Interview with Robert Ulrickson and John Nichols
April 19, 1995
Santa Clara, California

RW: Well, let's go do an interview. Well, we're here today in Santa Clara at Logical Services, Inc. or LSI with Bob Ulrickson and John Nichols, the founders. John, why don't you start off and tell us about your early pre-Fairchild days.

JN: Well, I'm an Okie from Muskogee, that's a little too early, isn't it? But I tried to find a job down in San Diego when I got out of college, but nobody down there would hire me so I ended up interviewing up at Philco Western Development Labs and Al Pound who ended up with some semiconductor outfit...

BU: GME?

JN: Yeah. He took me out in the lab and told me he wanted to see the signal on the oscilloscope and he hired me because I was the first one that showed up that actually knew how to use an oscilloscope. So, I worked there writing computer programs, designing IO equipment, doing systems design for satellite tracking stations in the mission control center in Houston, worked with a bunch of good people, George Byers, Paul Reif and lots of others but...

RW: And me. You didn't list me in there under the good people.

JN: Oh, yeah, yeah, yeah and Rob Walker too. We weren't really in the same group, were we?

RW: No, were for a little while.

JN: We were for a little while. But anyway, George Byers, he ran off to some other company and then he quit there and went to Fairchild. We keep running into each other periodically. And he finally, they wanted him to do some job and he didn't want to do it and they told him he had to go out and find somebody to do this job because he wanted to go up with Seids and work on the Semble II computer up in R and D. So he came over and asked me to do that and at the time my wife was pregnant with Lisa and I'd lose my medical insurance if I went to Fairchild, so I declined. And another fellow in our group decided to go to Fairchild named Paul Reif. And he was at Fairchild three or four months, something like that, just adopted a baby and married and he took this airline down to South Shore Tahoe and the dispatcher lied to him when he left here about what the weather at South Shore was and it ended up he couldn't, he didn't land, picked up a lot of ice, tried to get out of the canyon there, ran into the side of the mountain and killed everybody. So, that opened up the job after my wife had had Lisa and so I was available and I went to Fairchild and that's how I got to be at Fairchild.

RW: Well Bob, I know that you went to MIT and that you were chasing after Joan Baez, that's the part that I remember. So tell us about your early days.

BU: Well, I graduated MIT in 1959, EE, and went into, straight into the Coast Guard Reserve, Officer Candidate School; came out of there and was stationed at the Coast Guard Academy in New London, Connecticut. And then I got married and drove across country to Silicon Valley, which was not Silicon Valley at that time. But I accepted a job with Lockheed who was kind enough to put me through school for four years over here at Missiles and Space in their electronics organizations, and I began designing PCM telemetry systems for satellites, like the Gemini Agena Target Vehicle was my project. And I worked in a group there with a bunch of good guys, like John. Among those guys was Jim Kubineck and Tony Holbrook, they worked in what ended up being my group when I was a supervisor at Lockheed after about five years. Kubineck quit and went to this outfit called Fairchild Semiconductor and was hired by John Hume over there in the applications department. He was going to design some memories. And I got to looking around and talked to Jim and Jim said, hey, there's an opening over here. So I went into my boss and quit at Lockheed and was about to join Jim with Fairchild when Tony Holbrook who worked for me, walked into me and quit. And Tony said he was going to Philco to design some aerospace thing and I said, Tony, you can't do that. The semiconductor industry is where it's at. I called Kubineck and Kubineck got Tony a job at Fairchild. Tony recently retired. I guess he was president of AMD for a while.

RW: A long time.

BU: But I went over to work for John Hume in the applications department, which was what Fairchild called it. It was the engineering department that designed all the integrated circuits at the time. And that's how I got to Fairchild.

RW: Well then, how did you get into custom at R and D?

BU: Well, the job I was hired to do at first was for John Hume, but that was kind of a holding pattern. I helped him write the business plan for the integrated circuits division for the applications department. But really I was kind of in a holding pattern waiting for Bob Shriner to get started with a new department at Fairchild R and D that was destined to put out the first LSI chips. And interesting you should ask because at that time they were having these big meetings and I have a verbatim transcript here of a meeting that occurred at Fairchild that caused Fairchild to get into the LSI business. I call it the LSI story. It's dated September 8, 1967 and the attendees at the meeting were Bob Noyce, Tom Bay, Gordon Moore, Dr. Bob Sieds that John mentioned a minute ago, Jerry Sanders, Jerry Larkin, Maurice O'Shea, Bob Shriner, John Sentuse and John Hume. And it's a very interesting story. Talking about gate arrays which people thought was Gatorade at the time because nobody had ever heard of an array of gates that could be connected together to perform any digital logic function that you want. And Maurice had curves that showed that the number of these gate arrays was going to go off into the sky and hardly anybody believed that at the time. I think it turned out to be true. There was an exponential growth. But the interesting thing about that is all the discussion that went on in this seventy eight page document resulted in Bob Noyce making the decision to start this department with Bob Shriner as the department head and working for Bob Sieds at Fairchild R and D and the R and D division, of course, was run by Gordon Moore. And that decision was made and the department was started and then shortly after that a bunch of us moved up to R and D. About that time Shriner had been talking to this young engineer named Rob Walker who came into the department about a week after I did. I think if he'd come in a week before me, he'd been my boss instead me being his boss. And so we took Rob and a couple of other guys with us up to R and D and started this LSI department.

RW: Which ended up embedding standard cell and gate arrays. Just to put it in perspective, the first gate array was thirty-two gates, so we started small.

BU: The 4500 was it?

RW: 4500.

BU: I have a picture on my wall in the other room of the 4700, which was the next larger one, the ninety six-gate array. But what we left behind at that time was, I think, more important at that time, and that was the group that John and Clive had started in the Mountain View organization to do MSI, medium scale integrated circuits. John and Clive would come up with about sixteen digital functions that all worked together in a compatible way, the four-bit chip register, the adders, the decoders and so forth. And those things became what I believe is still the largest selling and most profitable product line the semiconductor industry ever saw. Those were picked up and used in every mini-computer ever built.

JN: In fact, I think their company started AMD.

BU: Yes...

RW: Just for that.

JN: Just for MSI. And as I recall, it seemed to me that Bob had told me he had sold Sanders on the idea of MSI in the...

BU: Well, that's an interesting story because John and Clive had been working for a long time to try to get somebody at Fairchild to listen to their ideas about these circuits and they had gone to John Carey and Carey said we don't want to do those big chips, we can't even build the ones we've got. And so one day, and I had been fighting the battle for quite awhile on their behalf, because I was hired in to be their supervisor, and I wrote this procedure for product planning; sent it all the way up to the president of Fairchild in Syosset, New York, got it approved and then it came back down and it said, okay, the applications manager and the product marketing manager and somebody from sales and somebody from the factory could approve new products that were submitted on this form and we submitted a bunch of these products, six of them I think, to start with and we couldn't get the factory to sign off. And we were all very frustrated because we felt like this was the next generation of products that would be produced by Fairchild. So, at that time, Fairchild had a cafeteria in Building Three, and they served beer in the cafeteria, which was just kind of rare. We don't do that in companies much anymore and I sat down at the table and I sat across from Jerry Sanders. Jerry was the Military Marketing Manager. For some strange reason, the product marketing manager for all digital integrated circuits was Floyd Quamy, who reported to the Military Marketing Manager and then he had a number of guys like Dick Coors and Dick Bader and Dan Hower and a number of other guys reporting to him as product marketing managers for these different integrated circuit lines. So I said, Jerry, we can't get these guys to listen to us. So I told him the story about the products and how important they were and so forth. Jerry stood up, regally, and said, this conversation will not go unheeded. The next day we called the first product-planning meeting and approved the first six MSI products. A few years later AMD was founded based on those products and a set of linear products.

RW: And those basic organizations live on today, essentially unchanged; the pin-outs...

JN: You can go down to Fry's right now and buy a 74-195, which is the equivalent of the 9300, which was the first product.

RW: The designs were so elegant that they lived through various technologies, the Shocky, Shockley, Shocky and the low power and without change and they're produced all around the world.

BU: Even the CMOS products that first came out at MSI-size before it got to be an LSI phenomenon. The very first functions they chose to put in those were the 9300 Series MSI functions, which were renamed the 74-1 something or other. TI won the name contest, they had the numbers that were popular, but Fairchild's MSI's were the products that really made the product line.

JN: The architecture.

RW: Yeah, that won.

JN: Actually, it was obvious.

RW: John, that application group, that central application group was so strong, so fabulous, the people in it.

JN: What was the name of the guy that ran the sail board shop in Los Gatos?

BU: Boy, you got me on that one.

JN: He came by and saw us, he was in that applications group. Who else was in the applications group? He was a chemist.

BU: Don't remember.

RW: But that, I remember, before coming to Fairchild, reading those application notes and then using that as my guideline and they were just so valuable to me. It was incredible.

BU: Yeah, Fairchild put out more of these green binders than any other semiconductor company did. We had books for data sheets, books for application notes. I'll never forget John's first application note on the 9300. That's how we got to be friends. I came in and he didn't know I was his boss, nobody told him.

JN: Right, I've never had a boss.

BU: John's more of a friend than a partner, has been ever since, but it started because when I took his application note and I helped him edit it. There were a lot of red marks on it when I gave it back to him. He gave it back to me and all he had changed was he had crossed his name off and put my name on it. I said, no, John, you don't understand. I don't want any of the glory, I don't want your ideas, I don't want my name on your stuff, I just want to help you get it published and...

JN: I think it's really fascinating all the different things we remember.

RW: Well, what killed that central group? It was split up amongst the various fiefdoms and that would seem to me to start the downfalls, one of the starts of the downfall of Fairchild.

JN: I don't know. I got to Fairchild and they gave me some project, full ladder of some type, it was being done with RTL or something like that, they wanted me to analyze it. I analyzed it and decided something better do and I spent all the time I was at Fairchild picking out what I wanted to do and I did it. I read this ap note, Larry Blazer, Larry Blazer wrote an ap note and it said, hey, here's a really neat thing. You can digitally synthesize all the frequencies for a Citizen's band radio, twenty-seven channels. And being a pilot at the time I knew that aircraft radios had three hundred and sixty channels. I said, you know, this is a good deal for a twenty-seven channel CB, it must be fantastic for a three hundred and sixty channel. So I started working on digital frequency synthesizers and we ended up going around, I took the, I flew the airplane and I flew and I visited all the aircraft, major aricraft manufacturers, Keene, Narco, Arc, Aircraft Radio Corporation, Bendix and I gave them this pitch on, at that point in time, all the frequencies were synthesized by these huge barrel switches with tons of crystals on them and you switch this crystal and you mixed it with that crystal and you got this frequency there, so just to show them how simple. So it had ended up that Keene ended up using the Orbit shift register for the counters. Narco took our application note that Jake Steinhelfer and I had prepared, designed and implemented it just like it was, just copied the application note. So we ended up trying to figure out how to get a programmable controller at a very high frequency and saw what some other people was doing and it suddenly occurred to me that all I had to do was have a counter that would count by one value and that value plus one so divide by ten and eleven and you have one input that selects what it divides by ten or eleven. And I call it pulse swallowing, people call it dual modular prescaler now and Fairchild wasn't interested in patenting anything, but right now everybody that, hand-held radios, TV's, everybody uses this dual modular prescaler idea.

RW: Fairchild only wanted to patent a process.

JN: If it had something to do with processing silicon.

RW: I did the same thing. I came up with a bunch of ideas and, sorry, not interested.

BU: And yet, a lot of the system work that was done by the first systems engineering group, which was the group John was in, have proliferated throughout the industry and ended up in much larger semiconductors today than were available. Your question earlier, Rob, I recall how the central applications group, which was really a circuit engineering group, we were the only systems engineering group in the team. And it was based on different technologies available at the time. DTL was one organization, TTL was another, CTL, RTL and so on were the different teams in that group. And that eventually evolved so that they became divisions and the technologies became bipolar, MOS, linear, diodes, and so on. And the engineering groups that were part of the central applications group went off to those divisions, which each had its own vice president and its own engineering team. But then later it was decided that there was something missing from that and that was this engineering support to marketing which would provide a connection between the technology and the customers to write these application notes and to show customers how to use the integrated circuit technology. So then a central applications group was recreated out of that that didn't do the circuit engineering. The circuit engineering was part of the divisions and the applications group then became a part of marketing.

RW: That's what I remember.

BU: And the marketing organization was different.

JN: And when we quit basement of Building 20.

BU: Yes, I'll never forget that.

RW: It was before that, there was a change over in R and D when they decided to cancel the bipolar portion of the ASIC of the gate array and devote Bob Nevil to go devote himself to bipolar memories and it turned out to be exactly the right decision.

BU: Yeah, Bob was a bipolar guru.

RW: He was very successful.

BU: And we didn't think we were going to get the complexity on a chip that we needed for bipolar gate arrays, so it made sense.

RW: I have a story here. I was supposed to be the logic designer and it's pretty tough with thirty-two gates, but you can take them apart but they weren't complete gates, they were components and so you could play games with them. So these RFU's would come in and I would go through the analysis and go to Bob Nevil and say, well, it won't fit and he's supposed to be the circuits guy and he's say, well, what if you did this, and I'd say, oops, I'll be back. It turned out he was a better logic designer than I was, which was a little embarrassing.

JN: You're over the line there, you know.

RW: Anyhow though, but at that time, well, say something about the tester.

JN: Oh, sure.

RW: Because that the start of testing.

JN: When we were up at R and D with Shriner, there were three groups. One was the architectural group, which you came up with the architectural design for some of these gate arrays and/or invert technique and there was the CAD group, which really Hugh Mays was the leader of, but we were to take what he did and implement it in production, so we had our own little CAD group. And the third group was, we realized it was going to take a whole new generation of testers to test these large-scale integrated circuit devices because there were more pins and there were more complex functions. Well, Harold Vitale was the supervisor of that group and Harold was developing this thing called the 8000A, which was the very first LSI tester. A tester designed to go down there with a lead frame, and grab a lead frame and test these wafers while the chips were still on the wafers and also to test the packaged chips and chip handlers and to develop the, you needed a computer in it and it had to be a sizeable invention. Well, Harold and his team were coming along very well with this at R and D and the guys down at Fairchild Instrumentation said, wait a minute, we're supposed to be in the tester business. How come you guys are developing this tester? So through some political shenanigans, eventually the group that was doing the tester got transferred to Instrumentation. Later on in the larger scheme of things, Bob Shriner, who was our boss, I guess about three organization changes later, he took over Fairchild Instrumentation and renamed it Fairchild Systems Technology and that division then produced these large, that was the forerunner of the Century 400, which is a line of testers that's still being produced by Schlumberger.

RW: Hugely successful.

JN: Right.

RW: I mean, the success story in testers.

JN: And they sold those testers to everybody else that was building LSI devices, then VLSI devices.

RW: It all started there in our R and D.

BU: Harold the Rodge Vitale.

RW: And of course Harold went on to GenRad and to, I forget all the places, but he just travels around and designs VLSI testers.

BU: Yep, he and his team. There are a number of guys on his team that were expert at that.

RW: Yep. Well, that was the start of the LSI testers. Up to that point the highest pin count we could do was twenty pins, I think the 4000 was it? The tester? The Fairchild 4000.

BU: 4000 and 5000.

JN: I remember when John Hume decided to quit Fairchild and so they didn't want him to quit for some reason or another and they decided that what they were going to do was give him this big cushy job as they could possibly give him, and he was going to give him a job of being a technical liaison between all the different Fairchild Groups. They said he could have whatever he wanted and he picked me, and him and I went up to R and D and then we took this trip and we went around to all the different Fairchild Groups as being, you know, really VIP's, got the real treatment and found out what all the different Fairchild outfits were doing. And after John Hume had had the exposure to what all the Fairchild people were doing, he quit.

RW: So John, what happened after John Hume left?

JN: Well, they somehow couldn't bring themselves to be in this "group" which consisted of me. And so I just kind of hung out at R and D and looked for stuff to do and people would come by and I looked into building a unit that had four bad memories that looked like it was or maybe it was five bad memories, and it looked like it was one good one.
Reusing reject memories.

RW: Partials.

JN: Yeah, partials. So I worked on a deal for that and they wanted to do video analysis of masks for faults in the mask. The first place is, when you're making IC's that the masks are not perfect so you know there's going to be faults in the wafer because there's faults in the masks. And so they wanted to scan masks and look at them and figure out automatically if they were defective or defective too great a degree. So I worked on that for a while. And then somebody back in the, Fairchild had a group that did typesetting machines and these typesetting machines were being run by discrete RTL logic, I mean really old stuff. They had these big boards that did the, well, what they had was they had this huge wheel, I mean this wheel had about a diameter of, oh, four feet. It was kind of drum and it had on the end of the wheel was photo-optical because it was a film of all the different possible fonts. And this film went by the photographic paper that was put in this machine in strips and then it would, the whole thing about this, all this logic did was it got what you wanted to print and it fired the strobe light at the right time. So when that character was coming by it fired the strobe light and so they figured since I was the guru on MSI that I was going to show them how to take this RTL design and convert it into MSI. And I looked at what they were doing and at that time you could buy a very sophisticated mini-computer, many different things for about, oh, five thousand dollars, maybe four thousand dollars and you could do this whole thing with a simple mini-computer and that's what I told the guys. I said you don't want to do a MSI design to this RTL thing when you could just take a mini-computer and program it. They were selling this thing as having this capability, I forget just exactly, it was auto-justifancy. They actually had never built one that did that. In fact, they had never designed one that did that. They had ads that had this listed as an option with a price on it and they hadn't even designed it. What they had was they had a couple of slots in the rack where they could put the cards that was going to do this.

RW: Well, that brings up a point. Before microprocessors, computers were indeed used as controllers.

JN: Oh yeah, oh yeah. Well, that's what I did at Philco Western Development Labs. That's what we did at Philco Western Development Labs. When I went to Philco Western Development Labs there was this three-rack piece of equipment, which I inherited from somebody, in fact, I think I inherited it from George Byers. He had designed it and then I became the sustaining engineer for it. And I learned all about this thing and everything was done by dedicated logic and what it did was it controlled the antennas. The antennas had three axises and this thing read the encoders that got this serial stream of bits, which was a gray code, converted that to binary, compared that value with the command value that came off of a paper tape and then sent out an adjustment signal to these hydraulic motors to move this huge antenna. And it occurred to somebody that we could buy this 160A computer for one hundred and ten thousand dollars. It had four thousand twelve-bit words, a paper tape punch and reader and a console and it looked like a desk and we could take and program that computer to do this entire job. And that was cheaper than the three racks of equipment that was dedicated logic that was doing the job now. And from that point in time we ended up, everything was being done at this tracking station was being done in the computer. We were sending it telemetry and getting telemetry back, cutting it apart, displaying values, sending commands up to the satellite, all kinds of stuff and it was all being done by computers. We had these, programming was all done on IBM punch cards, one punch card for each instruction. And we had these big boxes of IBM punch cards and then we'd used these felt marker pens to mark which group of punch cards was which section of the program. And then when we would run, we'd send the program, all these punch cards down to this computer that would do the assembly. We'd get back the assembled code and execute it and come back and find out that, you know, it wasn't working, we'd find out, well, we have to take out these instructions and put in these instructions and then we'd punch out the cards, take cards out, put cards in.

RW: How much would that controller cost today?

JN: You couldn't buy that controller today.

RW: Well, no, a controller that would do the same job?

JN: Oh, probably a dollar and a half.

RW: Come down a little bit.

JN: Oh yeah. What we were doing there, well now, see that, excuse me. Let's see, we paid a hundred and ten thousand dollars it had 4K in memory, but it had a paper tape punch and a reader. Now the paper tape punch and reader really hasn't come down in price very much today, in fact...

RW: I don't think you'd use it.

JN: You probably couldn't find a paper tape punch and reader.

BU: You'd probably use a floppy or a hard drive or something.

RW: Yeah, right, just a PC would do fine. Well Bob, I remember when the bi-polar ASIC Micro-matrix program was canceled and you came to us and said, well, you know, I guess I'm not needed anymore on this because it was, I guess, Jim Downey was running the MOS and now it was going to be a strictly MOS show and so then you transferred to head of digital marketing.

BU: It was more a circuitous route than that. But there was a time when there was a huge reorganization in the custom Micro-matrix arrays department, which you and I worked in. And the bipolar went by the wayside and it became an MOS thing and it went to a different division and so on. By that time we had already produced something like three thousand designs in MOS and maybe a few dozen in bipolar because there wasn't that much demand for the complexity level that we could fit on a bipolar chip. I went from there to, and you took over the engineering department that I had run and I went and took over the central applications department which was about to move into the basement of Building 20. And I ran the applications department for two years. That was about fifty five engineers and technicians and people divided into linear consumer group headed by Norman Doyle and digital group headed by Peter Alfkey and a couple of other groups. I think Bob Ricks was in charge of one of the groups and there was another group that Bob Hood had, which was automotive applications. So central applications was not now a circuit design group like it was in the previous group, but now it was an applications group that was responsible for writing the application notes, coming up with ideas for new integrated circuits in all the different divisions of Fairchild. It didn't matter what technology we served them all. And going out and helping customers in the field to show them how to use Fairchild's integrated circuits in all of its products. It was a very interesting job because I reported to the vice president of marketing and I was the only guy in the company besides the president of the company who could think on behalf of the whole company because everybody else had a slice of the pie, a vested interest in either a technology or a geographical orientation or some other piece, and the president of the company and myself thought across all product lines and all world-wide operations. And as a result, Corrigan by this time, had taken over the presidency of Fairchild. And he said we've got to cut the number of products we make in half. And he gave me the job of arbitrating which products would go and which ones would stay and so I was dealing with all these division vice presidents who didn't want to lose their products and I got to tell them what to do and I was reviewed every weekend in the board room. But that was probably one of the best jobs I ever had, was running that applications department because it also had a view of the industry that was global and throughout all of the industries that used electronics because after all, semiconductors were pervasive into all those industries. And we talked to all those people and we could see what was happening out there in the electronics industry because they would tell us about what they were doing, what their new designs were going to be.

RW: It was hugely successful it seemed to me.

BU: Oh, yeah. And John came back into that department and made a lot more contributions in the digital applications area.

JN: We did the program, we designed a set of devices that were going to be programmable in this first kind of microprocessor kind of thing, only I didn't call it a microprocessor, I called it programed logical image or something like that.

BU: Large-scale programmable standards, I recall. This was a set of standard circuits that we could program to do almost any operation.

RW: Were they bit-sliced?

JN: It was slower than that.

RW: I don't quite remember.

BU: It was more of a functional system.

JN: We actually built some simulations of it. We had a card for a chip and we wrote some programs and I think we did a controller for a washing machine, was one of the applications.

BU: You know John kept telling us and we all in the digital side, believed that this was something that was coming but in order to handle functions as complex as some of these functions were in terms of the number of circuits, that they would have to be programmable otherwise the semiconductor industry couldn't use it because the semiconductor industry has to sell in high volume. And you couldn't get high volume by having a custom device for every application at this level of complexity so the solution that John came up with was make them programmable and, of course, that was, you know, some of the thinking that eventually became the microprocessor.

JN: Well, it was just like one of the things we were doing at Philco.

RW Really, it was just an extension of what you'd always done. Well, what happened to that central group?

BU: Well, a funny thing happened. We had all been told that each of us had to have a successor for ourselves in the organization so that the organization could grow and I had decided that, Norman Doyle was the manager who was going to step into my shoes if I needed to, if I got hit by the proverbial beer truck or had to get promoted, Lord forbid. So one day he walks in and quits. And I said, Norman, you can't do that, you're my successor, I mean you're going to take over when I get out of here. Without you I can't get promoted. He was going to go take over National Semiconductor's linear applications organization, which was a nice job. I had a heck of a time talking him out of leaving. It took over a week of cajoling. And finally he and I were having a beer at the Wagon Wheel and I turned to him and I said, Norman, what is it, what can I do to get you to stay with Fairchild? And he said, not much, I guess, Bob. What if I offered you my job? He said I'd have to think about that. I said, don't think very long because I'm offering you my job right now. He said, okay I'll take it. Now I hadn't consulted with my boss who was Dick Henderson, the vice president of marketing. So I went over to the phone in the crowded Wagon Wheel, in this bar, called up Dick Henderson at home and said, look, I just quit my job and Norman Doyle is going take over, what do you want me to do? He says you can't do that. He says, be in my office at seven thirty in the morning. And so I was in his office and he tried to talk me out of it, and I said, I can't, I've already offered Norman my job and he's accepted. So what do you want me to do? And he says, do you like digital or linear better, because the linear marketing job, which had been filled by Mike Markila for years had just been emptied when he took over the digital marketing job and he was in the digital marketing job for one month when he quit and went to Intel to take over their memory marketing job, so both jobs were kind of open and I said, digital. And that's when I got to be a marketing guy. I'd always been a technical guy up to that point. So I took over the digital marketing department and I was there for two years and again, there were more reorganizations and changes in the company and they split up the applications department shortly after that. When Norman I think, no, he stayed, but in any case, they decided that the applications department should be split up and go with the divisions. So I inherited the digital applications group, which Peter Alfkey was heading at the time and John was in that group and that continued for a couple years. We had some phenomenal success there because about that time this MSI circuit functions that John and Clive and a couple other guys invented had matured and we had this sales goal of ninety million dollars for the digital products division and a profit goal and a share of market goal and an average selling price goal. And in one year, which has never been done before, we beat the sales goal, we beat the profit goal, we raised the average selling price and we increased our share of market. And it was all done because of the MSI products that John had come up with.

RW: Yeah, clearly superior. Well, that also impacted me because I was off doing customs in MOS and of course MOS division we had this series of clowns that would come in, the vice president clown of the month, these guys would come stumbling in and then stumbling out.

JN: I was in MOS for a while.

RW: Jack Gates and Roy Pollock and oh boy, Pollock later went on to destroy RCA so he was, not enough to destroy MOS, but Corrigan shut down the Micro-mosaic effort and everybody had to find a position and it was shut down really for two reasons. Number one, the microprocessor had come out and people perceived that they could use that instead of custom. But then also the price of MSI, there was a recession in early '70, and the price of MSI, which had been around five dollars or so, dropped to a buck and custom just wasn't competitive. We could only put about five hundred gates on a chip so it wasn't competitive and it was a lot slower than bipolar. And so John, you put me out of business there.

BU: John always said all along that standards were much better than customs.

RW: I know, I know, I know. He still does, as a matter of fact. So I was out of a job and you gave me one in product planning in digital marketing.

BU: Yeah, when I found out you were available I jumped at the chance to have you come over and at the same time, John was in MOS at the time and they laid him off.

JN: No, they didn't actually lay me off.

BU: They tried.

JB No, I'm trying to remember the name of the guy that did it, but, you know, Marcy worked for him in another company; said he was really great to her. And he had fond memories of me. But at the time, you see, he, I just didn't seem to go along with the program up there. These people would come in with these absolutely fantastically dumb designs and they would be buying MSI and other products from Fairchild, and the salesman that was getting the buddy for selling all these other products at Fairchild wanted to see Fairchild do the MSI thing in MOS form, the custom MOS thing. This is what I was told and so we ended up spending a lot of front-end money and my impression was to design these things and by the time we got the part for the design the customer would have discovered that what they thought they wanted would be this thing that really wouldn't do the job, so they never ordered any.

RW: Yeah. Well that's the dilemma of custom, a lot don't go into production, but going back to that. At LSI Logic our problem was getting the first designs because nobody wanted to trust the fate of their company to this start-up. And so the first ones came in and I looked at them and I said, this could be done with just a few MSI parts, I mean, this is not one that you want to do in custom. And Wilf took me aside and says, shut up and take the business. We'd been searching for somebody to take a chance on it, but they really were dumb designs I have to say. I think one of them could have been done with two MSI parts, to give you an idea.

JN: Well, anyway, the guy that was working under Pollock, and I can't remember his name right now, called me in and said, you know, he'd had this employee that had a nervous, you know, a mental breakdown and that everybody told him not to keep him, to let him go and he took a chance on him and he worked out and he was going to take a chance on me. And...

RW: Another mental breakdown case?

JN: I had thought that was really nice of him and I walked out of that meeting and I called Norman Doyle and found out he was in route someplace and somebody got a hold of him and he called me from a payphone at an airport and I said, Norman, I need a place to go to and he says, no problem. So I moved.

BU: Into central applications...

JN: Yeah, back into central applications.

BU: ...which Norman had taken over.

JN: So this guy took me into Pollock to talk me into not going and that wasn't a very pleasant interface.

RW: Pollock was literally, I had this secretary; attractive secretary. He would literally chase her around the desk. Literally. I'm serious, at work. That's the sexual harassment, man. Well, Bob...

JN: You never had a secretary that attractive.

RW: Well I thought so, actually. Anyway, Bob, in '73 you started Logical Services. What led you to do that?

BU: Well, John hired me to be his boss.

RW: Oh.

BU: It's true. But the genesis of that was, back to Fairchild for a second. John had these programmable chips and couldn't get anybody to listen to him in the divisions so I went to Roy Pollock and Bernie Marin who was the marketing guy for the MOS division at the time and helped John to set up a meeting with all of these technical guys. He had the ear of all the technical guys, but he couldn't get the ear of the managers to allow this meeting to occur. So we got this meeting going and John, for an hour and a half I guess it was, explained the nature of these programmable devices and how they would be a product line that the MOS department should use. And it goes on and on, but to make a short story of it, a few days after the meeting we got a letter...


JN: Bernie Marin.

BU: ...signed by Bernie Marin and Roy Pollock that says, I'm sorry but the world is custom and we're not going to put any energy into these programmable large-scale standard devices that you guys are talking about. I have that letter somewhere.

RW: Which today is how many billion-dollar market?

BU: Right.

RW: I mean, Intel alone...

BU: Intel, of course, did a very good job of that and is king of the hill.

RW: But its got to be many billions of dollar market. Roy Pollock...He was a clown.

JN: There's nothing wrong with any of these people, I don't know.

BU: Well at this point John and I looked at each other and we said, do we want to work for this company anymore?

JN: So we invited Peter Alkey to lunch at the Velvet Turtle and I told Peter I'm quitting and Peter turned to Bob and said, Bob, what are we going to do about that?

BU: I said, Peter, I got bad news for you because Peter worked for me and John worked for Peter and we were having lunch with Peter, both his boss leaving and his best digital designer is leaving. So we decided to start a company before there were really microprocessors available, but John's idea was that we would build these little building blocks on printed circuit boards, that we would do that same thing that you could do with a microprocessor when they became available. You would just take a CPU card and a ROM card and a RAM card and some IO cards of various types, put them in a chassis and program them to do whatever control function you wanted.

JN: We picked out plastic injection molding machines and said, well, we can build controllers for plastic injection molding machines, they're doing discrete logic for relays or programmable logic controllers so figured that we could do much better than that.

BU: Got a good reception too for the product. This was the first application of our general-purpose capability. We went to McCormick Place in Chicago where they had this big show with all these huge presses from Van Doren and Cincinnati Milacron and Husky and all those guys that made plastic molding machines and they just had this brand new idea with solid state control. They were just catching on to getting away from relays and when they saw our little controller was programmable and they could really adjust the machine to do almost anything they wanted, they really liked it. But it turns out, that was 1973 and just a symptom of what the problem was, all of the plastic feed stocks that were used to make the samples, got stolen off the loading docks at McCormick Place because this was the oil crises and, of course, oil is where the plastic feed stocks come from and when the plastic dried up, all of the orders for injection molding machines got cancelled and the whole industry went down the tubes. And of course, there we are with our one shot at this small company, working out of my garage, left in the lurch. That's how we got into the consulting business.

RW: I was also very shrewd in that period of time. I bought an Olds Cutlass with the 454 engine in it, excellent, excellent timing. Let's break.

RW: So Logical Services is, the oil crisis hits. Now did you have any venture capital money?

BU: No, we cashed in our Fairchild stock options and used that to essentially fund the company. John, you put some money too.

JN: I sold the house.

BU: Sold the house and came up with some cash.

JN: Burned the boats.

BU: And we put it together. I bought a condo and we started the company in the garage of the condo.

RW: Well, that is a Silicon Valley tradition, isn't it.

BU: You bet.

RW: So then what did you do?

BU: Well, we found ourselves with this disaster on our first product line and we needed some income so I want back to my old boss, Tom Longo at Fairchild and Dick and John came up with some ideas for a next generation of integrated circuits beyond the MSI circuits, we called them Super MSI and we convinced him that that's what he needed so he hired us and we designed that, you know, four or five chips that were another generation beyond MSI.

JN: Dick Derickson did a new implementation of a FIFO.

RW: Yes, that sold very well. It was a super product.

JN: Not only that but it was patented. Yeah, got a patent on that.

BU: First one, first..

RW: Yeah that was a super product.

JN: I forget what I did. I did...

RW: So you were essentially consulting for Fairchild for a while.

BU: That was really our start in designing new products for people other than ourselves and that became kind of a way of life for us and that we were able to go find somebody that wanted some new product and they'd hire us to design it and that's what grew the company from John Nichols and Dickie Derickson and myself and Frank LaVato to, we've got twenty five people aboard now.

JN: You still don't remember the name of the secretary.

BU: I can't remember her name.

JN: She looked like a Raggedy Ann doll. Can't recall her name.

RW: You did the Jacuzzi control didn't you?

JN: I remember...

BU: At one point we worked on a Jacuzzi control.

RW: And were you working on a refrigerator...

BU: We designed the only refrigerator that can take the load off the power company for six hours without melting the ice cream. It was a thermal energy storage device. Of course our client came up with this chemistry for the thermal energy storage and we did all the controllers for it.

JN: Then Bob was supposed to go by late at night and do something about the refrigerator and somehow ... it was supposed to do a defrost, didn't have a defrost, somehow he forgot and left this thing on. I come in in the morning and there's fine soot all over the place and there was this big hole burnt in the middle of this table.

BU: I managed to destroy the first refrigerator that we designed by failing to turn off the switch somewhere along the line.

RW: Well, is there a product that, that people would know that you designed?

JN: Lifescan.

BU: There's probably a whole bunch of them but they don't have our name on them.

RW: Oh, sure.

BU: Most people have heard of the company called Lifescan. That makes the glucose scan meter for diabetics, measure blood sugar ratio.

JN: Pocket blood sugar ratio.

BU: Yeah, fits in your pocket. The chief engineer of Lifescan, Ray Underwood, when they were three guys in a garage like us came by and found us and said okay, you're going to be my engineering department. The reason he chose us, I think, more than anything else was because when he came in he said I want to work real closely with you. John pipes up and says we'll give you an office, you can move right in here. And he did and I think that's the thing that really closed the deal. So for six years we designed four generations of their meters for Lifescan.

JN: And did other things at the same time.

BU: Yeah, and a bunch of other products in parallel with that.

RW: That's great. Now you're still privately held?

BU: Yes.

RW: And no venture capital money?

BU: No, it's just what we've, the earnings...

JN: Retained earnings.

BU: A couple of stories you've missed though Rob. When we were on Walsh Avenue, this guy named Wilf Corrigan, who used to be president of Fairchild, came by.

JN: Oh yes.

BU: He called me up and said, hey Bob, I want to talk to you. He came by and sat in my office and he says, Bob do you know any good CAD guys? Yeah, Koford and Jones, they're the best. And he says, where are they? I said well I'm not sure but I'll call Rob Walker and we'll find out. And later on he wanted to pursue them and he also asked me who'd make a good vice president of engineering and I said Rob Walker would. He came back a week later and he said are you sure Rob Walker would be a good VP of engineering? He's been a marketing guy over at Intel for five years. And I said Wilf, that's exactly what you need because you're going to have to have your engineers sell those custom circuits to your customers. So...

RW: You convinced him. And I had been out of engineering for seven years, which is an eternity so not only did you give me my first job in applications for specific IC's, you then got me the second job at LSI Logic so I had a second chance and not too many people in this life have a second chance.

JN: I put you out of work and he gets you a job.

BU: We had a bet going at the time.

RW: But I did give you some stock.

BU: Do you remember the bet? We had a bet that the first one to reach a net worth of a million dollars would treat the other to an all expense paid trip to Hawaii for himself and his girlfriend. And you paid off with 250 shares of LSI Logic that you bought for four cents and I sold for eighteen dollars a share.

RW: It's now around sixty.

BU: Yeah.

RW: Finally, after I sold all of them.

BU: Great.

RW: Now it's amazing that somebody, you've had just this huge influence on my life and just this sort of random meeting here and then going off.

BU: It's a very small valley.

RW: Well, as you see the semiconductor industry now, what are your ideas? I mean you're a user now of semiconductors and...

BU: In a sense we're still back in the application days showing customers how to use the new technology that comes out. So we have to keep abreast of what the latest developments are in the new technology so that we can design it into systems that our customers use. So it's very similar to what we were doing before.

RW: But you get paid for it.

BU: Now we get paid for it.

RW: Ahhh, I never thought about that. That's interesting. That is interesting. You've been doing the same job now for all these years.

BU: Where's it going? Well, Moore's Law has shown us that you really can double and keep doubling the complexity of what you can fit on a chip for quite a long time. It's gone on much longer than most technological phenomenon.

JN: It won't last another twenty years though.

BU: No, it won't. Unfortunately it can't.

RW: More than the number atoms in the universe.

BU: We'll run into physical constraints on the size of a molecule and we can't do much better than that at the moment.

RW: You know some of these isolation layers, gate widths and things like that can be measured in atoms, you know, like ten atoms thick, things like that. It really is incredible.

BU: Well, I still think we have a long way to go in the electronics industry which is driven by the semiconductor industry. Its enabled all of the revolutions that have occurred in making things smaller and cheaper. We wouldn't have PC's today if we weren't able to put this kind of logic on a chip and do it economically, so I would never speculate beyond five or ten years because I think that technology will surprise us. But the way I do product planning is I always use the Star Trek model in the product planning, you know, beam me up Scotty. If you can see something or imagine something in science fiction and you think of it as how you'd like it to be, I'd like a communicator were I can touch my badge and be in touch with anyone in the world and access any information, computer, tell me whatever I want to know. I think that'll be reality and I think it'll be reality in a hundred years. So extrapolate back from that and what can we do with the technology today and what can we do five years out. And go along that line and you'll see what's going to happen as you approach the science fiction of the future. If you know it can be done in a hundred years, we'll probably be doing something like that in ten years.

RW: Yeah. We used to joke about hot rodding ROMs, hot rod ROMs and of course those are now offered.

JN: We knew about that. We did the heating control. The first microprocessor, we did the emulators for them. We built the emulators that Ford Motor Company and others used to develop the engine strategy. Intel, actually, what happened was, is we got a job doing the Intel, helping the Intel develop the microprocessor which is the, I forget the number of it is.

RW: It was custom.

JN: Not the standard.

BU: 8096, wasn't it?

JN: 8096 or something like that.

RW: It was specifically designed for automobile use.

BU: Yeah, Ford.

JN: Well, yeah, it was. But it was specifically designed for that kind of application, actually. But, we built an emulator wire wrap and then we ended up building multi-wire boards and a Nova box and we built about twenty five of them. Didn't we end up selling them?

BU: About a dozen actually.

JN: But you see, Intel told Ford to come to us and we'd build them for them. Ford came to us and says that's a great idea. You ship them, when we receive them and check them out then we'll pay you. We didn't have the cash to do that operation so we said, hey, we can't do that you have to pay us as we build them. So Intel came back and they said okay, we'll buy them and sell them to Ford and others. So we did that, but as we were doing this I looked at that and I says it's obvious here that the way the engine works now is it's going to be in this ROM.

RW: Sure.

JN: And one of the products that's going to be coming out is speed ROMs, hot rod ROMs.

BU: You can make the engine as clean or as dirty as you want or make a square cam out of it.

JN: Well, you can hop it up.

RW: It's interesting that today a '95 Caddy, I think it's called the Concourse, it's the big six passenger huge Cadillac, it gets better gas mileage than a Ford Pinto of the '60's and that's electronics, largely.

BU: John wanted to go into the racing ROM business.

JN: Yeah, that's what it was called: a racing ROM.

BU: And sell them through Goody's Speed Shop.

RW: Well, going back to something, we wanted in the first standard cell, we wanted a device to build a design that would demonstrate the technology and so Dick Derickson and I came up with a tic tac toe player...

BU: Yes.

RW: ...which would have been the first electronic game. And so we were busily designing this thing, which we were going to build at MicroMosaic to demonstrate that the technology existed. And I got a call from you and you said, if anyone gets wind that, because we called it something other than a tic tac toe plan, you know, it was a decision...

JN: It had to have a code name.

RW: Decision circuit or something like that and you called me and said stop, stop working on that immediately. If anyone finds out about it we'll all be fired. And that would have been the first electronic game. Well, great. Thanks so much and its been a long association and has been great fun the whole time.

BU: Well, I wish you well with the story. It needs to be told.

JN: I want to apologize for walking out on that meeting back then...

RW: Logic Design, it was.

JN: Philco Western Development.

RW: Yes, yes. Okay, you're forgiven.