Preface

This is a book of lore.

The word evokes many images: Merlin singing spells under the oak trees, a 'San bushman drinking from a hidden sip well through a reed, an angler trying to hook the wisest old trout in the lake. What I mean by it is altogether more homely: a mixture of rules of thumb, experience, bits of theory, and an indefinable feeling for the right way to do things, a sort of technical taste. It is what makes the difference between merely being able to analyze a design once completed, and being able to rapidly synthesize a good design to fit a particular purpose. Course work and textbooks teach analysis reasonably efficiently, but most contain no lore whatsoever.

One of the odd things about lore is that it lives in the fingers more than in the brain, like piano playing. Often in writing this book, I have run up against the difference between how I do something and how I think I do it, or how I remember having done it once. Since it's the actual lore of doing that is useful, I have where possible written or revised each section when I was actually doing that task as part of my job, or was consulting with someone who was doing so. I hope that this gives those sections a sense of immediacy and authenticity.

Apologia
Lore is usually acquired slowly through experience and apprenticeship. Beginners pester experts, who help fairly willingly, mostly because they're kept humble by stepping in potholes all the time themselves. This mutual aid system works, but is slow and unsystematic. As a beginner, I once spent nearly six months trying to get a laser interferometer working properly, a task which would now take about a week. The reason was a breakdown in the apprenticeship systemeveryone consulted said "Oh, that comes with practice"--perfectly true, and by no means unsympathetic, but not too helpful. Conversations with many others in the field indicate that this experience is the rule and not the exception.

This book is an attempt to provide a systematic and accessible presentation of the practical lore of electro-optical instrument design and construction--to be the book I needed as a graduate student. It is intended for graduate students at all levels, as well as practising scientists and engineers: anyone who has electro-optical systems to build, and could use some advice. Its applicability ranges from experimental apparatus to CD players.

The very attempt reeks of hubris. The range of topics covered here is enormously broad, and I do not pretend to be master of it all; it presents the work of a huge number of others. Lack of space and defect of memory prevent my adequately acknowledging even those contributions whose source I once knew. My only defence is that a volume like this can erect bridges between subdisciplines, prevent common mistakes, and help all those working on an instrument project to see it as a whole. So much good stuff gets lost in the cracks between physics, electrical engineering, optical engineering, and computer science, that a salvage attempt seemed justified. I ask pardon of those whose work has been inadequately acknowledged, and hope that they can remember once needing a book like this.

Mission
Designing and constructing electro-optical instruments is without a doubt one of the most interdisciplinary activities in engineering. It makes an absorbing and rewarding career, with little danger of growing stale. On the other hand, the same interdisciplinary quality means that instrument building is a bit scary, and keeps us on our toes. The very broad range of technologies involved means that at least one vital subsystem lies outside the designer's expertise, presenting a very real danger of major schedule slippage or outright failure, which may not become apparent until very late in the project.

We in electro-optics rely on whatever subset of these technologies we are familiar with, together with a combination of outside advice, collaboration, and purchased parts. Often, there are many ways of reaching the goal of a robust, working system; then the problem is where to start, among a range of unfamiliar alternatives. It's like the classic computer game Adventure: "You are in a maze of twisty little passages, all different." Some judicious advice (and perhaps a map left by a previous adventurer) is welcome at such times, and that's what this book is about, the lore of designing and building electro-optical instruments that work.

In order to have confidence in an instrument design, we really need to be able to calculate its performance ahead of time, without needing to construct an elaborate simulation. It is a nontrivial matter, given the current fragmented state of the literature, to calculate what the resolution and SNR of a measurement system will be before it is built. It's not that there isn't lots of information on how to calculate the performance of each lens, circuit, or computer program, but rather the complexity of the task, and the very different ways in which the results are expressed in the different fields encountered. For example, what is the effect of fourth-order spherical aberration in the objective lens on the optimal band-setting filter in the analogue signal processor, and then on the signal to noise ratio of the ultimate digital data set?

Somebody on the project had better know that, and this book aims to make you that somebody. It is intended in the first instance for use by oppressed graduate students in physics and electrical engineering, who have to get their apparatus working long enough to take some data before they can graduate. When they do, they'll find that real-world design work has much the same harassed and overextended flavour, so in the second instance, it's intended for working electro-optical designers. It can be used as a textbook in a combined lecture-laboratory course aimed at graduate students or fourth-year undergraduates, and as a self-teaching guide and professional reference for the working designer.

Organization
Textbooks usually aim at a linear presentation of concepts, in which the stuff on page n does not depend on your knowing pages n+1...N. This is very valuable pedagogically, since the reader is initially unfamiliar with the material and usually will go through the book thoroughly, once, under the guidance of a teacher who is presenting information rapidly. Reference books are written for people who already have a grasp of the topic, but need to find more detail or remind themselves of things dimly remembered. Thus they tend to treat topics in clumps, emphasizing completeness, and to be weak on explanations and on connections between topics.

Those two kinds of presentation work pretty well in some subject areas, but design lore is not one of them. Its concepts are not related like a tree, or packed like eggs in a crate, but rather interlinked like a fishnet or a sponge; thus a purely linear or clumped presentation of lore is all but impossible without doing violence to it. Nonetheless, to be any use, a lore book must be highly accessible, both easy to work through sequentially and attractive to leaf through many times.

The book is organized into four main sections: Optics; Electronics and Signal Processing; Special Topics (front ends, thermal design, and system integration and debug); and Digital Stuff (digital signal processing and software) and an appendix which is mainly about where to find more information. The material is presented in varying levels of detail. The differences in the detail levels reflect the amount of published lore and the measured density of deep potholes that people fall into. For example, there are lots of potholes in software, but weighty books of relevant advice fill shelf after shelf, and anyway software problems aren't usually what cause instrument projects to failunexamined assumptions, inexperience, and plain discouragement are. To get the software job done, we talk instead about how to code something simple that works, gives the right answer reliably, and can be maintained easily by non-experts.

Computer scientists use the concept of locality of reference--it's a good thing if an algorithm works mainly with data near to each other in storage, since it saves cache misses and page faults, but all the data has to be there, regardless. That's the way this book is organized: most of the lore on a particular topic is kept close together in the book for conceptual unity and easy reference, but the topics are presented in a sufficiently linear order that later chapters build mainly on earlier ones, and important connections are noted in both forward and backward directions. A certain amount of messiness results, which (it is to be hoped) has been kept close to a minimum.

The one big exception to this general scheme is Chapter 1, Basic Optical Calculations. It pulls in strands from everywhere, to present the process and the rhythm of conceptual design, and so will contain things that many readers (especially beginners) may find somewhat unfamiliar. Don't worry too much about the technical aspects, because there's more on all those things later in the book, and pointers to other sources. (Teachers may want to leave this chapter until late in the course.)

Although a really complete instrument design course would probably have to wait for a first or second year graduate class, the material in Chapter 1 through Chapter 10 could be taught selectively as a fourth year undergraduate survey course in optical instruments, following the usual upper-level electromagnetics and physical optics classes and a really solid grounding in Fourier transforms, at the level of Bracewell. In order to get the most out of it, the audience should be people with instruments of their own to build, either in a lab course, a senior project, or a part of their graduate work. Because of the complicated, interdisciplinary nature of instrument building, the laboratory part of the course might best be done by teams working on an instrument project rather than individually, provided each designer knows enough about everybody else's part to be able to explain it.

Exercises

Making complicated tasks intuitive is the true realm of lore--knowing the mathematical expression for the fringe pattern of a defocused beam is less useful than knowing which way to turn which knob to fix it. The most powerful method for gaining intuition is to use a combination of practical work and simple theoretical models that can be applied easily and stay close to the real physics.

Accordingly, the emphasis in the problems is on extracting useful principles from theory and discussion. Most of the problems have been taken from real design and scientific work, and so tend to be open-ended. Most students will have had a lot of theoretical training, but nowadays most will not have the skills of a Lightning Empiricist, a gimlet-eyed designer who's fast at mental rule-of-thumb calculations and who sanity checks everything by reflex.

Perhaps this book can help fix that.

Errata

A certain number of errors and misconceptions--hopefully minor--are bound to creep into a book of this type, size and scope, unfortunately. I welcome your comments and corrections, large and small: errata and omissions will be periodically posted at http://www.pergamos.net and and the John Wiley FTP site. Send email to hobbs@pergamos.net

Acknowledgements

To acquire lore, one needs a big sandbox, and long uninterrupted stretches of time to spend there, absorbed in the play. I am forever grateful to my parents for providing that sort of environment in my growing up, and for believing in me even when only the mess was visible.

I learned most of this material through participating in the stimulating and supportive technical cultures of the places where I've been fortunate enough to study and to work: the Edward L. Ginzton Laboratory at Stanford University, Stanford, CA; the departments of Physics and of Geophysics & Astronomy at the University of British Columbia; Microtel Pacific Research in Vancouver, B.C., and IBM Thomas J. Watson Research Center at Yorktown Heights, New York. I owe a special debt to IBM and to my managers there, Arthur Ciccolo and John Mackay, for supporting this project and for generously allowing me time and resources to work on it.

I also wish to thank the many gifted men and women I have been privileged to have as close colleagues, teachers, and friends, particularly J. Samuel Batchelder (who first suggested I write these books), Kurt L. Haller, Gordon S. Kino, Brian A. Murray, Martin P. O'Boyle, Marc A. Taubenblatt, and Theodore G. van Kessel. Without them, I'd still be stuck in one of those potholes way back along the road.

Most of all, I wish to thank my wife Maureen and our children Bronwen, Magdalen, and Simon, for their patience and encouragement while I wrote and wrote.