The 1978 comedy science-fiction work Hitchhiker’s Guide to the Galaxy describes a computer that was specially constructed to determine the “Answer to the Ultimate Question of Life, The Universe, and Everything.” After 7.5 million years of processing (remember, this was written in 1978 when computers were pretty slow), the computer determined that the answer was "42." However, the computer was unable to explain that response to its creators and would need to build a new computer (called "Earth") to compute the explanation.
At my previous employer, we had a designated New Product Coordinator. I won’t embarrass her by using her real name, but it’s the same as a popular flower. I’ll call her “Violet.” Her job was to keep track of where new-product lots were in the manufacturing line, and keep them moving so we’d stay on schedule. She had to work with all aspects of manufacturing to make that happen, including splitting lots for process corners, holding lots at poly or metal for possible changes, etc. She also had to pester the engineers to finish their simulations, then characterizations, etc., in preparation for our weekly new-product reviews.
Management relied on Violet to push the development process hard, to speed it up. She kept the schedules, had all the commitment dates, and had the full backing of all the managers when she asked why someone missed a deadline.
Many an engineer, especially among the younger ones, would get his first chip back from assembly, run to his bench, put a chip in a test fixture, and see if it worked: a quick look for instability, turn up the clock or input signal frequency to see where the chip stopped working, load the output past the spec, then declare success to the rest of the lab -- usually to thunderous applause and high-fives all around. Then a quick write-up to distribute at the weekly review meeting.
Violet did not have an engineering degree and was not a design engineer, but she quickly learned the seven-word Ultimate Question of Analog Circuit Design that every one of our design engineers would learn to dread hearing at a review meeting:
“Did you look at it over temperature?”
Either on the bench with assembled parts or in simulation, the question needed to be asked and Violet knew just when to ask it. “Uhhh… no. But this is a commercial temp device, and I’ve used all these subcircuits and this process lots of times before. Nothing can go wrong” was not an acceptable answer, although some engineers tried it. Few made the same mistake twice.
Violet is retired now. But perhaps someone reading this will get the benefit her genius by asking himself The Ultimate Question before the next tapeout.
Tweet This
