Backplane Design

Don't believe every myth you hear. For example, RF radiation from computers does not kill either brain cells or hair follicles. To directly test this theory, you might check out the pate on your local EMC specialist, who probably spends as much time as anyone with his head next to (if not inside) a computer, but I must warn you that such an unscientific sampling will likely lead you astray. The reason the "sample-of-one-EMC-guy" method won't work is because most EMC engineers are already bald for entirely unrelated reasons.

What I know about baldness and EMC is the result of painstaking research conducted in 1999. This research consisted of standing for ten minutes in the back of an IEEE EMC chapter meeting in Seattle waiting to make a brief presentation. From that vantage point I could easily study the levels of cranial exposure. My observations corroborated the existence of a pronounced "baldness effect," which I initially assumed was due to some little-understood RF principle. As I approached the podium, however, turning to confront my audience for the first time, the inescapable truth finally dawned on me. The short-sleeved white shirts, the coffee-stained ties, and the pocket-protectors all drove one common thought into my mind: Most of these guys are nearing retirement. To put it bluntly, you'd see the same number of bald guys at a meeting of World War II veterans. The problem isn't radiation—it's age.

This revelation leads me to a very interesting question. Why is it that digital engineers last only five to seven years on the job before going obsolete, whereas EMC professionals carry on for decades? The answer to this question has more to do with what's inside the head than on top: EMC engineers spend a lot of time educating themselves. They have to. The whole field of electromagnetic compatibility, specifically the limits on radiation from computing devices sold or used in the United States, was cooked up by the FCC in 1982. Since then, it's grown into quite a fast-changing, burgeoning business. Engineers who don't keep up go obsolete. That means switching into marketing, sales, or administration. Most EMC folks understand that if you want to keep doing what you love to do, you must constantly re-educate yourself.

This reason is why the EMC society hosts such great IEEE chapter meetings, and it explains why so many people write books for the EMC audience. Even though there are fewer EMC engineers than digital designers, EMC engineers buy more books and take more classes. As a result, the industry enjoys a number of terrific books and courses on the subject. Some examples of these excellent books are Clayton Paul's Introduction to Electromagnetic Compatibility, ISBN 0-471-54927-4; Henry Ott's Noise Reduction Techniques in Electronic Systems, ISBN 0-471-85068-3; and Douglas C Smith's High Frequency Measurements and Noise in Electronic Circuits, ISBN 0-442-00636-5. In the EMC training department, don't overlook the courses by Bill Kimmel and Daryl Gerke at

On the topic of training materials, let me tell you about a CD-ROM tutorial I saw this week that flipped my bits. The CD-ROM, entitled EMCT Tutorial, is written by my favorite EMC guru, W Michael King of Costa Mesa, CA. I was lucky enough to work with King early in my career. He took me out to the test range and delivered some very effective personal tutorials on "how EMC works" and "why you should have done what I told you to do in the first place." If you can't get King to personally train you, the next best thing is to check out his CD-ROM, available from Especially for those of you who respond well to pictures and moving diagrams, EMCT Tutorial is a very high-quality offering. It covers a wide range of pc-board design issues of great interest to anyone involved in EMC compliance.

I hope this training discussion helps you understand the key reason why most digital engineers, in contrast to their EMC colleagues, have such a short half-life. Don't let it happen to you. I plan to keep reading and keep learning for a long time, whether my head's covered or not.