Bob Pease didn’t hate Spice simulations

Larry Lewicki, an analog IC designer has some interesting remembrances of Bob Pease. Larry relates:

• In the late 80s when I worked for National, I had a part in production for a couple years and the process (CP-88) was cleaned up and that reduced leakage currents.  Lo and behold - the bandgap refused to start at cold temperature.  Bob gave me the best advice for how to “SPICE SIMULATE A BANDGAP STARTER” all in his trademark capital letters. He would probably never admit to this. I used his methodology for years and mentored many a younger engineer in the black art of using Spice to evaluate startup circuits in bandgap references.

.

                Bob Pease at the EDN 2010 Analog Aficionados party (courtesy Fran Hoffart).

This confirms that, while Bob Pease did love having the reputation as a crusty old guy that hated Spice, he not only knew how to use Spice, he knew how to use it well. I remember writing an article about strain gauges. I will never forget when one of my sources, Tom Rummage, a senior application engineer at Vishay, told me stories about how engineers would look at some big broken part on the test stand and blame him, saying that he must have screwed up the test. He said these engineers kept pointing out how he must be wrong since the computer simulation showed the part would not break. These dilettante book-learned engineers would trust some pretty-colored finite element analysis rather that a part broken in two sitting on a bench.

What was going on was that the casting or welding processes used to make the part would add tremendous internal stresses as the part cooled off. Those loads are real, so when you add the design load the part is supposed to carry, it exceeds the structural strength of the metal.

What Bob Pease and Tom Rummage are saying is that you have to make computer simulations earn your trust. There is some weird human psychology where we tend to believe things on the computer screen more than we should. It is the same with calculators. You see 13 decimal places and figure that number is somehow better and trustworthier than something you did with an antique slide rule, even though your source numbers only had 3 significant digits to begin with. I have been 500 feet from my house as a passenger in a car and my pal would believe his confused GPS rather than me telling him “No. 237 West is THAT WAY.”

So if you want to honor the memory of Bob Pease, who passed away three days ago, all you have to do is use your head and try to understand the engineering principles behind the 13-digit numbers and pretty Spice plots and snazzy 3-D finite model renderings. It was this blind faith in computers that bothered Bob Pease, not computers themselves. Well, maybe. Larry Lewicki had another Pease story:

• It was in the mid 90s when I started working for Sing and Lee in the Data Acquisition group. I came into work one morning at about 8:15am. Lee was standing in the hall of building D holding his pipe.  Lee said he just needed to talk to someone, and proceeded to explain what it was like to be Bob’s manager.   That morning Lee had an urgent voice mail from National Semiconductor security. It seems that the previous evening a pile of surplus monitors that were stacked in the hallway of bldg D had gone missing.  I told Lee that, indeed, I remembered seeing the monitors when I left to go home the previous evening.  Apparently someone “matching Bob’s appearance” had gotten a group of people together and carried the monitors to the top of parking structure by Bldg. E.  Then a co-conspirator videotaped Bob throwing monitors off the top of the parking structure while yelling something like “these computers will never lie to anybody every again.”

Oh gosh I can just picture it. Bob was a beautiful man, willing to be wacky and subject himself to teasing, ridicule and even termination, all because he so passionately cared about teaching us some good common sense. I know what Bob is talking about. When I was at National there were two kinds of IC designers. One group was young folks that exclusively used computers and never ever would even do IC layout. These, well lets call them children, or perhaps even infants, felt that their job, the reason they should get paid, was to run a Spice simulation on the company-provided computer using company-provided models. If Spice showed the part worked. Well, they were done. Anything else was “not my fault”. These engineers seemed to feel their job was like playing a video game all day. They sat in front of the computer, played with Spice, and when things seemed to work, well, that meant they won the game, and could get paid and go home.

The other group of IC designers understood they where paid to get working silicon in production. They knew that Spice was a good place to start, but that it was just that- a place to start. You would see this group of IC designers in the lab when their silicon came out. I heard stories of Bob Widlar, how he was like a pit bull and mother grizzly bear around his bench. He would shoo off the managers and marketing types, and keep beating on his parts until he knew exactly how they worked. He and his kindred spirits would never blame the process, or Spice or Cadence or the tool group or the layout people or anything else. These IC designers took responsibility and used their brains.

That’s all Bob Pease and Tom Rummage want. Just use your brain, your common sense, and listen to your experience. I think it was Richard Feynman that said that the highest calling of a scientist is to prove himself wrong. So don’t just trust those pretty computer screens. Put a picture of Bob Pease in your cubicle. Whenever you are going to just toss some design over the wall to layout or manufacturing or process, look at Bob smiling down on you and think. Think about what is really going on in the circuit, in the silicon, and in the package. Bob did, and that is why he was a great man.

Oh, one other little remembrance Larry Lewicki had about Bob Pease:

• I first became aware of Bob in probably 1978 or ‘79 when I was working in Cedar Rapids for Rockwell-Collins.  There was an EDN article (that I wish I had kept) with a picture of Bob’s VW with stegosaurus plates on the roof. This was very funny for a guy in Iowa like me.  The article quoted Bob about northern Califonia earthquakes. Pease said something like “…of course it’s unstable here - look at the Transamerica Pyramid - it looks like an op amp and needs a compensation capacitor.”

So for Larry and the rest of you, here is that picture, courtesy of Fran Hoffart, applications engineer over at Linear Tech and former National Semiconductor coworker of Bob Pease.

              Bob Pease, standing next to his trusty 1967 Volkswagen Beetle (courtesy Fran Hoffart).

Paul Grohe, Bob Pease’s good friend noted a news item about Bob’s tragic death.

My obituary is here and here.

Please come remember Bob Pease and Jim Williams at EDN’s celebration of life, Tuesday June 21, 2011 at David’s Restaurant.