 Integrated circuit packages: (a) hermetic metal can, (b) plastic DIL, (c) plastic small outline (surface mount), (d) PLCC, (e) PGA, and (f) ceramic DIL. (Courtesy of National Semiconductor. These package depictions are for example only and should not be used to design with. For current packages and accurate dimensions, please refer to the National Semiconductor Web site at: http://www.national.com/.)

application for the frit-seal package is UV-erasable memories, where the chip is visible through a quartz window in the top slab.

When DIL packages were rst manufactured, ICs typically had 14 or 16 external connections, making a package about 20 mm × 7 mm. As LSI chips became available with 40 or more connections, the DIL package became unwieldy: a 40-pin DIL package measures about 50 mm × 16 mm, yet may house a chip about 5 mm square. To reduce the size of IC packages for large chips, the pin-grid array (PGA) was developed, with pins still on a 2.54 mm (0.1 inch) pitch but arranged in several rows all around or all over the underside of the package on a rectangular grid. PGAs are constructed in the same way as a side-brazed ceramic DIL package. The PGA can have up to 400 leads and yet be only 25 mm square. Larger PGAs with over 400 leads are used with sides of 47 mm or more.

At about the same time that pin-grid arrays were introduced, the technology of surface mounting was also being developed. Because several manufacturers were working on surface mounting simultaneously, several types of package emerged. The small-outline (SO) package is essentially a scaled-down DIL package with leads spaced at half the pitch of a DIL package and folded out at rather than projecting beneath the package. The chip-carrier package has leads spaced at the same 1.27 mm (0.05 inch) pitch as the SO package, but on all four edges. Plasticleaded chip carriers (PLCCs) are moulded in a similar way to SO packages but have J-shaped leads rolled under the package body. Leadless ceramic chip carriers (LCCs) have pads rather than leads and are similar to a side-brazed ceramic DIL package in construction. There are also quad at packs that are similar to SO packages but with leads around all four edges of a roughly square package. Lead spacings vary, but can be as little as 0.5 mm. Flat packs are usually moulded from plastic. PLCCs and LCCs can be housed in sockets. SO packages and at packs are not socketable. Ceramic chip carriers are not suitable for soldering to epoxy PCBs because of the difference in thermal expansion coefcient of the ceramic relative to epoxy, which can cause stress cracking of solder joints. Ball-grid arrays (BGAs) are packages designed for direct surface mount soldering by reow — they have small solder bumps arranged in a rectangular grid on the underside of an epoxy plate. The chip is mounted on the top surface of the plate under a plastic cover. BGAs can also be mounted in sockets. Some specialized packages are used for processor chips such as the Intel Pentium, where the heat dissipation at full power is signicant (50 W or more).

This point is discussed further in

This point is discussed further in Chapter 10.

Chapter 10.

Handling of semiconductor devices

Many types of semiconductor devices and integrated circuits are damaged fairly easily by physical or thermal shock, overheating during soldering, and, especially, electrostatic discharge. The damage caused by mishandling is often not catastrophic — the device does not fail immediately, but is weakened by the damage, and eventually fails weeks or months later after being built into a product and shipped to an end user. Careful quality control is therefore essential on an electronics production line. Production operators must be made aware of handling

Test methods are discussed in

Test methods are discussed in Chapter 10.

Chapter 10.

Germanium diodes, for example,

Germanium diodes, for example, nd application in some

nd application in some analogue circuits because of

analogue circuits because of their low forward voltage drop

their low forward voltage drop compared to silicon diodes. The

compared to silicon diodes. The 1N34A and OA90 are examples

1N34A and OA90 are examples of germanium diodes that are still

of germanium diodes that are still available.

available.

CMOS and MOS logic circuits

CMOS and MOS logic circuits have gate-protection diodes

have gate-protection diodes connected to external gate

connected to external gate electrodes to provide a path

electrodes to provide a path for static charge to leak away.

for static charge to leak away. This may not, however, prevent

This may not, however, prevent damage caused by a discharge

damage caused by a discharge into the gate terminal from an

into the gate terminal from an external source.

external source.

Symbol denoting electrostatically

Symbol denoting electrostatically sensitive devices. (Courtesy

sensitive devices. (Courtesy of the Electrostatic Discharge

of the Electrostatic Discharge Association.)

Association.)

Electrical safety is discussed in

Electrical safety is discussed in Chapter 11.

Chapter 11.

precautions and provided with the correct tools and equipment for handling semiconductors. Completed electronic products can also be tested to reveal defects such as weakened semiconductor devices before they leave the factory.

Thermal damage

Germanium semiconductors were the earliest solid-state devices, and they were easily damaged by overheating during soldering. Although silicon devices are used for nearly all applications today, a few germanium devices are still available and are used in specialized applications. Heat-absorbing pliers are recommended for holding the leads of these devices while soldering. Silicon semiconductor devices are more robust thermally, but they still require care if soldered by hand. Mass soldering is more easily controlled both in temperature and in time, to keep within the limits specied in a data sheet (usually under the heading “Absolute maximum ratings”).

Certain types of integrated circuits and discrete semiconductors are sensitive to damage by discharge of static electric charge. Not all device types are sensitive to damage (although none is completely immune). Field-effect devices with insulated gate electrodes are the most sensitive types. These include MOS and CMOS logic circuits and MOSFET transistors. Damage to the devices results from electrostatic discharge (ESD) into the gate electrodes, causing dielectric breakdown and perforation of the gate insulation. Electrostatic charge can build up on clothing, shoes, and oor coverings as a result of surfaces rubbing together. Figure 3.7 shows a typical ESD protected area (EPA) for electrostatic sensitive devices as recommended in an international standard. The workbench surface, compartment trays, oor mat, and operator’s stool are all electrically conductive. The operator is wearing electrostatically conductive overalls and special shoes, and is connected by a wrist strap to earth. Because of the danger of electric shock in an environment of earthed conductors, the electricity supply to the workbench is isolated from the mains supply by an isolation transformer and is tted with a residual current device (RCD). The earthing straps have a resistance of 0.5 to 1 M

Ω,

sufcient to conduct static charge safely to earth, but high enough to limit current ow in the presence of an electrical fault to a safe level.

Electrostatically sensitive devices are protected in storage and in transit using special packaging materials. Dual-in-line ICs, for example, are inserted into conductive foam or kept in an electrostatically conductive plastic tube. Assembled PCBs containing sensitive devices can be placed in electrostatically conductive plastic bags for shipment and storage.

Morant (1990) has discussed

Morant (1990) has discussed the IC design process in greater

the IC design process in greater detail.

detail.

Custom integrated circuits

There is a wide variety of standard integrated circuits available, both for logic and for analogue applications. If an electronic product is to be manufactured in quantity, however, it may be economically worthwhile to design a tailor-made custom IC specically for the product. These circuits are known as ASICs, for application-specic integrated circuits. A custom-designed IC is cheaper per IC, more reliable, and of smaller size and power consumption than the equivalent circuit implemented with

1 2 3 4 5 6 7 8 9

Groundable wheels Groundable surface Wrist strap tester, shall be displayed outside the EPA Footwear tester, shall be displayed outside the EPA Footwear tester foot plate Wrist cord and wrist band (wrist strap) EPA ground cord EPA ground Earth bonding point (EBP) Groundable point of trolley

11 12 13 14 15 16 17 18 19 20

ESD protective footwear Ionizer Working surfaces Seating with groundable feet and pads Floor Garments Shelving with grounded surfaces Groundable racking EPA sign Machine

standard SSI and MSI components. Also, the reduced size or greater functionality (or both) of the resulting product may give the company using a custom IC a market advantage over its competitors. There are several different approaches to custom IC design, each applicable over a certain range of production volumes, because production costs per IC are inversely related to setup and design costs. The most expensive option to set up (full custom design) produces the lowest-cost ICs, but only if hundreds of thousands of ICs are to be made. Conversely, the cheaper techniques such as gate arrays are much less expensive to set up, but produce more expensive chips. For small production runs of perhaps 10,000 ICs, however, they may offer the cheapest total cost.

Full-custom integrated circuits

The most expensive form of ASIC is the full-custom integrated circuit. These are designed and manufactured in exactly the same way as standard ICs. They are justiable only for very high-volume applications

Trade-offs among unit cost,

Trade-offs among unit cost, setup or design costs, and

setup or design costs, and manufacturing quantity occur

manufacturing quantity occur in all elds of engineering.

in all elds of engineering.

(hundreds of thousands of ICs per year) because of the high cost of design. They also are the cheapest type of custom IC per unit manufactured.

Companies using full-custom ICs in their products either have their own design and manufacturing facilities or contract out the work to a semiconductor manufacturer.

Standard-cell integrated circuits

Computer-aided design (CAD) has made possible a cheaper approach to custom IC design, known as the standard-cell system. A computer database holds a library of common circuit elements such as logic gates, ip-ops, operational ampliers, and so on stored in the form of their physical layout within the chip. The library is similar to the standard small- and medium-scale ICs used in noncustom design, but it may also include larger components such as microprocessors (known as “cores”). A standard-cell library, however, can hold many more designs than are available as standard ICs. A standard-cell chip is designed by assembling the required cells from the computerized library (using CAD software). This involves not only the logical circuit design but also the physical placement of cells on the chip. Since the shape and size of the

Figure 3.9 A Xilinx eld-programmable gate-array chip showing a regular array of congurable logic blocks. (Courtesy of Xilinx Inc.)

cells are xed, there will be unused areas among cells (Figure 3.8). A standard-cell design will therefore occupy more chip area than a fullcustom equivalent. Production costs will be higher than for a full-custom design because there will be fewer chips per wafer, but on the other hand design costs are lower because a lot of the design work has already been done in designing the standard cells (and the cost of that work is shared among all the purchasers of the cell library).

Once designed, a full set of masks has to be fabricated exactly as for

a full-custom IC, and the manufacturing process is identical.

Gate arrays

A third method of custom IC design and manufacture exists, with a lower design cost than the standard-cell system. It also differs from fullcustom and standard-cell techniques in the manufacturing stages. The gate-array manufacturer designs a standard chip with a xed layout of transistors or logic gates and ip-ops, and produces a full mask set with the exception of the metallization masks. Wafers are processed in quantity by the normal processes, but no metallization is applied. The customer designs the metal interconnect using CAD software. A custom mask is then made for the interconnect, and preprocessed wafers are metallized to the customer’s design (Figure 3.9).

The setting-up costs of design and mask manufacture are comparatively low for gate arrays, so that a gate-array chip design may be viable for production quantities of only a few thousand per year. Gate arrays have several disadvantages, however, compared to standard-cell design,

Small batches of gate-array ICs

Small batches of gate-array ICs may be metallized by electron

may be metallized by electron beam lithography, which was

beam lithography, which was described earlier in this chapter.

described earlier in this chapter.

including longer interconnection paths on the chip due to the xed layout of the array elements.

Programmable logic and gate arrays

The cheapest form of custom IC in terms of design costs is programmable logic. Several types of device exist, but all have in common the fact that they are manufactured in large quantities, fully packaged but uncustomized. They contain an array of basic elements interconnected by user-programmable links. The links can be fuses or insulated-gate MOS transistors that can be selectively burnt through or charged respectively to dene the connectivity and therefore the function of the logic. Field-programmable gate arrays (FPGAs) are similar to the gate arrays described in the previous section except that interconnections among logic elements are programmable. Programmable logic arrays (PLAs) consist of a xed AND–OR structure with programmable connections through a diode matrix. The same technology can be used for programmable read-only memories (PROMs) used in computers and microprocessor systems to store machine-code programs and processor microcode.

Programmable logic and read-only memories (ROMs) can also be fabricated as mask-programmed devices in which the logic function or memory contents are dened by the metallization mask. Single-chip microprocessors with on-board ROM can also be mask-programmed. Mask programming allows lower-cost, high-volume production once the logic or program design has been proven.

Any electronics manufacturer, whether using standard ICs or custom-designed parts, needs more than one source of supply to guard against component shortages caused by technical or other problems at the IC manufacturer’s facilities. The semiconductor industry recognizes this need and tries to set up second sourcing wherever possible. Typically, two semiconductor companies exchange mask sets so that each can manufacture and sell the other’s designs. In some cases, very popular parts such as the 741 operational amplier or the 555 timer IC become available from ve or more manufacturers. Problems can occur, however, especially with microprocessors and other LSI chips, if the different manufacturers have designed their chips independently rather than exchanging mask sets, because of detail differences among the nominally identical products.

Summary

Integrated-circuit technology is a very important part of modern electronic engineering. It makes possible low-cost, highly reliable products with a high level of functionality. Most integrated circuits are fabricated from silicon. Integrated transistors and other devices are formed within a silicon wafer by incorporation of dopants into the silicon crystal structure. The regions to be doped are dened by lithography and etching of a silicon dioxide layer grown on the surface of the wafer. Dopants can be introduced by diffusion or ion implantation. Additional silicon may be built up on a wafer by epitaxial growth, in which silicon atoms are deposited on the wafer from a gas. The epitaxial layer is an

extension of the existing crystal structure of the wafer. Interconnections among devices on an IC may be made by a deposited metallization layer. Throughout all of these processes, whole wafers containing hundreds of ICs are processed together often in batches of dozens of wafers. There are several different styles of IC package, each with its own advantages and disadvantages. These can be broadly divided into through-hole or surface-mounting types and into plastic and ceramic types. Surface-mounting types are smaller than equivalent through-hole types, and ceramic packages are more reliable and expensive than plastic packages.

Semiconductor devices and ICs are sensitive to thermal and electrostatic damage and correct handling precautions must be taken in a production environment to reduce product reject rates.

ICs for custom applications may be designed in the same way as standard production ICs provided large quantities are required, or by assembling a design from predesigned standard cells, or by designing a metallization layer to interconnect a mass-produced, but unmetallized, gate array. For small batches of custom circuits, eld-programmable gate arrays or logic arrays may be used.

Problems

3.1 At the time of publication of the rst edition of this book in 1987,

VLSI ICs with 106 transistors per chip were reaching production. Using the smallest discrete transistors available, which are surface mount devices measuring about 2 mm × 3 mm, what area of PCB would be required to realize the equivalent function, making no allowance for interconnections? What area of PCB would be required to realize the equivalent of a 2006-made chip with 150 million transistors?

3.2 Moore’s law says that IC complexity doubles approximately every

18 months. How many transistors would have been expected on the largest chips in production in 2006, when this third edition was published? Assume 106 transistors per chip in 1987 when the rst edition of this book was published. Compare the result with the actual number of transistors on the largest chips made in 2006 (see Figure 3.1).

Power sources and power supplies

Objectives

To introduce the main sources of electrical energy used in

□

electronic systems, including mains supplies, batteries, and

photovoltaic cells.

To introduce the concept of a power supply.

□

To discuss the characterization and performance of power supplies.

□

To explain the functions of the main subcircuits found in a power

□

supply.

To explain the operation of linear and switching voltage regulators.

□

All electronic circuits and systems require energy to operate. Energy is required to move electric charge; to produce heat, light, or sound; to produce mechanical movement; and to manipulate information (as in a computer). Energy is a conserved physical quantity: in a closed system energy can be neither created nor destroyed, although it can be converted from one form to another. In electronic engineering, we are usually concerned with electrical energy, although other forms of energy are also important. Heat, for example, is produced in electronic circuits, usually as a by-product of a useful function, and is discussed in Chapter 7. Energy may be stored as chemical energy in a cell or battery. Chemical energy sources are discussed later in this chapter.

While the importance of energy should not be forgotten, electronic engineers more frequently use the concept of power. Power can be used to quantify the rate at which heat is produced in a resistor, the mechanical output of a motor, or the rate at which an electronic system takes energy from its energy source.

The terms a.c. and d.c. stand for alternating current and direct current respectively. We customarily talk about a.c. voltage and d.c. voltage, even though technically this is nonsense — what is an “alternating current voltage?” We can avoid the term

d.c. by talking about steady voltages and currents, and, in the frequency domain, zero frequency (for example, “A low-pass lter has unity gain at zero frequency”).

Let us examine the form in which electronic circuits of various types require their electrical energy supply. If we leave aside a.c. power control systems such as thyristor motor controllers, most circuits operate from one or more d.c. supply rails. Some circuits require only one rail and associated return (usually at 0 V), while others require two rails that are often symmetric with a 0 V return. Logic circuits and microprocessors, for example, usually operate from a single +5 V or +3.3 V rail, or as low as 1.8 V for some modern circuits, while linear circuits such as active lters using operational-ampliers (op-amps) require perhaps ±12 V rails. Most low-power electronic circuits use voltages below 20 V with respect

Energy is the capacity to do

Energy is the capacity to do work. The International System

work. The International System of Units (SI) unit of energy is the

of Units (SI) unit of energy is the joule (J).

joule (J).

Power is the rate of conversion,

Power is the rate of conversion, utilization, or transport of energy.

utilization, or transport of energy. The SI unit of power is the watt

The SI unit of power is the watt (W), which is 1 J s–1.

(W), which is 1 J s–1.

A typical ripple waveform with

( )V

A typical ripple waveform with d.c. component VM, ripple

d.c. component VM, ripple component V r, and ripple

component V r, and ripple frequency 1/Tr.

frequency 1/Tr.

to earth. Power ampliers may require supply rails of up to 300 V. Cathode-ray tubes (CRTs) used in older televisions, oscilloscopes, and visual display units required voltages of up to 5 kV or more, although normally at low current.

Apart from a nominal d.c. voltage, what other characteristics of a d.c. supply rail need to be specied? First, the voltage tolerance of the rail must be dened. A nominal +5 V rail supplying low-power Schottky transistor–transistor logic (LS TTL) logic, for example, could be allowed to vary from +4.75 to +5.25 V because this is the recommended operating voltage range for LS TTL circuits. Second, the allowable ripple on the rail must be dened. Ripple is a small a.c. component superimposed on the mean d.c. level and is usually due to the a.c. source from which the d.c. rail has been derived. Note that the ripple waveform is not usually sinusoidal. Ripple is specied by its frequency and its peak-to-peak amplitude.

Energy sources

As designers of electronic apparatus, we are concerned with the form in which energy is to be supplied to or contained within our equipment. The ultimate source of energy used by our design is of interest only in so far as this affects the form, availability, and cost of the energy delivered to our system. We can distinguish two main forms of electrical energy that might be supplied to an electronic system: a.c. or d.c.; and several ways in which an energy source can be built into an electronic apparatus. Let us examine the characteristics of each of these.

Electronic equipment is usually

Electronic equipment is usually operated from a single-phase

operated from a single-phase a.c. supply.

a.c. supply.

R.m.s. stands for root mean

square — the square root of

square — the square root of the time-averaged value of the

the time-averaged value of the square of a waveform:

square of a waveform: V

rms

A.C. mains

Electronic equipment is often supplied with electrical energy from an external a.c. supply. The most common case is the mains supply derived from an electrical grid. The ultimate source of energy provided by a grid may be coal-red, oil-red, nuclear or hydroelectric power stations, diesel or gas-turbine generators, or wind turbines. Grid supplies are characterized (in the developed countries at least) by high reliability and low-cost energy (compared to other sources). A.C. supplies may also be found on board trains, aircraft, and ships where (except in the case of electric trains) the supply will be provided by generators driven from the engines.

A.C. supplies are widely used because of the efciency and ease with which a.c. can be transformed from one voltage to another. Alternating supplies are characterized by their voltage and their frequency. Three common nominal frequencies are used: 50 Hz on the grid systems of Europe, Africa, Asia, and Australia; 60 Hz in North and South America; and 400 Hz on board aircraft. The higher frequency used on aircraft is due to the reduced size and weight of transformers operating at the higher frequency.

Nominal mains voltages vary from one country to another. 230 V a.c. root mean square (r.m.s.) is normal in Europe, although 110 V is often found on building sites and in some factories where the lower voltage is used for safety reasons. In the United States, 120 V a.c. r.m.s. is commonly

used. A.C. supplies may vary in voltage, frequency, or both from their

1 2

nominal values. These variations are due to load variations on the supply — it is common, for example, in the United Kingdom for the frequency to drop by up to 1.5 Hz when a very popular television programme such as a major sporting event ends. This effect is caused by millions of people switching on electric kettles to make cups of tea and therefore creating a sudden increase in electrical load that the power stations cannot instantly meet. The frequency drop is usually corrected within a few seconds as the power stations supply more power. Local voltage drops may also occur, particularly in premises located some distance along a supply cable, as consumers nearer the substation switch on heavy loads. On small electricity grids, voltage and frequency variations may be more marked, and the design of electronic equipment may have to take this into account.

Apart from voltage and frequency uctuations, a.c. supplies may also be subject to momentary interruption (as, for example, when the local electricity company switches a substation from one circuit to another) and may also carry electrical interference in the form of high-frequency periodic disturbances, switching transients, or harmonics of the grid frequency. This interference may be caused by other electrical or electronic equipment connected to the supply or by natural phenomena such as lightning discharges, and could have a serious effect on the operation of electronic systems that are not designed to cope with it.

Internal energy sources

For a sinusoid, the r.m.s. value is

times the peak value, so

the peak value of the European

the peak value of the European 230 V a.c. r.m.s. mains supply is

230 V a.c. r.m.s. mains supply is 2302V or 325 V.

2302V or 325 V.

Harmonics of the grid frequency

Harmonics of the grid frequency may be caused by thyristor-

may be caused by thyristorcontrolled equipment. For further

controlled equipment. For further details, see Bradley (1995).

details, see Bradley (1995).

If a piece of electronic equipment has to be self-contained with its own power source, there are only two options (neglecting portable diesel engines and generators). These are electrochemical cells and photovoltaic cells. Hand-held mobile phones are an example of self-contained equipment powered by electrochemical cells. Electrochemical cells directly convert chemical energy into electrical energy. They can be divided into two types: batteries and fuel cells. Photovoltaic cells directly convert the energy of visible or ultraviolet (UV) light into electrical energy. They are, of course, tremendously important on spacecraft such as communications satellites.

Fuel cells convert chemical energy from reactants supplied externally to the cell into electrical energy. Their best-known application is on board manned spacecraft, such as the NASA Space Shuttle where hydrogen– oxygen fuel cells supply electrical power (and drinking water as a useful by-product). Fuel cells will not be covered further in this book.

Batteries

Batteries are closed electrochemical power sources. They convert chemical energy from reactants incorporated into the device during manufacture to electrical energy. Originally the term cell was used in this context, a battery being a collection of cells wired in series or parallel. In modern usage, cells are often referred to as batteries.

Two main types of battery exist, known as primary and secondary. Primary batteries can be used once only: when the chemical reactants have been consumed as a result of electrical discharge of the battery, the

A very specialized alternative,

A very specialized alternative, used for deep space probes,

used for deep space probes, is a nuclear thermoelectric

is a nuclear thermoelectric generator that uses the heat

generator that uses the heat from radioactive decay of a mass

from radioactive decay of a mass of plutonium to provide electrical

of plutonium to provide electrical power.

power.

Many batteries contain toxic

Many batteries contain toxic metals such as lead, cadmium,

metals such as lead, cadmium, and mercury, and their safe

and mercury, and their safe disposal or recycling should be

disposal or recycling should be considered when considering

considered when considering their use. Alternatively, the

their use. Alternatively, the use of less environmentally

use of less environmentally damaging batteries should be

damaging batteries should be considered wherever possible.

considered wherever possible. Batteries are exempted from the

Batteries are exempted from the European Union’s Restriction

European Union’s Restriction of Hazardous Substances

of Hazardous Substances (RoHS) Directive because of the

(RoHS) Directive because of the lack of alternatives to lead and

lack of alternatives to lead and cadmium.

cadmium.

1.5

0.8 1 2 3 Service

life (h)

Continuous discharge

A typical discharge characteristic for two

diﬀerent loads.

6Ω load

3Ω load

1.5

0.8 1 2 3 Service

life (h)

Continuous discharge

A typical discharge characteristic for two

diﬀerent loads.

6Ω load

3Ω load

Whether the capacity is stated

Whether the capacity is stated in Wh or Ah, it is a measure

in Wh or Ah, it is a measure of the energy available from

of the energy available from the battery, not its ability to

the battery, not its ability to store electric charge (as in a

store electric charge (as in a capacitor): the mechanism of

capacitor): the mechanism of energy storage in a battery is

energy storage in a battery is chemical, not electrical.

chemical, not electrical.

device must be discarded. Secondary batteries are based on a reversible chemical reaction: the battery may be recharged by passing electrical current through the device in the opposite direction to the discharge current. Despite their ability to be recharged, secondary batteries have a nite useful life: eventually, recharging fails to store sufcient chemical energy in the battery for useful operation.

Different types of battery have different nominal open-circuit voltages, depending on the electrochemical reaction used in the cells and the number of cells in the battery. The actual voltage provided by a battery falls as the energy stored in the battery is used. The change in voltage is shown on a discharge characteristic for stated conditions of discharge, such as continuous discharge into a specied resistance or intermittent discharge for a specied period per day.

When comparing and selecting batteries, we are usually interested in the amount of energy that the battery can supply before it is fully discharged. This quantity is called the capacity of a battery and can be stated in watt-hours (Wh), which is a unit of energy, or more frequently, in ampere-hours (Ah), which is a unit of electric charge.

Since a battery normally has a fairly constant voltage during discharge, we can calculate the approximate energy content of a battery by multiplying its capacity in Ah by its nominal voltage, remembering that an ampere-hour is 3600 coulombs because there are 3600 seconds in an hour.

Worked Example 4.1

The concept of the C rate is an

The concept of the C rate is an approximation, valid only over a

approximation, valid only over a limited range of discharge rates.

limited range of discharge rates.

Calculate the energy content of a 2 Ah 12 V battery in (a) watt-hours and (b) joules.

Solution

(a) 2 Ah × 12 V = 24 Wh (b) 2 Ah × 3600 × 12 V = 86.4 kJ

As in many other areas of electronic engineering, the concept of a normalized quantity is useful when comparing different systems. Battery discharge rates are often expressed in a normalized form known as the “C” rate. A discharge current of 1 C will discharge a battery in 1 hour. A rate of C/5 will discharge a battery in 5 hours, and a rate of 5 C will discharge a battery in 12 minutes. For a 2 Ah battery, the C/5 rate is 400 mA. The C rate can also be used to quantify charge rates for secondary batteries.

The capacity of a battery is not a precise measure of the energy available because the amount of energy that can be extracted from the battery depends upon how it is used. Some types of battery provide more energy in total if they are used intermittently and are allowed to “rest” between discharges, while others are more suited to continuous discharge at a steady rate. When discussing capacity, it is usual to say that the capacity of a battery is dependent on the pattern of discharge, although in reality, it is the amount of energy that can be extracted that is variable.

All batteries will self-discharge to some extent when not in use, because chemical reactions occur within the battery even when no

current is being drawn. This limits the performance of secondary bat-

0.5 hours W V

×=3

0 1.

7 5 0 5

0 25

. .

hours W

teries and the shelf-life of primary batteries (typically to a few months or years), except for some types of lithium primary battery, which have very long shelf lives.

Worked Example 4.2

A portable VHF radio transceiver consumes 3 W of power at 15 V when transmitting and 0.5 W when receiving only. If the unit is to operate from a secondary battery over an 8-hour shift, what battery capacity is required, assuming the radio is transmitting for a total of 30 minutes during the shift? What has been assumed about the battery capacity?

Solution

Transmitting:

Receiving only:

Total energy = 0.35 Ah

A battery of 0.35 Ah capacity is needed, assuming that the intermittent loading will not reduce the battery’s capacity and making no allowance for loss of capacity with life.

Main primary battery types

The most widely used primary battery is the Leclanché cell and its variants. Originally, Leclanché cells were wet systems in glass jars, but nowadays the cells are made using chemical reactants in paste form and are referred to as “dry” cells. There are three cell types in common use based on the Leclanché cell. The rst of these is known as a zinc–carbon cell from its construction with a carbon rod electrode down the centre of a zinc can that serves as a case and the outer electrode. The paste electrolyte in the cell consists of ammonium chloride and zinc chloride mixed with manganese dioxide. This is the cheapest of the three Leclanché cells and is best suited to applications where current is drawn from the battery intermittently (say, for 1 hour per day). The available capacity of the zinc–carbon cell varies with the pattern of discharge, so that it is not possible to state a single capacity for a cell. The capacity is also reduced at low temperatures, making these cells unsuitable for use in equipment to be used out of doors in freezing conditions. The second variant of the Leclanché cell is known as zinc chloride and has all the ammonium chloride of the zinc–carbon cell replaced by zinc chloride. This gives greater capacity at high current drains and better low-temperature performance, although at greater cost than with the zinc–carbon cell. The third variation on the Leclanché cell is the alkaline–manganese cell, which has a potassium hydroxide (alkaline) electrolyte. These cells

Zinc–carbon cells are now

Zinc–carbon cells are now available in a limited range of

available in a limited range of sizes. They have been replaced

sizes. They have been replaced by zinc–chloride cells.

by zinc–chloride cells.

are suitable for continuous high-current discharge and have about four times the capacity of a similar sized zinc–carbon cell used in this way. Their low-temperature performance is similar to that of zinc–chloride cells but at higher cost. The shelf life of alkaline–manganese cells is good — they will retain up to 80% of their initial capacity after 4 years in storage at 20°C. All three Leclanché types have a nominal open-circuit voltage of 1.5 V.

Several types of cell are based on lithium. These were originally developed for military applications in, for example, munitions, where their exceptional shelf life of over 10 years was important. Lithium cells have a very wide operating temperature range down to below –20°C and up to over 50°C. Lithium cells are potentially more hazardous than other types of cells, and manufacturer’s data sheets and safety advice should be followed carefully.

For miniature equipment such as watches and calculators, there are three primary cell types manufactured as “button” cells. These are the zinc–mercuric oxide cells with an open-circuit voltage of about 1.35 V, the zinc–silver oxide cell with an open-circuit voltage of about 1.6 V, and lithium cells with an open-circuit voltage of 3 V. The rst two types have a at discharge characteristic: the cell voltage remains fairly constant until the battery is almost fully discharged. Zinc–silver oxide batteries are also made in large sizes for military applications such as missiles and torpedoes, where they are used for their high capacity per unit mass despite their high cost.

Table 4.1 summarizes the main types of primary battery.

Main secondary battery types

The main types of secondary battery are the lead–acid type used in road vehicles for starting, lighting, and ignition (SLI) and the nickel–cadmium (NiCd) type used in aircraft and military vehicles, both of which are also available in smaller sizes for powering portable equipment; and the nickel metal hydride (NiMH) type widely used in laptop computers.

The lead–acid cell consists of metallic lead electrodes and sulphuric acid electrolyte. Lead–acid cells have a nominal open-circuit voltage of 2 V. The capacity of lead–acid batteries drops very rapidly below 0°C. Vehicle batteries account for a large proportion of lead–acid battery production. They must support short intense discharge of up to 5 C on engine starting. Automobile batteries are rated at 30 to 100 Ah at 12 V, while commercial vehicle batteries have capacities of up to 600 Ah at 24 V. Larger batteries are used for traction applications such as electric road vehicles and railway locomotives.

For portable equipment, sealed lead–acid batteries are available with capacities from 2 to 30 Ah. These are of lower cost than nickel–cadmium batteries, but are heavier for the same capacity. They have a life in excess of 300 charge–discharge cycles and are maintenance free, needing no topping up of the acid electrolyte. They exhibit a fairly constant voltage during discharge at up to the C/4 rate, and can withstand short high-rate discharge.

Nickel–cadmium batteries are based on cadmium and nickel oxide electrodes with a potassium hydroxide electrolyte. The open-circuit

Table 4.1 Main primary battery types

Nominal open-

Type

circuit voltage Main characteristics

Zinc–carbon

(Leclanché)

Zinc–chloride

(Leclanché)

Alkaline–manganese 1.5 V Suited to high-current

Lithium–thionyl

chloride and

Lithium–manganese

dioxide

Zinc–mercuric oxide 1.35 V Available as “button” cells for

1.5 V Low cost. Best used intermittently. Poor performance at low temperature.

1.5 V Improved capacity at high current drain compared to zinc–carbon, and better lowtemperature performance.

continuous discharge. Long shelf life. Similar lowtemperature performance to zinc chloride but at higher cost.

3.5 V

3 V

High cost. Long shelf life.

Wide operating temperature range down to –20°C or less and up to +50°C or more.

miniature equipment such as watches. Flat discharge characteristic. Good shelf life.

Zinc–silver oxide 1.6 V High cost. Flat discharge

characteristic. High capacity per unit mass in large sizes. Military applications.

voltage is about 1.2 V. Nickel–cadmium batteries are more expensive than lead–acid batteries and are the most important alkaline secondary type. Unlike lead–acid cells, they can work well at temperatures down to less than −30°C. They have a at discharge characteristic and can accept continuous overcharging at a low charge current. (This is known as “trickle” charging.)

Nickel metal hydride batteries are similar to nickel–cadmium ones but are less environmentally hazardous. Their capacity is greater than that of NiCd cells, and they are widely used in hybrid vehicles (for example, the Toyota Prius) and in portable electronic devices such as phones and laptop computers. They have a higher internal resistance than NiCd cells and therefore are less suited to applications with high current demand.

For further reading on batteries,

Photons

n-type silicon

p-type silicon

Cathode

Anode

A simpliﬁed p–n junction

photovoltaic cell.

Photons

n-type silicon

p-type silicon

Cathode

Anode

A simpliﬁed p–n junction

photovoltaic cell.

e direction of current ﬂow

from a photovoltaic cell.

−

e direction of current ﬂow

from a photovoltaic cell.

−

For further reading on batteries, see Vincent and Scosati (2000).

see Vincent and Scosati (2000).

Other cells are based on

Other cells are based on cadmium sulphide, selenium,

cadmium sulphide, selenium, and gallium arsenide. All include

and gallium arsenide. All include a p–n or semiconductor–metal

a p–n or semiconductor–metal junction.

junction.

The arrow in the diode (and

The arrow in the diode (and junction transistor) symbol

junction transistor) symbol represents the direction of

represents the direction of forward current ow when the

forward current ow when the diode is driven by an external

diode is driven by an external circuit. Conventional current, by

circuit. Conventional current, by historical accident, ows in the

historical accident, ows in the opposite direction to electrons.

opposite direction to electrons.

Lithium ion batteries are also widely used for portable devices, particularly laptop computers. Their light weight and slow self-discharge are useful, but they lose capacity due to ageing, with or without use.

Photovoltaic cells

Photovoltaic cells convert the energy of visible or ultraviolet light into electrical energy. They are used on board communications satellites to supply electrical power where no other energy source other than a radioisotope generator or small nuclear reactor could supply the power needed over the many years of the satellite’s life. Solar cells may also be used on Earth, for example to power communications relay stations located in remote regions far from electrical grids, or roadside illuminated signs. In this case, electrochemical cells are needed to store energy for use during the night when direct solar power is not available. On a smaller scale, some electronic watches and calculators are powered by photovoltaic cells with secondary electrochemical cells providing energy storage to power the device while in darkness or low light.

The most important photovoltaic cells are silicon junction diodes of large area with a thin n-type region on the exposed face.

A p–n junction with no externally applied bias develops a space charge region on either side of the junction as mobile charge carriers diffuse across the junction under the inuence of concentration gradients. In equilibrium, there is no net charge transport across the junction because of the presence of an electrostatic potential.

If electromagnetic radiation of suitable wavelength now illuminates the junction, electron-hole pairs can be created by photon absorption. This process can be thought of as the ionization of a silicon atom, creating a free electron and a positively charged silicon ion. The silicon ion can attract an electron from a neighbouring atom, so that the positive charge (a hole) is mobile. The free electron and the free hole are, however, inuenced by the electrostatic potential difference across the junction: they move in opposite directions, the electron towards the cathode and the hole towards the anode. If the diode is connected to an external circuit, current can ow and supply power to the external circuit. The direction of this current ow is that of a reverse current through the diode: the positive potential is developed at the anode. This means that in a circuit where a photovoltaic cell charges a secondary battery, a blocking diode must be connected in series with the cell to prevent the battery from driving forward current through the cell during darkness. The blocking diode must have a low forward-voltage drop, and often a Schottky barrier (semiconductor–metal) diode is used. The electromotive force (e.m.f.) generated by a silicon photovoltaic cell is around 0.5 V, so that practical circuits using secondary batteries have several photovoltaic cells in series to raise the e.m.f. to a practical level. Solar cells may also be made from amorphous silicon, and although cheaper, they are less efcient.

The abbreviation PSU for power-

supply unit is often used.

Power supplies

So far in this chapter, we have looked at energy sources and the forms in which energy may be used by electronic circuits. We can now look

at the way in which electrical energy can be converted into the required form. In many electronic systems, this conversion is performed by a subsystem called a power supply. Even in battery-operated equipment where the energy source provides energy in almost the required form, there may still be some form of power supply.

A wide range of power supplies are available commercially, usually from manufacturers specializing in this eld. Many larger electronic systems use commercial power supply units bought off the shelf. In other cases, however, a custom design may be needed, because for example a special physical form, low weight, or high reliability is required. For high-volume mass production, a custom power supply, as with any other component or subsystem, may be cheaper than any commercial off-theshelf design because of a better match between the power-supply design and the system requirement. Often such custom design is contracted out to a power-supply rm with special expertise.

We can now look at the main functions of a power supply. The most obvious and common function of a power supply is to convert electrical energy at the source voltage to some other voltage, higher or lower than the source voltage, and with or without a change from a.c. to d.c. or vice versa. A computer, for example, might be designed to operate from an a.c. mains supply and yet contain circuitry operating at 5 V or 3.3 V d.c. A power supply would be needed, therefore, to reduce the voltage and convert the energy to d.c. On board a communications satellite, d.c. will be available from batteries charged from a solar panel. For economy of space and weight, only one voltage will be available from the batteries. Electronic circuits and systems such as microwave ampliers, attitude controllers, and computers, however, may need a variety of voltages both lower and higher than the battery voltage. D.C.–d.c. converters can produce these voltages at high efciency without wasting valuable energy as useless heat. Inverters generate a.c. from a d.c. input. One common application is on board small boats to generate 120/230 V a.c. 50/60 Hz from a 12 V battery, allowing low-powered domestic mains-operated equipment such as radios and shavers to be used.

Voltage conversion, whether to a higher or lower voltage, is possible in practical terms only in an a.c. circuit, using a transformer. D.c.–d.c. power supplies (or converters) are in reality, then, d.c. to a.c. to d.c. power supplies.

Power supplies operating from an a.c. source also have to provide energy storage during the parts of the source cycle where little or no energy is available from the supply. It may also be necessary to store energy to supply the output current during momentary loss of the a.c. supply, such as happens when a substation is switched. Energy storage is usually in the form of electric charge in the power supply’s reservoir capacitors and is usually practical for only a limited time in most systems. Longer-term energy storage requires the use of batteries. A power supply with batteries is an example of an uninterruptible power supply (UPS). A UPS produces constant output even during breaks in the a.c. mains supply. Laptop computers work in this way, since loss of the mains supply does not interrupt the supply to the computer — this comes from the battery.

A change in voltage is always

A change in voltage is always accompanied by a compensating

accompanied by a compensating change in current: the power

change in current: the power output of a power supply is

output of a power supply is always less than the power input.

always less than the power input.

Voltage increase at low

Voltage increase at low current can be achieved

current can be achieved without a transformer using a

without a transformer using a diode multiplier such as the

diode multiplier such as the Cockcroft-Walton multiplier.

Cockcroft-Walton multiplier. These circuits do, however,

These circuits do, however, require a switching action.

require a switching action.

Larger uninterruptible power

Larger uninterruptible power supplies may use diesel

supplies may use diesel generators as well as, or instead

generators as well as, or instead of, batteries.

of, batteries.

Exercise 4.1

Load regulation =

− ×

( )

V V

10 100

100

0 10 50 100

I/I

max

(%)

A telemetry transmitter uses 4 W of power at 5 V d.c., derived from a mains supply. It is to be capable of continued operation despite momentary interruptions in the mains supply lasting up to 125 ms. Show that a capacitor of about 1 F is needed if connected across the 5 V rail, assuming that the rail voltage must not drop below 4.9 V during the supply interruption.

An example of a load that

An example of a load that would demand varying power

would demand varying power is a power amplier driving a

is a power amplier driving a loudspeaker with speech.

loudspeaker with speech.

Varying denitions of such

Varying denitions of such practical parameters are quite

practical parameters are quite common — be cautious when

common — be cautious when using such parameters.

using such parameters.

Load regulation is normally

Load regulation is normally expressed as a percentage.

expressed as a percentage. An alternative denition might

An alternative denition might use the full-load and no-load

use the full-load and no-load voltages, rather than the full load

voltages, rather than the full load and 10% of full-load voltages.

and 10% of full-load voltages.

The third main function of a power supply is maintenance of a constant output irrespective of varying load. This is known as regulation or stabilization, and a power supply having this feature is called a regulated or stabilized power supply. Not all power supplies have this feature as some applications can tolerate variations in voltage or current. A d.c. supply for low-voltage lament lamps and relays is an example of such an application.

Power-supply performance can be represented graphically as a characteristic, which shows output voltage as a function of output current. Ideally, the output voltage would be independent of output current, but in practice some drop in voltage is unavoidable and is quantied by a parameter known as load regulation. The denition of load regulation varies somewhat, but a fairly common denition is given below and in Figure 4.1.

is the output voltage at 100% of the full load current. Load regula-

100

V10 is the output voltage at 10% of the full load current, and

tion is then dened as

(4.1)

Output voltage variation might also be due to variation in the source voltage. The a.c. mains voltage, for example, can vary, particularly for consumers at the end of a cable some distance from a substation. The terminal voltage of a battery falls as the battery becomes discharged. It is possible to dene a regulation factor to quantify the performance of a

Figure 4.1 Output characteristic of a voltage-regulated power supply showing quantities used in dening load regulation.

power supply under variations of input voltage, but this can be done in

0 I

max

(c)

(b)

(a)

max

many ways and it will not be discussed further here.

A nal factor that can inuence the regulation of a power supply is ambient temperature. Normally, the variation in output voltage with temperature is small and roughly linear over the operating temperature range of the power supply, so that the normal idea of a temperature coefcient can be used to quantify this aspect of the power supply’s performance.

Apart from the precisely controlled voltage, there is another benet to be had from a voltage-regulated power supply. This is a low a.c. or dynamic impedance and is very important in the operation of linear circuits. In analysing, say, a common-emitter amplier circuit, we assume the power supply to be of negligible impedance in deriving our smallsignal model. We must not forget that this assumption may not always be valid.

Overload protection

We can now turn our attention to the part of the output characteristic beyond the full-load current. What happens when the full-rated current of a regulated power supply is exceeded? Figure 4.2 shows some of the possibilities. If no special provision is made in the design of a power

Temperature coefcients are

Temperature coefcients are discussed in the next chapter

discussed in the next chapter in the context of passive

in the context of passive components.

components.

Figure 4.2 Output characteristics of voltage-regulated power supplies: (a) with voltage reduction beyond full-load current, (b) with crossover to constant current at full-load current, and (c) with “foldback” limiting.

Current limiting, as shown in

SCR

triggering

voltage

Overvoltage

sensing

circuit

Regulated

supply

−

SCR

Load

Current limiting, as shown in Figure 4.2b, is a popular option

Figure 4.2b, is a popular option for bench power supplies used

for bench power supplies used in laboratories. Often the limiting

in laboratories. Often the limiting current is adjustable so that

current is adjustable so that limiting can be used to protect

limiting can be used to protect the circuit being tested.

the circuit being tested.

Overvoltages may also occur

Overvoltages may also occur at switch-on and switch-off

at switch-on and switch-off unless the power-supply design

unless the power-supply design deliberately includes proper

deliberately includes proper control such as a dened

control such as a dened “power-up” sequence.

“power-up” sequence.

The characteristics of thyristors

The characteristics of thyristors or SCRs are discussed by

or SCRs are discussed by Bradley (1995).

Bradley (1995).

supply, damage or destruction of some components may occur due to overheating. The power supply may include a fuse that will rupture and disconnect the power supply from its energy source. This possibility is not shown in the gure. Current limiting is the next most likely possibility — the power supply will be designed so that beyond full-rated load current, the output voltage will be reduced. This may be done by limiting current as in Figure 4.2b. A third, more elaborate possibility is known as foldback limiting and is illustrated in Figure 4.2c. Here, once the full-rated load current has been exceeded, the power-supply output voltage and current are reduced, bringing both down independently of the load.

We have considered protection of the power supply itself from damage due to excessive current being drawn by the load. Foldback limiting also protects the load to some extent. One other eventuality that must be considered, however, is overvoltage. Suppose that a regulated power supply supplies current to a large system containing many expensive integrated circuits (such as a computer). If the power supply becomes faulty and the output voltage becomes too high, a large amount of expensive circuitry could be damaged beyond repair if the voltage exceeded the absolute maximum ratings of the integrated circuits. To guard against this prospect, overvoltage protection is normally included in the powersupply design. Figure 4.3 shows a common solution known as a crowbar circuit. The essence of this circuit is to create a short circuit (or at least a very low-resistance path) across the power-supply terminals as soon as an overvoltage is detected (within microseconds). When the overvoltage-sensing circuit detects that the regulated voltage has risen beyond its normal limit, a triggering signal is generated, switching the SCR or thyristor into its conducting state. RA limits the maximum current through the SCR and is typically a fraction of an ohm. Once triggered, the SCR effectively short circuits the regulated supply and prevents damage to the load. The SCR continues to conduct until current is switched off.

We have now examined the main functions and characteristics of power supplies and can turn our attention to the circuits used within power supplies. Figure 4.4 is a block diagram of a typical mains-oper ated power supply. The transformer changes the input a.c. voltage to a higher or lower a.c. voltage, which is then converted to d.c. by the

Figure 4.3 A crowbar overvoltage protection circuit.

Unstabilized d.c.

a.c.

supply

Transformer

Rectiﬁer

Reservoir

capacitor

Regulator Load

−

Figure 4.4 Block arrangement of a typical mains-operated d.c. power supply.

full-wave bridge rectier and reservoir. At this stage, before the regulator, the d.c. voltage is unstabilized (it will vary as the load current varies) and may also have greater ripple than the regulated voltage. The regulator stabilizes the load voltage and may also include overload protection circuits. Practical power-supply designs may not show such a clear distinction among the functional blocks. Let us now consider each functional block in turn.

Transformers

Transformers are widely available commercially in a range of physical sizes, power ratings, and electrical congurations, and are designed for operation at a specic frequency (usually 50, 60, or 400 Hz). The theory of transformer operation is covered elsewhere, and will only be discussed briey here. Transformers consist of one or more electrical windings of low d.c. resistance, wound onto a core of magnetic material (commonly iron). A changing current in one winding induces a changing magnetic eld in the core, which links the same or another winding and induces a changing e.m.f. in that winding. Two main types of transformer are shown in Figure 4.5. The autotransformer (Figure 4.5a) has one winding with intermediate tappings, while the double-wound transformer has separate primary and secondary windings. The double-wound transformer is almost universally used in electronic power supplies, because of the increased safety obtained by having electrical isolation between the windings. The ratio of turns on the secondary winding to the number of turns on the primary winding determines the ratio of secondary voltage to primary voltage. The amount of power that can be drawn from a secondary winding is always less than the power supplied to the primary winding, because some energy is lost as heat in the windings due to their resistance (copper loss) and as heat due to circulating electric currents (eddy currents) in the core (iron loss). Iron losses can be reduced by constructing the core from thin at sheets of iron, or laminations, stacked together and insulated from each other with a layer of varnish, or by constructing the core from ferrite that is nonconductive. Copper losses can be reduced by increasing the crosssectional area of the wire used for the windings. Copper is an expensive metal, however, so the transformer designer must trade off copper losses against the cost of the transformer. For small transformers such as are

See for example Senturia and

See for example Senturia and Wedlock (1993).

Wedlock (1993).

The primary winding is the

The primary winding is the energy-input side of the

energy-input side of the transformer. The secondary

transformer. The secondary winding is the energy-output side

winding is the energy-output side of the transformer.

of the transformer.

Ferrites are discussed

Ferrites are discussed in Chapter 5.

in Chapter 5.

Core

“Shell” type transformer

with all windings on centre-limb.

Core

“Shell” type transformer

with all windings on centre-limb.

Core

“Core” type transformer

with primary and secondary

windings on separate limbs.

Core

“Core” type transformer

with primary and secondary

windings on separate limbs.

Toroidal transformer.Toroidal transformer.

The number of volt-amperes

(a) (b)

(d)(c)

The number of volt-amperes (VAs) is the product of the r.m.s.

(VAs) is the product of the r.m.s. current and the r.m.s voltage. It

current and the r.m.s voltage. It is not equal to power (in watts)

is not equal to power (in watts) because the current and voltage

because the current and voltage are out of phase in an a.c. circuit

are out of phase in an a.c. circuit (unless the circuit is purely

(unless the circuit is purely resistive).

resistive).

Figure 4.5 Types of transformer: (a) autotransformer, (b) double-wound transformer, (c) double-wound transformer with centre-tapped secondary winding, and (d) double-wound transformer with multitapped primary winding and separate secondary windings.

used in electronics, cost is likely to be more important than a small energy loss.

There are three common constructional designs for double-wound transformers shown diagrammatically in the margin. The shell type is the most common for small transformers and has all the windings wound on a common centre limb of the core. Small core-type transformers are less common and have primary and secondary windings located on separate limbs of the core. The chief advantage of this arrangement is reduced capacitive coupling between the windings. The third transformer shown consists of a toroidal (doughnut-shaped) core and is a compact design popular for low-prole equipment. It also has reduced ux leakage compared to other designs, a factor that can be important in audio-ampliers and low-frequency instruments because of their tendency to pick up and amplify mains frequency signals. When mounting toroidal transformers, it is most important not to create an electric circuit through the centre of the toroid. This would constitute a short-circuited secondary and would cause overheating and possibly destruction of the transformer. Mountings should either use nonconducting fasteners (such as nylon screws) or be electrically insulated at one or both ends.

The isolation between the primary and secondary windings of a transformer can be less than perfect at higher frequencies because of capacitive coupling between the primary and secondary windings. This coupling may result in transmission of electrical interference (unwanted noise and transients) into the secondary circuit. To reduce the coupling effect, some transformers are tted with an electrostatic screen between the primary and secondary windings, brought out to an external terminal that is usually grounded.

Transformer ratings are usually expressed in VA (volt–amperes), rather than watts, or by stating the maximum r.m.s. current rating of each winding. Transformer secondaries should be fused, or otherwise

protected against damage by a short-circuit load, because a secondary

Load

(a)

(b)

−

Load

−

Load

−

winding can be supplying heavy current under fault conditions without the primary current exceeding the rating of the primary winding. It is therefore not sufcient to rely on a fuse in the primary circuit to protect the secondary.

Rectication

Rectication is the conversion of a.c. to pulsed d.c. Several rectier circuits are shown in Figure 4.6. The most important rectifying component in modern use is the semiconductor diode. Power diodes may sometimes be referred to as rectiers to distinguish them from signal diodes. Fullwave rectication is universally used in electronics and produces pulsed d.c. at twice the supply frequency. This can be achieved with two diodes if a centre-tapped transformer is used, or with four diodes connected in a bridge conguration. The bridge conguration is the most widely used nowadays, requiring a secondary winding without a centre tap on the transformer. Four diodes arranged in a bridge with four terminals or leads are commonly available and known as bridge rectiers.

The centre-tapped full-wave circuit was widely used in the days of thermionic valve electronics, the two diodes being implemented within one valve envelope (a double diode). This is a good example of how cost inuences design: at one time extra copper (in the form of a centre-tapped secondary) was cheaper than extra diodes (to form a full-wave bridge).

Rectier or power diodes have separate ratings for average forward current and surge current. As will be seen in the next section, the inclusion of reservoir capacitors in power-supply circuits can cause large surge currents to ow when the diodes start to conduct in each cycle. Power diodes also have greater forward-voltage drop than small-signal diodes

Figure 4.6 Rectier circuits: (a) half-wave, (b) full-wave, (c) full-wave bridge, and (d) alternative representation of the full-wave bridge.

= ≈

≈

∆

mA ms

100 10

−90°

6.5

10 ms

6.45

∼83°

ωt

0° 90°

−90°

6.5

10 ms

6.45

∼83°

ωt

0° 90°

Rectiﬁer

current

(a)

(b)

Figure 4.7 (a) Output waveform of a full-wave rectier circuit with reservoir capacitor, and (b) rectier current.

so that power dissipation in a rectier diode can be signicant. Lastly, the maximum reverse voltage or peak inverse voltage (PIV) rating of rectiers must be adequate both for the normal reverse voltages present in the circuit and for any abnormal transient voltages on the supply.

Reservoir capacitors

For the majority of electronic applications, the pulsed d.c. output from a rectier is unsuitable directly. In some applications, a reservoir capacitor or smoothing capacitor connected between the supply rails is sufcient.

Figure 4.7 shows the effect of a reservoir capacitor on the output of a full-wave rectier. The ripple voltage, vr, is determined by the value of the reservoir capacitor, the load current, and the supply frequency (which determines the time over which the capacitor discharges).

Worked Example 4.3

A reservoir capacitor is to be connected across the output of a full-wave bridge rectier. Calculate:

(a) The phase angle at which the rectiers start to conduct. (b) The value of capacitor required. (c) The peak rectier current, neglecting any series resistance in the

circuit.

The peak value of the rectied waveform is 6.5 V; the supply frequency is 50 Hz. The ripple must not exceed 50 mV peak-to-peak, and the load current is 100 mA maximum.

Solution

–1

(a) The phase angle is sin

(6.45/6.5) ≈ 83°.

(b) Assume the capacitor discharges linearly between the peak of one

cycle and the angle in (a):

= 6 5. cos ,ω ω

max

. ( ) cos= × ×

−

6 5 2 50 83 250





I C

mA mF V mA A

max

.= + = × +

−

100 20 250 100 5 1



∆V

∆I

Rz =

∆V

∆ I

The rate of change of

= 6.5 sin ωt, where ω = 2π × 50 rad/s, 83° ⩽ ωt ⩽ 90°.

and the greatest rate

of change occurs when ωt = 83°.

The peak rectier current is given by:

Voltage references

All regulated power supplies require a voltage reference: a device or circuit that can maintain a constant voltage between its terminals independently or nearly independently of variations in current or ambient temperature. Before considering the various types of regulator, let us consider the most commonly used voltage reference component, the Zener diode.

Figure 4.8 shows the V-I

characteristic of a typical Zener diode. The

characteristic shows only the reverse-biased region because Zener diodes are operated in reverse breakdown. Their forward characteristic is of no interest. The main feature of the characteristic to note is that, for voltages greater than the breakdown voltage, VZ, a small increase in voltage produces a large increase in current. Looked at another way, the voltage across the diode is nominally constant over a wide range of currents.

The V–I characteristic in the breakdown region is not precisely parallel to the current axis (a change in current does produce a small change in voltage). This variation is expressed as the diode’s slope resistance,

RZ (in ohms), and is the ratio of incremental voltage change to incremental current change at a specied current. RZ is not constant, but

There are low-voltage reference

There are low-voltage reference diodes (which are actually

diodes (which are actually integrated circuits) that use the

integrated circuits) that use the band-gap energy of silicon to

band-gap energy of silicon to provide a reference voltage.

provide a reference voltage. These are known as band-gap

These are known as band-gap references.

references.

All junction diodes exhibit

All junction diodes exhibit reverse breakdown, but usually

reverse breakdown, but usually at much higher voltages. Zener

at much higher voltages. Zener diodes are specially fabricated

diodes are specially fabricated to break down at a precise

to break down at a precise voltage and to withstand

voltage and to withstand continuous power dissipation

continuous power dissipation in the breakdown region. The

in the breakdown region. The name Zener diode is actually

name Zener diode is actually a misnomer for diodes with

a misnomer for diodes with breakdown voltages above

breakdown voltages above about 5 V. The Zener effect

about 5 V. The Zener effect occurs when electrons within the

occurs when electrons within the space-charge region of the diode

space-charge region of the diode are dislodged from their atoms

are dislodged from their atoms by the electric eld intensity.

by the electric eld intensity. The dominant process in most

The dominant process in most diode breakdown is avalanche

diode breakdown is avalanche multiplication, where the

multiplication, where the energy of dislodged electrons

energy of dislodged electrons is sufcient to dislodge further

is sufcient to dislodge further electrons from their bonds.

electrons from their bonds.

Figure 4.8 V-I characteristic of a Zener voltage-reference diode.

varies with current. Typically, Zener diodes have slope resistances from

∆

and ∆V V

Z Z

R R= = = ±60 0 1Ω .

∆ mAIV= ± ±

0 1

1 7..

Ω



a few ohms to a few tens of ohms. Breakdown voltages also vary with temperature, the variation being expressed in the usual way as a temperature coefcient.

Worked Example 4.4

A 5.1 V type BZX79 500 mW Zener diode has a slope resistance of 60 Ω maximum at 5 mA. What would be the allowable range of currents if the diode voltage was not to vary by more than ±0.1 V?

Solution

thus

The allowable range of currents is thus 3.3 to 6.7 mA.

Zener diodes are available in power ratings from 400 mW to over 20 W. Note that a Zener must be operated at a suitable current (the operating point must be beyond the “knee” point of the V-I characteristic). Manufacturer’s data usually state the slope resistance at a specied current, and generally, the diode should be operated near to the stated current.

Linear regulators

Linear voltage regulators operate by dropping an unregulated voltage through a dissipative element (either a resistor or a transistor), controlling the voltage drop so as to maintain a constant output voltage. Two possible arrangements of linear regulator are shown in Figure 4.9. The load represents the electronic circuits to be supplied with a constant voltage. It contains active circuits, and its impedance varies with time. In the series regulator, a regulating element is placed in series with the load. The voltage drop across the regulating element is varied as the load current or the unregulated voltage varies in order to maintain a constant voltage across the load. In the shunt regulator, a resistor is placed in series with the load and a regulating element is placed in parallel with the load. The current drawn by the regulating element is varied in order to alter the voltage drop across the resistor and keep the load voltage constant. Both series and shunt regulator circuits are used in practical designs, as we shall see below.

The simplest arrangement of linear voltage regulator is a shunt circuit and is shown in Figure 4.10. The regulated voltage directly across the Zener diode DZ. When the load current IL or the unregulated voltage VU changes, the Zener diode current IZ changes to compensate.

VR is obtained

225

3 3

VmA.

≈

5 3 3

0 068

25−=

Ω

Series

regulating

element

Shunt

regulating

element

Load

(a)

(b)

Unregulated

d.c.

Unregulated

d.c.

Figure 4.9 Two arrangements of linear voltage regulators: (a) series, and (b)

Load

IZ + I

shunt.

Some complementary metal-oxide semiconductor (CMOS) logic circuits operating at +3.3 V are to be included in a system and powered from a simple Zener shunt regulator circuit of the type shown in Figure 4.10. There is a regulated +5 V rail available. The designer decides to use a 300 mW 3.3 V Zener and nds from a data sheet that a typical recommended operating current is 5 mA. Since CMOS logic draws negligible current (micro-amperes) when quiescent, voltage regulation must be maintained down to zero load current. Select a value for the resistor, and calculate the maximum allowable load current. What factors have been neglected in the calculation?

Solution

Maximum Zener current I

Assume load current IL = 0, then R =

Figure 4.10 A simple linear shunt regulator.

Worked Example 4.5

Z,max

IZ + IL is constant at 68 mA. IZ must not fall below 5 mA, therefore I

63 mA. The self-heating, slope resistance, and temperature coefcients of the

Zener diode have been neglected.

L,max

Vz − V

Load

Unregulated

d.c.

VBE − V

Load

Unregulated

d.c.

Figure 4.11 A simple linear series regulator.

Figure 4.12 A transistor shunt regulator.

The simple Zener shunt circuit has several shortcomings, among which are that the Zener diode current and therefore the power dissipated in the diode vary with load current and unregulated voltage and that there is no compensation for the temperature coefcient of the diode. The simple series regulator described next partly overcomes the rst of these problems. Temperature compensation of Zener diodes can be achieved by adding other devices with a similar temperature coefcient of opposite sign in series with the Zener.

Figure 4.11 shows a simple linear series regulator design using a bipolar transistor as the dissipative element. The transistor is in series with the load. The voltage at the emitter of the regulating transistor is constant at VBE less than the reference voltage across the Zener diode. R1 supplies operating current for the Zener diode and base current for the transistor. Variations on this simple circuit exist to overcome problems with temperature stability of the Zener voltage and VBE drop. Protection against short circuit of the load terminals may be needed, since under these conditions the transistor passes heavy current with the full unregulated voltage across the transistor.

Figure 4.12 shows a simple transistor-based shunt regulator. In this design, the dissipative element is the resistor R1 (although some power is also dissipated in the shunt transistor). The shunt regulator has the important advantage of being inherently protected against a short-circuit load, because under these conditions the transistor passes no current. Another useful feature of this circuit is that it provides a path for reverse current from the load and can actually absorb power from the load. This is an advantage when the load is a d.c. motor.

Integrated-circuit voltage regulators are available for commonly used voltages such as 3.3V, 5 V, and 12 V and with adjustable voltage outputs. They are available in both positive and negative polarities and include

a voltage reference, regulating element, and control circuits within the

Load

Control waveform generator

Error

ampliﬁer

ref

Unregulated

d.c.

package. They usually include current limiting and temperature compensation. Low-power types may be packaged in dual-in-line form or surface mount packages. Higher-power devices are packaged in the same way as power transistors and require heat sinking.

Switching regulators

Switching regulators operate at higher efciency than linear types by avoiding power wastage in a series- or shunt-regulating device. They are also smaller and lighter for a given power output than linear regulators. Their disadvantage is that, because of the switching action, output ripple is usually higher than with linear regulators. Also, because of their greater complexity and higher component count, switching regulators are usually slightly less reliable than linear regulators. (Even so, some switching regulators have mean time between failures [MTBFs] of over 200,000 hours, or more than 20 years.)

Switching regulators operate by chopping (switching on and off) the unregulated voltage, matching the demanded power with supplied power. A smoothing circuit produces continuous d.c. from the chopped waveform. The smoothed output is sensed and fed back to control the chopping frequency or pulse width.

Figure 4.13 shows a simple example to show the principle (in reality, circuits are more complex). The circuit operates as follows. The control waveform generator produces a control signal consisting of rectangular pulses that causes TR1 to chop the input voltage. When TR1 is conducting, current ows to the load through the inductor L1. Capacitor

C1 is charged to the output voltage of the regulator. When TR1 switches

off, current continues to ow through L1. The ywheel diode D1 is required to hold down the voltage at the collector of TR1 and to provide a path for the current while the transistor is off. During the time that the transistor is off, the load current is being supplied from the stored energy in the inductor and capacitor. The output voltage is thus a mean

MTBF, or mean time between

failures, is discussed in

Chapter 9.

Figure 4.13 A switching voltage regulator.

d.c. level with superimposed ripple. The mean level is regulated by comparing it with a voltage reference and generating an error signal to control the pulse generator. The output voltage may be varied by changing the frequency of the pulses, keeping their width constant, or by varying the pulse width, keeping the frequency constant. Either way, the markto-space ratio or duty cycle of the pulse generator output varies and controls the output voltage of the regulator.

Typical operating frequencies of switching power supplies are above 20 kHz, both to avoid generation of audible frequency acoustic noise and because inductors for higher-frequency operation can be made smaller and lighter than those for lower frequencies.

Summary

This chapter has looked at the main energy sources used to power electronic systems, starting with the characteristics of a.c. mains supplies and then examining energy sources such as electrochemical and photovoltaic cells, which can be included within self-contained systems.

The idea of a power supply as a subsystem within an electronic system has been introduced, and the performance characterization of power supplies has been discussed. The main elements of a power supply were then described, including transformers, rectier circuits, and reservoir capacitors. Voltage regulation was then introduced, and the two types of regulator (linear and switching) were outlined.

The chapter has thus given a broad introduction to the subject of powering electronic equipment.

Problems

4.1 A D-size nickel–cadmium cell has a nominal 4 Ah capacity and

a nominal open-circuit voltage of 1.25 V. The manufacturers recommend a 12-hour charge at 500 mA. What will be the average heat dissipation during charging assuming that the cell absorbs its nominal capacity in the form of chemical energy?

4.2 The energy stored in a capacitor is ½ CV

. Calculate the value of a capacitor required to store the same amount of energy as a 30 Ah 12 V car battery if the capacitor is charged to 12 V.

4.3 Why would a primary battery manufacturer advise users to store batteries in a cool place and keep them out of direct sunlight?

4.4 Recalculate (a) the capacitor value and (b) the peak rectier cur

rent of Worked Example 4.3 for a supply frequency of 400 Hz.

4.5 What is (a) the maximum discharge rate and (b) the minimum charge rate for a secondary battery used to power a miner’s helmet lamp if the lamp is to operate continuously over an 8-hour shift and then be ready for use at the start of the same shift on the following day? Assume that two thirds of the energy input to the battery is absorbed as chemical energy during charging. (Hint: The answers are expressed independently of the battery capacity.)

4.6 Calculate the efciency of the circuit designed in Worked Example

4.5 at (a) maximum load current and (b) a load current of 5 mA. Efciency is the ratio of power delivered to the load to power drawn from the source (in this case, the +5 V rail).

4.7 For the application given in Worked Example 4.5, an engineer decides to try the circuit of Figure 4.11, and sets the Zener diode current at 5 mA by suitable choice of R1. Calculate the efciency for the same load currents as Problem 4.6. Neglect the transistor base current.

Passive electronic components

Objectives

To emphasize the differences between real components and ideal

□

circuit elements. To introduce the properties and characteristics of real passive

□

components. To survey the main types of passive electronic component and to

□

discuss their fabrication.

Passive component characteristics

A fully detailed data sheet for an apparently straightforward component such as a capacitor contains information on a great many aspects of its behaviour. In any particular application, some of the component parameters will be of the utmost importance, while others will be of little consequence. In some circuits, the ultimate performance that can be achieved is limited by component behaviour. Figure 5.1 shows such a circuit, a single-slope analogue-to-digital converter, or ADC. In this type of ADC, a ramp generator or integrator circuit produces a ramp waveform starting at a voltage slightly below 0 V. Two comparators are used, one to compare the ramp waveform with 0 V and one to compare the analogue input voltage with the ramp waveform. As the ramp waveform increases from below 0 V to greater than the analogue input, the two comparators switch one after the other, and the time interval between the two comparators switching is accurately proportional to the difference between the analogue input voltage and 0 V. The comparator outputs are used to start and stop a counter driven from a clock circuit. The nal digital value in the counter is proportional to the time interval between the switching of the two comparators and therefore to the analogue input voltage. Figure 5.1b shows the details of the ramp generator circuit, which integrates a constant reference voltage to produce a ramp. A eld-effect transistor (FET) switch is needed to reset the integrator at the end of each cycle by discharging the capacitor. (The control signal for the FET has been omitted from Figure 5.1b for the sake of clarity.) The overall linearity of the ADC depends critically on the quality of the ramp waveform: any signicant deviation from an ideal ramp, as in Figure 5.1c, will cause linearity errors in the ADC output. One possible cause of ramp nonlinearity is the integrator capacitor: leakage current through the capacitor dielectric causes the ramp waveform to droop. A further problem is that capacitor dielectrics can absorb charge, a phenomenon that is discussed later in this chapter. The choice of capacitor for this circuit is thus very important as the linearity of the ADC depends on the capacitor.

The single-slope analogue-

The single-slope analogueto-digital converter (ADC) is a

to-digital converter (ADC) is a

classic circuit. Typical ADC

circuits used in practice are more

circuits used in practice are more elaborate.

elaborate.

Linearity is a very important

concept in electronic

concept in electronic engineering. A linear system

engineering. A linear system obeys the principle of

obeys the principle of superposition, such that if input

superposition, such that if input

x1 causes output y1 and input x2

causes output y2, then an input

causes output y2, then an input of x1 + x2 will cause an output of

of x1 + x2 will cause an output of

y1 + y2. In the case of an ADC,

linearity means that the value of

linearity means that the value of the digital output is accurately

the digital output is accurately proportional to the analogue

proportional to the analogue input.

input.

Analogue

input

Start

Digital output

Done

Ramp

generator

0 V

Clock

Counter

control

logic

−

(a)

(c)

(t) = − ∫

0Vref

⋅ dt

ref

(b)

−

FET switch

−V

−

Figure 5.1 A single-slope analogue-to-digital converter: (a) block diagram, (b) ramp generator circuit, and (c) ramp nonlinearity.

Exercise 5.1

Explain why two comparators are used in the single-slope ADC discussed above, and why the counter is not started at the start of the ramp.

An understanding of component parameters is essential to the circuit designer who needs to compare different types of component for a particular application. A good circuit designer comes to know the type of component that will be needed for a particular purpose and understands why that type of component will do the job.

This chapter starts by looking at some general aspects of passive components, and then moves on to discuss specic types of component.

Tolerance

In engineering, no manufactured value or dimension can ever be exact. Engineers express the closeness of a value or dimension to the desired value by a tolerance. Smaller tolerances are more difcult to achieve, and components with small tolerances are therefore more expensive. The acceptable range of a value can be specied in several ways. One is to state the upper and lower limits of the range within which the value must lie. This method is sometimes used in mechanical dimensioning, but is rarely used for component values in electronic engineering, where

the normal practice is to state a nominal value with a tolerance. Often, but not necessarily, the nominal value lies in the middle of the acceptable range and the tolerance is quoted as a percentage of that nominal value. A 100 Ω ± 5% resistor, for example, could have any value from 95 to 105 Ω. For some components, the tolerances above and below the nominal value are unequal, implying perhaps an uneven probability distribution resulting from the fabrication process.

The tolerances described so far have been relative: they represent a

fractional deviation from the nominal value. For some components such as capacitors in the 1 to 10 pF range, the tolerance may be expressed as an absolute value thus: 3 pF ± 0.5 pF.

Preferred values

Clearly it is not possible for a resistor manufacturer to produce economically every possible value of resistance, even to a 1% tolerance. Accordingly, resistors (and capacitors) are manufactured to limited ranges of preferred values. The most common range is known as E12 and contains values of 10, 12, 15, 18, 22, 27, 33, 39, 47, 56, 68, 82, and decimal multiples and submultiples of these values. These are generally sufcient for most purposes. In analogue circuits where a precise ratio of two resistors is required (for example, to dene the gain of a precision amplier), further values are available. The E24 range includes all the E12 values plus values of 11, 13, 16, 20, 24, 30, 36, 43, 51, 62, 75, and 91. There are also E48 and E96 ranges whose values are specied to three signicant gures. Generally, components from the E12 and E24 ranges will be the cheapest and should be used wherever possible.

Electrolytic capacitors commonly

Electrolytic capacitors commonly have capacitance tolerances of

have capacitance tolerances of ±20% but some are specied

±20% but some are specied asymmetrically, for example as

asymmetrically, for example as –10% to +30%.

–10% to +30%.

The E ranges of component

The E ranges of component values are specied in an

values are specied in an EIA (Electronic Industries

EIA (Electronic Industries Association) standard.

Association) standard.

Temperature dependence

The values of electronic components often vary with temperature because the electrical properties of materials vary with temperature. For most components, the fractional change in value is roughly proportional to the temperature change and the temperature dependence of the component value can be expressed as a temperature coefcient, often in parts per million (p.p.m.) per °C.

Worked Example 5.1

A range of ¼ W resistors has a stated manufacturing tolerance of 5% and a temperature coefcient of resistance of ±200 p.p.m. °C–1. Ignoring any other effects on resistance value, assuming the nominal value applies at 20°C, and making no assumption about the sign of the temperature coefcient, what would be the worst-case deviations from nominal value over a temperature range of 0 to 70°C?

Solution

The worst-case values at 20°C are 1.0 ±5%, or 0.95 and 1.05. The greatest deviation due to temperature will be at 70°C, where the change in resistance from the value at 20°C will be (70°C – 20°C) × (±200 p.p.m. °C–1) or ±10,000 p.p.m or ±1%. Taking the worst-case combinations of manufacturing tolerance and temperature dependence:

Maximum deviation = 1.05 × 1.01 = 1.06 or +6% Minimum deviation = 0.95 × 0.99 = 0.94 or –6%

Ionizing radiation, in the form

Ionizing radiation, in the form of cosmic rays and subatomic

of cosmic rays and subatomic particles from the Sun, is a

particles from the Sun, is a signicant problem in electronics

signicant problem in electronics for spacecraft.

for spacecraft.

Exercise 5.2

Rework Worked Example 5.1, assuming that the temperature coefcient of resistance is always negative.

Stability

The electrical properties of components vary with time, whether a component is in use or in storage, due to physical and chemical changes in the materials from which the component is fabricated. Another way of looking at these changes is to say that components age. Component ageing can be accelerated by applied stress. If a component is operated continuously at its full rated voltage, for example, the component’s value may change much more quickly than if the component were in storage unused. In some types of ceramic capacitors, for example, an applied voltage causes gradual changes in the crystal structure of the dielectric, resulting in a change in permittivity and hence a change in the value of the capacitor. Some other examples of stress that can accelerate ageing are: heat, which can speed up chemical changes; thermal cycling (repeatedly heating and cooling a component), which can cause joints to crack because of differential expansion; and ionizing radiation, which can disrupt the molecular and crystal structure of component materials. High-stability components have values that change comparatively little over time. Stability is expressed as a fractional change in value (usually in p.p.m. or %) over a stated time interval and under stated conditions.

There is, of course, a cost

There is, of course, a cost penalty in using a 16 V capacitor

penalty in using a 16 V capacitor for a 10 V application, but the

for a 10 V application, but the initial cost of the component may

initial cost of the component may be less important than the

be less important than the long-term cost of unreliability.

long-term cost of unreliability.

Component ratings

Electronic components have limitations on voltage, current, power dissipation, and operating temperature range. In some cases there may also be limitations that are more complicated such as rate of change of voltage. These limitations are known collectively as ratings. Component manufacturers usually state two sets of ratings for their products: an absolute maximum rating, beyond which the component will be damaged or destroyed; and a recommended rating, which is the manufacturer’s statement of the component’s capability. To say that a capacitor has a recommended rating of 16 V does not guarantee that the capacitor will work as well at 16 V as it will at 10 V: it will almost certainly be more reliable at 10 V than at 16 V, and it may also be more stable. For these reasons, design engineers normally use a component well within its recommended rating.

Absolute maximum ratings may be important under fault or transient conditions. If a fault occurs in a system, other components will be undamaged if protective devices (such as an overvoltage trip circuit) operate before absolute maximum ratings have been exceeded.

Parasitic behaviour

So far in this discussion of passive component characteristics, we have looked at nonideal properties that are due to the physical limitations of materials and manufacturing processes. There is another way in which passive components can be nonideal, which is due to their electromagnetic behaviour rather than to the limitations of materials.

jω C

jω L

Reactance

v = i . R

(Ohm’s law)

i =

⋅

dv dt

v =

⋅

di dt

Terminal relation

viR

−

viC

−

viL

−

Resistance

Circuit element and symbol

Capacitance

Inductance

Figure 5.2 The resistance, capacitance, and inductance parameters.

In lumped-parameter circuit theory, there are three simple circuit elements: resistance, capacitance, and inductance. Figure 5.2 summarizes their properties. Practical realizations of all three circuit elements exist in the form of resistors, capacitors, and inductors, but in all three cases the practical component possesses a little of the other two circuits’ properties. (Note the sufxes -ors for a component and -ance for a circuit property.)

Figure 5.3a shows the construction of a metal lm resistor, made by depositing a metallic lm on the surface of a ceramic cylinder and then cutting a helical track into the lm to obtain the desired resistance value. The helical construction of the resistive track suggests inductance, and there is also capacitance between turns of the helix. Figure 5.3b sug gests a possible equivalent circuit for this type of resistor, but in reality the resistance, capacitance, and inductance of the component are physically distributed and not lumped as suggested. The series inductance,

Ls, and the parallel capacitance, Cp, are known as parasitic properties

of the resistor. In many applications, their presence can be neglected. At low frequency, perhaps in an audio circuit, the reactance of Ls is low and the reactance of Cp is high, so that the resistor behaves almost as a pure resistance. At higher frequencies, however, the impedance of the resistor is lower than at low frequencies as the reactance of Cp decreases. It is important to realize that the series inductance cannot be eliminated by making the resistor a straight bar, although it is reduced. This is because inductance is associated with the magnetic eld induced by a changing current, not with a helical or spiral conductor shape. Similarly, the parasitic capacitances are associated with the electric eld between two charged conductors, not with parallel at plates.

In a lumped-parameter circuit,

In a lumped-parameter circuit, we model the circuit elements

we model the circuit elements as if they were localized and

as if they were localized and connected by zero-impedance

connected by zero-impedance wires. Contrast this with a

wires. Contrast this with a distributed-parameter circuit

distributed-parameter circuit such as a transmission line

such as a transmission line where capacitance and

where capacitance and inductance are distributed

inductance are distributed evenly along the line.

evenly along the line.

This point is developed further

This point is developed further in Chapter 8.

in Chapter 8.

L =

× × × × ×

≅

− −

−

10 4 10 1 5 10

10 10

2 7 3 2

π π ( . )

Ceramic body with

helical track in

metal ﬁlm

Outline of insulating

coating

Metal end cap

(a)

(b)

Figure 5.3 A metal lm resistor: (a) construction and (b) a possible equivalent circuit.

Worked Example 5.2

Estimate the parasitic inductance of a metal-lm resistor of the type shown in Figure 5.3, if the ceramic body is 3 mm in diameter and the helical track consists of 10 turns spaced over a 10 mm length.

Chapter 8 of Compton (1990)

Chapter 8 of Compton (1990) discusses calculation of

discusses calculation of self-inductance.

self-inductance.

Solution

If we assume the ceramic has a relative permeability µr of 1, we can use an approximate formula for the inductance of a single-layer air-cored coil:

= N2 µ0A/l

where N is the number of turns in the coil, µ0 is the permeability of free space, A is the cross-sectional area of the coil, and l is its length. Hence,

Exercise 5.3

Using the parasitic inductance value calculated in Worked Example 5.2, nd out at what frequency the inductive reactance of such a resistor becomes more signicant than the resistance, for resistance values of (a) 10

Ω, (b) 1 kΩ, and (c) 100 kΩ.

(Answers: [a] 18 MHz; [b] 1.8 GHz; [c] 180 GHz [which is so high that lumped parameter circuit models would no longer be valid, so the result is meaningless].)

Figure 5.4 shows a possible equivalent circuit for a practical capaci tor. The series inductance, Ls, and resistance, Rs, are due to the wire leads of the capacitor, while the parallel resistance, Rp, is due to the leakage resistance of the dielectric. If the capacitor is of wound construction, there is an additional contribution to Ls. In reality, as with the resistor discussed earlier, the parasitic properties are distributed within

Q = ×2π

((Maximum energy stored in component)

Ennergy dissipated per cycle)

Z(ω)

R(ω)

jX(ω)

Z(ω)

R(ω)

jX(ω)

Figure 5.4 A possible equivalent circuit for a capacitor.

the capacitor to some extent. The properties of a real capacitor or inductor cannot therefore be dened analytically, and an empirical approach is taken. A capacitor or inductor is represented by a pure reactance, X, and a series resistance, R, both of which depend on frequency. The impedance of the component is then a function of frequency:

(ω) = R(ω) + jX(ω) (5.1)

X(ω) can be recognized as the normal capacitive or inductive reactance,

while R(ω) is known as the equivalent series resistance (ESR). Both depend on frequency. Several other quantities are derived from R and X, and all depend on frequency. The dissipation factor (DF) is the ratio of R to X and is also known as tan δ. The angle δ is called the loss angle, and its relationship to R and X is shown by the diagram in the margin. The reciprocal of the dissipation factor is known as Q, for quality factor. Q is dened as

(5.2)

A capacitor or inductor with a high Q factor, or low dissipation factor, absorbs little power when used in an a.c. circuit. In tuned circuits, it is not possible to realize a high Q for the complete circuit unless the inductors and capacitors have high individual Q factors.

Some manufacturers may state a power factor (PF), a term that is more often used in a.c. circuit theory, rather than a dissipation factor or Q factor. If the product of the root mean square (r.m.s.) current and r.m.s. voltage owing into a circuit is multiplied by the PF, the result is the power dissipation in the circuit in watts. In general, the current and voltage are not in phase, so that the product of their r.m.s. values (in volt–amperes, or VA) does not represent actual power dissipation.

Selection of a passive component for a particular application requires a knowledge of component characteristics in general, as introduced so far in this chapter, and also an understanding of the different types of resistors, capacitors, and inductors and their fabrication and characteristics, which is what the remainder of this chapter covers.

The concept of Q factor is also

The concept of Q factor is also used in resonant circuits and

used in resonant circuits and waveguide cavities with similar

waveguide cavities with similar meanings in terms of energy

meanings in terms of energy losses.

losses.

Power factor is covered by Kip

(1969), and is an important idea

(1969), and is an important idea in electrical power engineering.

in electrical power engineering.

Resistors

Resistors are used in electronic circuits for limiting current, setting bias levels, controlling gain, xing time constants, impedance matching and loading, voltage division, and sometimes heat generation. The resistance, R(Ω), of a material of length l and cross-sectional area A is given by

lAl

= =

Table 5.1 Characteristics of resistor types

Typical

tolerance

Type Precision

wire-wound

Precision

metal lm

Metal lm

Carbon lm Power

wire-wound

Thick-lm

networks

Surface

mount “chips”

* Dependent on rating.

(%)

0.1

5 5

temperature

coefcient

(p.p.m. °C–1)

< 10 < 1

±15

50–100

150–800 50–250 Up to

100

Power

rating

⅛, ¼,

⅛ – 1

⅛, ¼

(W)

⅛

600

Range

values

10 Ω – 100 kΩ

10 Ω – 1 MΩ

0.1 Ω – 1 MΩ

10 Ω – 1 MΩ

0.1 Ω – 10 kΩ*

10 Ω – 100 kΩ

1 Ω – 10 MΩ

Operating

temperature

range (°C) Features

–55 to +145 High stability, low

noise

–55 to +155 Low noise, good

stability

–55 to +155 Low noise, good

stability, lower cost than precision metal

lm –55 to +155 Low cost –55 to +250 Higher-rated types

may require heatsinks –55 to +125 Multiple resistors per

package –55 to +155 Thick or thin lm on

ceramic construction,

low inductance

At high frequencies, the

At high frequencies, the geometric form of a resistor

geometric form of a resistor must be considered and special

must be considered and special geometries such as discs may

geometries such as discs may be needed.

be needed.

(5.3)

where ρ is the material resistivity (Ω m), σ is the material conductivity (Ω–1 m–1), and l and A are expressed in metres and square metres respectively.

There are three main ways of fabricating a resistor, each applicable over a range of resistivities and resistance values (Table 5.1). First, if a material of suitable resistivity can be made, Equation 5.3 can be realized directly in terms of a slab or rod of resistive material with metal contacts at each end. If this method of construction is not feasible, the resistor can be fabricated from a longer length of material of thinner cross-section. One way of doing this is to wind a wire onto a cylinder, and the other method is to use a lm, or thin layer, of the resistive material, deposited onto an insulating substrate. Slab or rod resistors for surface mounting have the lowest series inductance of any resistor types because of the absence of leads. Wound resistors tend to have high series inductance, although this can be reduced to some extent by special winding techniques in which one part of the winding cancels the inductance of the remainder.

The temperature coefcient of resistance and the stability of a resistor are determined primarily by the properties of the resistive material used to fabricate the resistor. Most pure metals have temperature coefcients of resistance of around 4000 p.p.m. °C–1 and fairly low resistivities, making them unsuitable for use in resistors. Lower temperature coefcients of resistance can be achieved by fabricating resistors from proprietary alloys with temperature coefcients as low as ±5 p.p.m.

°C–1. Many of these alloys are based on nickel, chromium, manganese,

A thick-ﬁlm resistor network.

Common

A thick-ﬁlm resistor network.

Common

and copper. Well-known examples are the alloys known as nichrome (80% nickel, 20% chromium), constantan (55% copper, 45% nickel), and manganin (85% copper, 10% manganese, < 5% nickel) with tem

–1

perature coefcients of < 100 p.p.m. °C

, < 20 p.p.m. °C–1, and < 15

p.p.m. °C–1 respectively. Metal alloys are used in the fabrication of pre- cision wire-wound and metal-lm resistors. These types of resistor have the lowest temperature coefcients and the best stabilities of any resistor type. High-value wire-wound resistors are not feasible because of the fairly low resistivity of resistance wire. Nichrome wire of only 0.02 mm diameter has a resistance of 3.44 kΩm–1, and this is about the thinnest practical wire for production resistor manufacture; over 3 m of wire is required for resistor values above 10 k

Ω. The range of values that can

be obtained by varying the length and pitch of a helical track in a metal lm is limited, and higher-resistance values require either thinner lms or lms of higher sheet resistivity, which can be obtained by including nonconducting materials in the lm during deposition from a vapour.

For general purpose use where low temperature coefcients and

high stability are not essential, cheaper resistors can be fabricated using carbon as the resistive material. Carbon lm resistors are fabricated by pyrolytic decomposition of a carbon-containing gas, such as methane, in a furnace, depositing a carbon lm onto a ceramic or glass substrate. The resistors are then tted with end caps and leads, and coated with a protective varnish, lacquer, or plastic.

A third type of resistive material used in resistor fabrication is

known generically as cermet, for ceramic–metal. These materials contain nely divided metals distributed in a glassy vitried ceramic, which can be printed or painted onto a substrate in the form of a paste and then red to fuse the constituents into a hard solid. There are two important applications for these materials. One is in variable resistors or potentiometers where a low temperature coefcient of resistance (< 200 p.p.m. °C–1) is required. The other is in thick-lm hybrid circuits and thick-lm resistor networks. Thick-lm circuits consist of resistors and small-value capacitors printed and red onto a ceramic substrate. Surface-mounted ICs and transistors can be soldered to the circuit, which can then be coated with epoxy resin and used either as a complete selfcontained circuit or as a component on a conventional printed circuit board (PCB). Thick-lm resistor networks are made in the same way but contain only resistors. They are especially useful where several resistors of the same value are required in the same location on a PCB. A common application is a group of eight pull-up resistors connected to a microprocessor bus with a common connection to a supply rail. A thick-lm single-in-line (SIL) resistor pack containing eight commoned resistors has only nine terminals and occupies a smaller board area than eight discrete resistors.

A sheet of conductor has a

A sheet of conductor has a resistivity, measured in ohm

resistivity, measured in ohm per square (Ω◽ –1), which is a

per square (Ω◽ –1), which is a constant for any sized square

constant for any sized square of the sheet. Metal lms can be

of the sheet. Metal lms can be fabricated with sheet resistivities

fabricated with sheet resistivities of up to about 20 k Ω◽ –1.

of up to about 20 k Ω◽ –1.

Thick-lm circuits are discussed

Thick-lm circuits are discussed in more detail by Till and Luxon

in more detail by Till and Luxon (1982).

(1982).

Noise

All resistors generate electrical noise, or small random uctuations of voltage or current. Noise is not necessarily due to material imperfections in resistors, although some types of resistor are noisier than others:

Senturia and Wedlock (1993)

Senturia and Wedlock (1993) discussed noise in more detail.

discussed noise in more detail.

CRC Electronic Components and Technology User Manual

Specifications and Main Features

Frequently Asked Questions

User Manual

Contents

Preface to the Second Edition

Preface

Acknowledgments

Author