Ernie's Story

Finally, Ernie couldn't stand it any longer. "Jim, you're the president, and you've got to know", Ernie blurted, "There's some sort of glitch on the main processor board. It might be a crosstalk effect or some kind of noise problem. We just don't know. We'll have to hold up shipments until we get it all sorted out." There was a long pause while everyone's eyes turned towards Ernie. His manager began slowly sliding away from him, over to the other side of the couch.

Jim turned white, then red. He opened his mouth and closed it again. The veins began to stand out on his forehead. He hesitated, then he stood up, and then he thundered, "So that's what you guys are trying to tell me?" He was used to problems, heck, he handled a crisis almost every day, but this was absolutely unbelievable. "Let's see, gentlemen, we have 145 manufacturing people, 35 technicians, and a warehouse staff waiting for this product. Our sales team is trained, advertising is in the works, and I have four press interviews starting tomorrow. The company burn rate is about $100,000 per day. If this problem takes a week, we lose five hundred grand. That's more than you're worth, Ernie. Doesn't anybody have any brains around here? Why didn't you tell me about this before!! How could we get this far and not know there was a problem??? Ernie?!!!??!"

Ernie's lips trembled. Now he didn't feel so bold. He tried to make everyone understand, "Hey, look, I did everything I could. It's not my fault. The simulator said everything was perfect. The timing was all checked out. How was I supposed to know we'd have crosstalk? How was I supposed to know that fast signals wouldn't go through that connector? "They don't even teach crosstalk and ringing in college. Those are the subjects we're supposed to learn through experience. That's just the way it works. We were unlucky, I guess."

It's a cruel joke how our educational system has failed us. Twenty years ago computer designers all went through a common electrical engineering curriculum. This standard curriculum included basic analog circuits, transmission lines, and linear systems theory--all the little details that make high-speed digital hardware really work. Trouble is, the computer hardware of that era was so slooowwww that few people needed to know anything about analog circuits to make their systems function.

Take a look in the first edition Texas Instruments logic catalog (if you can still get your hands on one). The typical LS-TTL logic gate had a rise/fall time of about 20 nanoseconds. By today's standard, that is very slow. It doesn't take an analog guru to plug together a lot of LS-TTL logic. As a result, over time colleges and universities felt they could safely drop the old analog curriculum requirements in favor of newer, more modern computer science classes. There were even those who said that analog design was becoming irrelevant. Not that I have anything against computer science, mind you. It's a wonderful discipline. Without it we wouldn't have the processor sophistication we have today, or have nearly as many people trained to work with complex computer architectures.

Neither can I blame the educators for their decisions. A modern university has a lot on its plate. Information overload is a real problem. There are too many subjects to teach, and too few hours available in which to teach them. The result, though, has been near disaster for many high-speed design projects. In today's world, modern processors are clocked at 100+ MHz. Plain-vanilla chip outputs sport sub-nanosecond rise times. Analog effects are beginning to dominate system design. Digital designers without basic analog training are operating at a serious disadvantage.

In reaction to this trend, some colleges and universities have begun re-introducing analog material in their computer science curriculum. For example, I particularly applaud the efforts of Clayton Paul at the University of Kentucky, and Henry Ott of Bell Labs to educate digital engineers about the tough new world of electromagnetic compatibility (EMC) standards.

The faster we go, the more important these analog effects will become. Other courses of importance to digital engineers include (1) Basic circuit theory, (2) E&M waves, and (3) transmission lines. A well-rounded digital engineer should understand enough about circuit theory to recognize in what ways two adjacent digital signals will behave like a loosely coupled, single-turn transformer. He or she should know enough E&M wave theory to understand why a ground plane is better for high-speed boards than a loose mesh of ground traces. And every digital engineer should have a good grasp of signal propagation, superposition, and reflections on transmission lines. A proper understanding of basic signal integrity as well as EMC is crucial to the continuing evolution of digital technology.

Engineers with a solid background in signal integrity and EMC will greatly enhance their chances for career success. Engineers without a basic understanding of high-speed effects will likely end up just like Ernie, sitting in somebody else's office, fidgeting and sweating.

Signal Integrity Training Classes taught exclusively by Dr. Howard Johnson for full schedule, see www.sigcon.com © 1993-2007 Signal Consulting, Inc. All rights reserved.