Tuesday, June 21, 2011

The Impact of One Man

This is going to be a departure from my usual fare - I don't think I've ever written anything quite like this. 

Have you ever wondered about the impact of one man on the world?  Just how much can the world be changed by one guy? 

Look around where you are.  Are you in front of a computer?  Laptop?  If you're reading this, you're using some form of electronics.  Do you have an MP3 player, like an iPod?  Do you listen to podcasts? 

The common thread I want to point out is analog electronics.  Not digital.  Not the computer itself but the sound system.  Although very few people are aware of this, the number of people who design the modern analog integrated circuits is shockingly small.   Two very bright stars in the history of analog IC design have both died unexpectedly in the last 10 days, and the engineers who know them by their work are in somewhat of a state of shock.

I don't want to get too off track, but I'll bet the vast majority of folks who stumble across this will know the names of the big name digital companies;  Intel, AMD, Texas Instruments (where the IC was invented).  You'll probably know Samsung, Fujitsu, Cypress or some other big players.  Digital is the hot technology, right?  Every generation, digital circuits get faster and more capable.  Everyone who has been around computers for more than a few years has stories of how much they spent for something that's cheap today.   I've had computers long enough to remember paying $350 for my first hard drive (30 Meg!) and $200 for my first megabyte of memory.  At a recent show, I was given a 1 Gig USB thumb drive as a promotion.  But have you heard of Fairchild Semiconductor?  Philbrick?  What about National Semiconductor, Linear Technology or Analog Devices? 

The analog world is a small piece of electronics, but indispensable.  Digital circuits are switches, and they function by rules of logic (Boolean algebra).  They work by switching between off and on at blazing speeds.  Modern circuits may only switch 1.8V or less, down from the 5V logic that was the norm when I was a kid - the lower voltage helps them speed up, and reduces power consumption.   Analog circuits, by contrast, are thought of as continuous operation.  While any signal below the high "turn on" level for a digital circuit (1.2V, for example) is the same, whole analog systems work with signals very far below that measured in millionths to thousandths of a volt.  Analog circuits power the digital world.  Essentially all modern systems use switching power supplies (also called switch-mode) to achieve the efficiencies we want in our electronics.  A simple power supply made of a transformer, rectifiers and filtering, may be 30% efficient; modern switchers can hit 90% efficient.  Think of the power saved!  (and the market did it on our own, without the .gov forcing us!  Imagine that...) 

One of the bright stars in analog IC design was a genius named Jim Williams.  Jim passed away unexpectedly June 12th after suffering a massive stroke.  Jim is best remembered for his years at Linear Technology, and they currently have a nice memorial page on their site.  I never met Jim in "meatspace", but I have read many of his applications notes describing the most intricate concepts clearly and concisely.  I think I first noticed Jim by name in the late 80s when I came across an applications note he authored called "Switching Regulators for Poets - a gentle guide for the trepidatious" (AN25).  An MIT guy, but not an actual graduate engineer, he is recognized widely in industry, including being inducted into the hall of fame.  Like all the great ones, he realized that filling a book page or whiteboard with math might describe the way the circuit worked, but didn't bring understanding; description and understanding are different things.

The second bright star, Bob Pease, died on the way home from Jim's funeral.  Bob first met Jim at Philbrick in the early days of analog ICs.  Bob cultivated his image as a curmudgeon, maintaining a "wizard beard", and at one time throwing computers off the National Semiconductor headquarters because the circuit simulation software "lied to me".  Again, I never met Bob in meatspace but I've read hundreds of his columns in Electronic Design magazine.  They have a tribute page up now, too, as does National Semiconductor.  Bob was involved in a lot of ground breaking design, but I think he was most proud of inventing the bandgap reference - still the state of the art approach for a voltage reference that varies the least with temperature - and started calling himself the Bandgap Czar.

In our modern, technical society, we tend to take things for granted.  We'll always have amplifiers and analog "black magic" electronics.  The community of people who actually do this is small, though.  I think it would surprise you.  National Semiconductor, where Bob and Jim both worked before Jim left for the upstart Linear Technology, has been acquired by Texas Instruments.  I'm not sure what will become of the name, but I wouldn't be surprised if it was gone in a year or two.  Of the 5700 employees at National, I'll bet there's less than 100 analog IC designers and one or two "staff scientists" who grok it all.  How many are there in the world?  Perhaps a thousand people on earth?   

Going back to my start, what's the impact of one man?  The philosophers argue that while a Jim Williams or Bob Pease may have invented much of modern circuitry, if they hadn't been there, someone else would have done it.  But how do we know?  Anyone who has done math has likely had the experience of struggling to solve a problem, only to have it look clear and obvious when the teacher solves it.  Oftentimes, the hardest part in solving a problem is seeing that it can be solved; and knowing a solution exists will drive someone on to find it.  I am always in awe of the great minds who do so much, and advance the standard of living for all of us - just because it's fun for them to do.  You'll see people saying that if we got through a really big SHTF event, that we'll be back to 18th century living.  We'll be surrounded by tech that no one knows how to use.  If the wrong thousand people die, society may "forget" how to design analog ICs. 

RIP, Jim, Bob.  Though we never met, you improved my life. 


LeverAction said...

I always enjoyed reading Bob's pieces in Electronic Design magazine where he frequently occupied the last page. I only wish I understood analog electronics half as well as he. In fact I've only ever known one man who did, and he himself is a curmudgeonly professor who's class it was my privilege to attend in college (it was required or I probably wouldn't have taken it because it was so difficult).

Men like Jim and Bob are an endangered species - who is coming up that understands the world of the electron like these men? Their contributions have shaped our world and hopefully we can keep what they taught and maybe in some way contribute a little as well.

Godspeed Jim and Bob. Though the world at large may not remember you, there are a handful of engineers who will never forget...

Graybeard said...

Well said.

Anonymous said...

Absolutely well said. Jim and Bob were two of the greats. I never met them in-person, either, but I have read probably all but just a handful of their application notes and articles. They definitely will be missed in the analog and even the digital world; there are no more of them around or being minted in the universities, in all likelihood. The passing of an era.