Fall of 1968 found me at the Massachusetts Institute of Technology (MIT), preparing courses, negotiating students' theses topics, and assembling a laboratory. My activities were fairly unremarkable behavior for the locale, but, for a 20-year-old college dropout, the circumstances were charged: This was my one shot at any sort of career. For reasons I'll never understand, my entire education, from kindergarten through college, had been a nightmare, perhaps the greatest impedance mismatch in history. When I got hot, the Detroit Board of Education didn't. Leaving Wayne State University after a dismal year and a half seemed to close the casket on my circuit-design dreams.

All this history conspired to provide me with an unusual outlook, a mix of terror and excitement. But mostly terror. Here I was, back in school, but on the other side of the lectern. Worse yet, my research project, although of my own choosing, seemed open-ended and unattainable. I was scared to death. The capper of this scenario was my social situation: I was younger than some of my students, and my colleagues had at least 10 years over me.

The architect of my unique opportunity was Jerrold R Zacharias, eminent physicist, Manhattan Project and Radiation Lab alumnus, and father of atomic time. Zacharias had waved the magic wand to get me an MIT appointment, a lab, and operating money. He also made it clear that he expected results. He was not the sort to tolerate looking foolish, and, in my mind, to fail him promised a far worse fate than dropping out of school.

Against this backdrop, I received my laboratory-budget request back from review. Zacharias permitted me untrammeled freedom. Everything I requested, even very costly items, had been approved without comment or question. He included just one restriction: no allocation for instrument repair and calibration. His handwritten comment read, "You fix everything."

It didn't make sense. Under pressure for results, scared to pieces, I was supposed to waste time by screwing around fixing lab equipment? I went to see Zacharias. I negotiated. I pleaded. I ranted. But I lost. The last thing I heard chasing me out of his office was, "You fix everything."

I soon cooled off, and the issue became irrelevant because nothing broke. At least for a while. Finally, a high-sensitivity, differential-scope plug-in, a Tektronix 1A7, died. Life would never be the same.

The problem wasn't particularly difficult to find, once I took the time to understand how the thing worked. The manual's level of detail and writing tone were notable; communication was the priority. This seemed a significant deviation from most academic publications, and I was impressed. The instrument more than justified the manual's efforts. It was gorgeous. The integration of mechanicals, layout, and electronics was like nothing I had ever seen. Hours after I fixed the thing, I continued to probe and puzzle through its subtleties. A common-mode bootstrap scheme was particularly interesting; it had direct applicability to my lab work. I resolved to steal the techniques for reducing input current and noise.

Over the next month, I found myself continually drifting away from my research project to take apart test equipment and see how it worked. The practice, alone, was interesting, but what I really wanted was to test my understanding of a piece of equipment by having to fix it. Unfortunately, Fluke, Hewlett-Packard, Tektronix, and the rest of that ilk had done their work well; the stuff didn't break.

I offered free repair services to other labs that would bring me instruments to fix. I had few takers. People had repair budgets and were unwilling to risk their equipment to my unproven care. Finally, in desperation, I paid people, in standard MIT currency (Coke and pizza), to deliberately disable my test equipment so I could fix it.

A few of my students became similarly hooked, and we engaged in all forms of contesting. After a while, the "breakers" developed an armada of incredibly arcane diseases to visit on unsuspecting working instruments. The "fixers" countered with ever more sophisticated analysis capabilities. Various games we created took points off for every test connection made to an instrument's innards, the emphasis being on how close you could get utilizing panel controls and connectors. Fixing without a schematic was highly regarded-a macho test of analytical skill and circuit sense. Still other versions rewarded pure speed of repair, regardless of method. It was great fun. It was also a form of efficient--and serious--education.

The inside of a broken, but well-designed, piece of test equipment is an extraordinarily effective classroom. The age or purpose of the instrument is of minor concern. Its instructive value derives from several perspectives.

It is always worthwhile to look at how a designer dealt with problems using available technology-and within the constraints of cost, size, power, and other realities. Whether the instrument is three months young or 30 years old has no bearing on the quality of thought behind it. Good design is independent of technology, essentially timeless. The clever, elegant, and often interdisciplinary approaches found in many instruments are eye-opening, often directly applicable to your current design work. More important, they force self-examination and, with some luck, prevent rote approaches to problem solving (and the attendant mediocre results). The specific circuit tricks you find are certainly useful and adaptable but not nearly as valuable as studying the thought processes that produced them.

The fact that the instrument is broken provides a unique opportunity; a broken instrument (or whatever is at hand) is a capsulized mystery, a puzzle with a definite and very singular "right" answer. As a result, you are forced to measure your performance against an absolute, nonnegotiable standard. When you're finished, the thing either works or doesn't work.

The reason this scenario is so valuable is that it brutally tests your thinking process. Fast judgments, glitzy explanations, and specious, hand-waving arguments cannot be costumed as "creative" activity or true understanding of a problem. After each ego-inspired lunge or jumped conclusion, you confront the uncompromising reality that the damn thing still doesn't work. The utter closedness of this reality prevents you from fooling yourself. When it's finally over, when the box works-and you know why-then the real work begins. You get to try to fix yourself. Poor technique, crummy arguments, and inaccurate conclusions all demand review. It's a humbling and sometimes embarrassing process, but valuable nonetheless. You learn to dance with problems instead of trying to mug them.

It's scary to wonder how much of this sort of sloppy thinking slips into your own design work. In that arena, the system is not closed. There is no arbitrarily right answer, only choices. Things can work, but not as well as they might if your thinking had been better. In the worst case, things work but for different reasons than you count on. This situation is a disaster and more common than might be supposed.

For me, the most dangerous point in design comes when it works. Ostensibly, this "proves" my thinking correct, which isn't necessarily the case. The luxury the broken instrument's closed intellectual system provides no longer exists. In design work, results are open to interpretation and explanation, which is a dangerous time. When a design "works" is a very delicate stage; psychologically, you are ready for the kill and, consequently, less inclined to continue testing your results and thinking. That's a precarious place to be, and you have to be careful not to get into trouble. The very humanness that drives you to solve a problem can betray you near the finish line.

What all this means is that fixing things is excellent exercise for doing design work, a sort of bicycle-with-training-wheels that prevents you from getting into too much trouble. In design work, you have to mix your willingness to try anything with what you hope is critical thinking. This seemingly immiscible combination can lead you to a lot of nowheres, but it can also force you to learn, which is the major reason I've been addicted to fixing since that semester back in 1968. I'm fairly sure it was Zacharias' reason for bouncing my instrument-repair allocation. I couldn't understand it then, but he had initiated me. He introduced me to what my life would become for the next 10 years. And no apprenticeship was ever more necessary, better delivered, or, years later, as appreciated.

There are, of course, less lofty adjunct benefits to fixing. You can often buy broken equipment at absurdly low cost. I once paid $10 for a dead Tektronix 454A 150-MHz portable oscilloscope. It had been systematically sabotaged by some weekend-bound calibration technician and tagged "beyond repair." The machine required 30 hours to uncover the various nasty tricks played within its bowels to ensure that it would be scrapped.

This kind of devotion highlights another benefit of fixing. There is a certain satisfaction, a kind of service to a "moral" imperative, that comes from restoring a high-quality instrument. Sure, I'll admit that this is unquestionably a gooey, hand-over-the-heart judgment, and I confess a long-term love affair with instrumentation. But, for me, it seems sacrilegious to let a good piece of equipment die.

And, finally, fixing is simply a lot of fun. I'm probably the only person at an electronic flea market who pays more for the busted stuff than for the equipment that works!

Oh boy, it's broken! Life doesn't get any better than this.

This article is part of Jim Williams' new book, Another Look at Analog Circuit Design. It is published by Butterworth-Heinemann as part of the EDN Series for Design Engineers. Contact (800) 366-2655 to order.