I’m sitting in a grassy meadow, surrounded by snowy mountain peaks, at the home of Dr. Howard Johnson, author of High-Speed Digital Design: A Handbook of Black Magic. It’s so quiet up here in the Cascade Mountains of Eastern Washington State that I can hear my own heartbeat. The only other sounds come from the horses, contentedly munching on the tall prairie grass. If there is any magic left in the world, surely there is some here.
“Dr. Johnson,” I begin, “You are an author, a columnist, and an international lecturer. What brought you here?”
“Please,” he replied, “Call me Howie. That’s what all my friends call me.
“As to how I got to this place, that story begins, as do the stories of many engineers I have met, in my youth. Curiosity was always my strongest impulse. I wanted to know everything. I remember discovering at an early age that some old farts hold secrets, powerful secrets, and that if you ask in a sincere way they will tell you what they know. That is the key to learning.
“My first great source of knowledge was, of course, my Dad. His brain held a treasure trove of information about basic science and, especially, physics. The only problem with his knowledge was that it related mostly to big government projects, like rockets, satellites, and missiles. Observing the ups and downs of our family finances during the 1970’s, I resolved that my career should not depend on government funding.
“As a teenager, Dad introduced me to the electrical engineers that worked at his laboratory. They were a jolly bunch, and some were quite skilled at shooting skeet, a sport in which I heartily indulged if only to have more opportunities to hear technical tidbits of their various projects.
“By the age of 15, I felt well-supported in my career aspirations by the men at the local ham radio club, who taught RF principles, the guys at my Dad’s lab, who gave me many electrical parts and a scope with which to investigate their behavior, and a man named Thibodaux (Tee’-bo) who worked at the TV repair shop where I landed a job as a tube tester.
“Thibodaux taught me a very important lesson: do not stick your head into the back of a color console TV without first discharging the display tube. This is especially important if you have a rather prominent nose. The mental image of Thibodaux’s proboscis unexpectedly receiving a discharge of 30KV, which caused him to rear his head, striking a nail on the inside top cover of the console, driving his head back down for a second, third, and God knows how many other discharges, until finally subsiding, has stuck with me all these years.”
Pulling another beer from the ice tub, I asked, “You had a mentor at some point, is that right?”
“Yes. That was Professor Martin Graham, retired, from U. C. Berkeley. We worked together off and on for twenty years. Marty held an unbelievable repository of historical knowledge about computer architecture—he lived through most of it—and it became his mission to pass it on to me. When the time came for me to take on the role of ‘old fart’ I decided to write down, as best I could, the lessons Marty taught. That is the inspiration for all my books. I hope that his personality, tempered as it is by the limitations of my writing style, still shines through.
“Marty was a trickster. He loved causing trouble, especially with management. He also liked to look at a complex, intractable problem and pinpoint the one, main, most important thing you could do to instantly fix everything. That was his genius, focusing on the main effect and ignoring all else.
“That’s a common thread among great thinkers. My bookshelf holds many history books. Copernicus, Galileo, Newton, Faraday, Heaviside, Maxwell, Einstein, Bode; all these men perceived the simplicity of nature and communicated their vision to the rest of the world.
“My goal on every project, whether I’m working on voicemail, robotics, or Ethernet, is to identify one or two main, shining principles that carry the design. With those fixed, I then work out the myriad details necessary to bring the project to life.
“Too many engineers become lost in their design process. In the field of signal integrity, for example, they may begin to believe that their objective is to build a better SPICE model, or better correlate measurements with simulation. They forget that their overriding purpose is to build products that work, at an acceptable level of failure, at the minimum overall cost including the cost of their own design efforts. That’s the way management thinks about your job. If you want a career, you’d better think the same way.”
Among the hills behind the house, the sun slowly sets. The clouds turn all beautiful shades of purple and orange. I pause before asking, “So, what’s next with your career?”
Dr. Johnson took a long swig. “After working damn hard for a lot of years, with a lot of startups, I’ve moved beyond the need for a regular salary. As a result, I spend an increasing amount of time making mine a better community. Mostly, I do that through music. For example, on Friday nights you’ll find me at a local club playing upright bass in a jazz trio. Most small towns don’t have that. We do.
“Recently, I finished converting a barn into a 200-seat semi-outdoor auditorium. There we host an international chamber music concert each summer.
These projects satisfy my need for community involvement, but not my need to pass on the knowledge that Marty shared with me. I’ll probably always respond to people who come up to the ranch and ask, sincerely, about high-speed electronics.
Dr. Johnson responded, “Oh. Marty told me that in 1979. He pointed out that the whole computer industry moves in great cycles. He lived through four: electromagnetic relays, vacuum tubes, discrete transistors, and now integrated circuits.
Each new technology brings with it a new package, fundamentally smaller than the one before. That smallness confers a tremendous advantage in speed. At first, the new technology seems perfect. The parts snap together like LEGO blocks. As long as the package performance exceeds the operating speed of the underlying technology, you needn’t know much about packaging to make it work.
As the technology matures, it eventually outstrips the performance of its packaging. The focus of engineering then shifts from a concentration on just using the technology to an interminable struggle against the physical and electrical limitations of the packaging. Problems with signal quality, crosstalk, power supply noise, ground bounce, and heat dissipation stymie further progress.
Large-system design is at that point now. Marty would say that integrated-circuit technology, now getting a bit long in the tooth, is due for a fundamental revision. What that revision will be, I don’t know. Terry Morris at HP thinks we need optical backplanes. Joe Fjelstad thinks we need ways to make smaller, denser PCB’s.
I think that the answer, when it comes, will surprise all of us. When that happens, a fundamental understanding of physical principles will serve you well. Keep reading, keep learning, and keep an ear to the ground for new technologies. ”