Historical Document · 2014
Oral History of Bill Mensch (2014)
Mensch on designing the 6800 and 6502 support chips, the 6530 RIOT and the KIM-1, and the Western Design Center.
Historical Document · 2014
Mensch on designing the 6800 and 6502 support chips, the 6530 RIOT and the KIM-1, and the Western Design Center.
........ Computer • History Museum Oral History of William David “Bill” Mensch, Jr. Interviewed by: Stephen Diamond Recorded: November 10, 2014 Mountain View, California CHM Reference number: X7273.2015 © 2014 Computer History Museum Oral History of William David “Bill” Mensch, Jr. Stephen Diamond: OK, it's November 10, 2014, here at the Computer History Museum. I'm Steve Diamond, and we're doing in oral history of Bill Mensch. Thanks, Bill, for joining us. William David “Bill” Mensch: Well, thank you. Diamond: We'll be talking about a variety of subjects and your perceptions of what's happened in the past and, perhaps, where things are going in the future. Why don't we start out by talking about your youth and your education, and then we'll follow that up into your transition into the semiconductor world? Mensch: All right. Well, thanks for inviting me. It's an honor to be here right now, and I will enjoy telling you what I know about my life, how I got here. And we'll start, then, with my growing up on a farm in Pennsylvania, Bucks County, about 35 miles north of Philadelphia, very rural. The dairy farm had like 26 head of cattle, and I was a middle child of eight children I grew up with. And as a result, I got to explore because it was more fun being outside of the house rather than inside the house. When I was probably about 10 years old, we may have gotten a TV. We had a radio, liked to listen to the Lone Ranger. Soon as the cow's milking was done for the evening, we'd run in and try to get a good spot in front of this tall radio, and so I started out what you would call, I call, pleasant poverty. So the dairy farm to support eight children, that's the source of income, 26 cattle. So my father was a small businessman, and the farm was his business. He was given that farm by his father, and so there's a lot of history associated with that happening. I found out, later in life, that my father wanted to be an engineer and wanted to go to Drexel. He wanted to be an electrical engineer. So when you look at my past, I don't think there's many indications that I'd be sitting here talking to you. I don't know what, actually, was the path and what the guide posts were because it was like I was following something, but I didn't know what it was. Diamond: When is the first time you thought of yourself as possibly an engineer? Mensch: It was probably when I was about-- when I was an engineer. In other words, the reason why I went to college was because I took the academic classes in my high school, and I really didn't know, even, what that meant. So therefore, I took the classes that I was interested in, I suppose, math and science. But for me, I never thought I was going to be an engineer. I thought I was going to be a laborer because that's what you do on a farm. CHM Ref: X7273.2015 © 2014 Computer History Museum Page 2 of 54 Oral History of William David “Bill” Mensch, Jr. So you learn how to do things that you're told to do. We need this project done. We need to plow that field. We need to cut that hay. We need bale the hay. We need to bring it into the farm. We need to milk the cows in the morning and the evening. And all these things you were doing, and you're expected to be. So hard work was what I knew, and I enjoyed that. I enjoyed being outdoors. So for me to think that I was going to be in an office environment, that's nothing I ever thought I would do. So maybe when I was 18, and the guidance counselor said, well, you've prepared yourself to go to college, I'm thinking this is a mistake because I wanted to get out of class. I didn't like English, didn't like history, and didn’t like social studies. I just liked math and science because you could get real answers. If somebody asked me what did this poem mean, I'd go why do you ask a poet? I mean, really-- I mean, what did I get from it? That wasn't what I remembered the question to be. So to me, going to college was not expected. None of my family ever went to college. Diamond: So you find yourself studying engineering? Mensch: Yes, and I think I studied engineering because my dad had engineering magazines. Maybe IEEE maga-- well, I don't know IEEE was around. Maybe it-- Diamond: Oh, it was. Mensch: It was around, but some of these magazines, like Spectrum, started 50 years ago, and this was- well, maybe that was something. But what I'm getting at here is he had some electronic magazines, and also he equipped the farm with a generator that, when the Philadelphia Electric Company power went out, we needed to keep the milk cold. So then he had a big circuit breaker from a ship, I think, and he installed that and then started the tractor and started generators, so to keep the milk cold. So that's electronics, so I suppose I got that from my dad. Diamond: Do you remember what classes interested you most? Mensch: In college, or-- Diamond: In electronics, in engineering? CHM Ref: X7273.2015 © 2014 Computer History Museum Page 3 of 54 Oral History of William David “Bill” Mensch, Jr. Mensch: Ah, no. Well, Temple was engineering technology, so it's more like learning how to use equipment technology, not the theory and the physics. And so therefore, the classes that I took were just to get my degree, get out, get a job as a technician, thinking I'd be able to make more money as a technician than as a laborer, so I was rudely awakened when I'm working for Henkels & McCoy making about $150 a week with overtime and then working for $85 a week. It didn't make sense after I got my degree. So I needed to make some changes, and I decided to be an engineer to pay my mom back. Diamond: So what was your first job after college? Mensch: OK, I worked for-- and I very seldom tell people this. But for three months, I worked for a company called Polyphase Corporation in Bridgeport, which is across the Susquehanna-- Susquehanna? The Schuylkill River from Norristown, and it was an old manufacturing community, and it was in an old factory that this Polyphase was. It was depressing situation for me. That I didn't start until-- it would have been in the spring of 1967, and I first went into the army for National Guard training. And that's why I did not interview right out of college. So it was after that that I got a recommendation from a friend of mine that found a job there, and he got $100 because I took the job. So then I decided this wasn't for me, so I quit without another job. Then I worked for Philco Ford, and that's where I was introduced to semiconductors. Diamond: So tell us about that. Mensch: Well, at Philco Ford-- now Ford bought Philco, and I think it was for the electronics that was going into automobiles. And then Ford-- or I don't know if was Ford or Philco-- acquired GME, which was the first MOS manufacturer of semiconductor products. And in fact, they had a 1024-bit ROM back in probably '67, something like that, because there was a jet engine controller that was built out of these with eight 1024-bit ROMs. Of course, there would've been memory in there-- I mean, RAM in there. So that is where I learned about semiconductors, and then I studied the physical characteristics of semiconductors, for instance you want to know the f tau of an amplifier and the frequency, and then the thermal coefficients of materials. So if you put a semiconductor, a regulator, solid-state regulator, on a heat sink, what's the thermal coefficient from the surface of the semiconductor to the actual heat sink? And so these kind of basic characteristics of semiconductors is what I worked on as a technician, so then I was interested in their characteristics. I went to Villanova night school. Took four different engineering classes to see if I could be an engineer, and then I did well in those four courses. Then I wanted to transfer my credits from Temple. I had like 109 credits from Temple-- you need about 130, 135 to be an engineer-- thinking I could finish in one or two years. They didn't accept my credits. CHM Ref: X7273.2015 © 2014 Computer History Museum Page 4 of 54 Oral History of William David “Bill” Mensch, Jr. That's when I came to Arizona because I had a fraternity brother that came to Arizona. And he went to ASU, and I went to the U of A, interestingly enough. I never saw him again. But that's how I got into semiconductors is through Philco Ford. Diamond: And how did you get from there to Motorola? Mensch: Well, good question. At Philco Ford, being in the radio business, television business-- [CLEARS HIS THROAT] excuse me. Motorola was a leader in those circuits, so Philco Ford was trying to figure out- or Philco was trying to figure out how Motorola was developing some of their high voltage circuits. So the breakdown voltage and other things, characteristics, of the circuitries in TVs and radios, is what Philco was looking at. So then I knew who was the leader because we wouldn't be trying to figure out what the follower was doing. We're trying to figure out what the leader's doing, so I knew then Motorola was the leader. So I wanted to work for a leader, and so my whole goal was set on working for Motorola after I got my engineering degree. Diamond: So how did that happen? Mensch: Well, that's another interesting question. That was kind of bizarre, actually. In my senior year-so it took me two years to get my engineering degree at the University of Arizona. Very good school back in those days. The professors were writing the textbooks. And actually, I took a course, and it could've been one of his first courses or early course because Temple-- Temple-- University of Arizona professor started systems engineering. I didn't know that until recently. But I took a course from him, and I did very well in that course. So what happens was, when you have businesses coming to interview the students, it was clearly stated that if you're not graduating you-- if you're graduating in the spring, you interview in the spring. So if you graduate in the fall, you interview in the fall. Well, for whatever reason, I really wanted to get interviewed at Motorola, so I went to the fall interviews and found that there was nobody signed up. And Motorola was looking for someone with over a 3.3 grade point average, and so I'm thinking, hey, there's nobody signed up. I'll sign up because I have a 3.3 grade point average. So the head of college recruiting came. His name was Jerry Fulton. And Jerry interviewed me, and we had a wonderful conversation. CHM Ref: X7273.2015 © 2014 Computer History Museum Page 5 of 54 Oral History of William David “Bill” Mensch, Jr. As it turned out, Jack Haenichen. Jack Haenichen? Yeah, Jack Haenichen was a manager from Philco that went to-- no, I'm sorry, it wasn't Jack Haenichen, it was another name. Right now, I can't think of his name. He was at Philco Ford. He was like two levels above my boss at Philco Ford. He was at Motorola, so I had something in common. There was somebody I knew. Oh, I almost had his name. [It was John Ekiss.] So what happened was I'm talking to Jerry Fulton, and Jerry and I get along really good. He says we've got a job for you, partner, because I had two years’ experience in, and so I had this engineering technology background and an engineering degree. So they were looking for people with experience like that. So when the spring semester came, I didn't do very well in that area, and I found out later why. But I don't need to get into that right now. It just didn't work, and I left in the middle of the interview thinking I already talked to the boss. And so that spring, when you get your responses back, I got an employment offer from Martin Marietta in Denver. I promptly put it back in the mail, thanks but no thanks. Got one from NCR in California working for a computer company. Both of them were computer. I promptly put that in the mail, thanks but no thanks, waiting for the Motorola. Motorola gets a letter signed by Jerry Fulton, said thanks for interviewing, but I'm sorry. We don't have a position for you. I wanted to jump in the mailbox to get back those other letters. So instead I called Jerry and say, hey, Jerry, this is Bill at the University of Arizona. He goes right. I said, well, I got this turn down letter, and I thought you wanted to hire me. Oh, Bill, there's a mistake. Throw that away. I said, Jerry, you signed it. He said that's a mistake. Throw it away. There's another letter coming. That's how I got the job. Diamond: What was the job that you were hired to do at Motorola? Mensch: Well, at Motorola-- Diamond: Presumably what interested them when they interviewed you? CHM Ref: X7273.2015 © 2014 Computer History Museum Page 6 of 54 Oral History of William David “Bill” Mensch, Jr. Mensch: Well, you know, Motorola is an awesome company. Was an awesome company? It's kind of fractured right now. But Motorola had a training program, and you spent three months in four different areas of the company on that training program. You were treated like royalty. Everybody wanted a trainee working for them. The reason why is the training department paid your salary, so you're free labor for the department that you're working for. So I thought about it, and I thought, well, why don't I learn Motorola. So my first one, first three months, they called them rotations, was in the applications department which is part of marketing. So I worked on the modem, 6860 modem, working with a consultant. I think his name was either Anderson or Jacobsen. There was Anderson-Jacobsen was the name of the modem of manufacturer at the time. Well, they wanted to do an integrated circuit modem, low speed, 300 baud and so I got to flowchart that and look at some of the circuits that were needed to make the modem what it needed to be. So then the next three months I spent in the design group, designing what I specified in the first one. Then I wanted to get into MOS processing because I wanted MOS because I knew MOS was the future. Diamond: And how did you know that? Mensch: Again, it was about some basic characteristics that I learned at the University of Arizona and saw at Philco-Ford, that the bipolar process was slower switching if you go to the-- the highest performing circuits were going to be voltage controlled, which is the FET. And so therefore, the analog, or the bipolar, was current controlled, so it's slower. But at the time, it was the fastest, but you knew in the future, it was going to switch from bipolar to MOS field effect transistors. So what happens is I know that I want to be in MOS. I knew that was a future. I was working for a leader in a business. I wanted to be in process design next, so then I would manufacture what I designed, so to speak. Not really, but that's the idea. And so that wasn't available to me. There wasn't an MOS spot available as a trainee, so I worked for a bipolar part of Motorola that was doing emitter couple logic, which was the fastest switching circuits in the world at the time. And so that ECL-- Motorola called them MECL, Motorola ECL. And so I designed a triple diffused process -- we called it MECL-N, which was the end of the line, so to speak, and I was defusing into a bulk background the collector. Then next was the base and then the emitter, and that was really hard to do. Nobody had done it before. So I decided I'm going to create a program model so I could change times and temperatures, and then I would be able to see the diffusion of the ions into the background and into the collector by the base and then the emitter into the base. CHM Ref: X7273.2015 © 2014 Computer History Museum Page 7 of 54 Oral History of William David “Bill” Mensch, Jr. So if you punch through the emitter through the base, you've got a diode. Or you've got a short, really, because they're the same type of material. So that's what I did, and so then I could get all the design characteristics with five different programs that I wrote. And there was a guy named Jim Dunkley that had a cookbook for bipolar processing, so what I did was automated his cookbook. So then the next rotation was marketing. So then I wanted to see how you sell these things, and so I worked on data sheets. And then there was a quote needed for IBM for various size memories and various volumes of memories. So it's a learning curve over a seven-year period, like seven different-- all I remember is 7 7 7, like seven different volumes, seven different years, and seven different densities, and so there's seven pages of numbers. So I didn't think the secretary would type up the numbers very accurately, so I decided I would write a FORTRAN program. And all I did-- it was a spreadsheet I wrote -- so you plug in the numbers, and you print it out. It was a FORTRAN program on a Silent 700, which was a terminal at the time. And then I copied on a-- Diamond: From TI. Mensch: Yeah. So then I copied it on regular paper because that rolled thermal paper wasn't fun to handle. So then I did that within a couple of days, and then onto the next project. But marketing is how I was forced by the marketing people into the logic group because there wasn't an opening for me in the microprocessor group, which I wanted to be in. Diamond: So why did you want to be in the microprocessor group? You got to Motorola. You worked on designing the modem chip. You worked on the processing for the bipolar chips. How did you decide that microprocessors were where you wanted to be? Mensch: It wasn't really microprocessors at the time. I was focusing on the fact that I love the design characteristics of MOS, but as it turns out, I got four offers. In every department that I worked in, they gave me an offer. The fourth one-- the fifth one actually, I'm sorry, would bring out five offers. The fifth one came from the high voltage bipolar organization that had the TV circuits. They wanted me because I could model the junctions, and so we could figure out what the breakdown voltages were, which was a very important characteristic. The reason why I didn't take the job is because, when I was told by Jerry, hey, Bill, you need to get off an extra month on this, and you need to pick, and so I needed to pick. CHM Ref: X7273.2015 © 2014 Computer History Museum Page 8 of 54 Oral History of William David “Bill” Mensch, Jr. Well, I wanted to work for the bipolar group. I sent them a letter. Never got a response. This was when you have interoffice memos sent around. We didn't have email, and so what happens is I never got a response. And so then the marketing group, the MOS marketing group, said we haven't really gotten circuits to work out of that MOS logic group, so we want you to work there. So they're the ones that said you need to take Bill. Well, meanwhile, I never got a response from the bipolar group. So then after I'm working for a week, I get this call from the head of the bipolar group that wanted me to work there. He says, Bill, where are you? I said in the logic group. Oh, he says, well, we want you to work here. We told you that. I said, yeah, but you didn't respond to my letter, my memo. And I thought, because you didn't respond, you didn't want me. So that's why I'm in the logic group. Now, I made a commitment. I'm here. He said, Bill, did you see the TV last week or the week before, when that tornado set down in North Scottsdale? Well, maybe. He says that was my house. I was putting my house back together. That's why I didn't respond to you. And I'm thinking a tornado changed my path. Diamond: So what did you do? Mensch: I worked in the logic group, and then what I did there was something that-- another one of these opportunities. Nobody wanted to design just logic gates. Motorola had a local logic, I called it loco logic, in Boston for some of the computer companies, and so you could design your own custom circuit with a standard cell library. So it was the job that I took over from somebody else that probably didn't really want to do it because it wasn't that much fun. But you had to put in the capacitances on the inputs and outputs. You have to have the drive capability and the performance. If you remember the old logic libraries that had their various characteristics that you had to have in a sheet that described the performance of the various logic gates. So I designed the local logic library for PMOS metal gate, PMOS silicon gate, CMOS metal gate, CMOS silicon gate, NMOS metal gate, NMOS silicon gate. So I learned all the processes, so it was kind of like, wow, now I know how it all works. So then because I did that, then they wanted me to design the process control monitor that monitored all of the wafers for the memory products and for the microprocessor products that ran on the NMOS CHM Ref: X7273.2015 © 2014 Computer History Museum Page 9 of 54 Oral History of William David “Bill” Mensch, Jr. enhancement load process that we designed our chips on. So then I designed the process control monitor that was stepped into the wafer in five different locations. Then you probe that to get to know whether the process was in spec or not, so then I designed that. Then I was asked to join the microprocessor design group, and that's why I worked on that. But then all the engineers were working on that, and then they needed somebody to do the peripheral interface adapter, so I was pulled off the microprocessor to design the peripheral interface adapter, which connected the microprocessing element, the memory elements, to the real world, driving solenoids, driving just reading external switches. And the PIA was used, also, for a keyboard interface so that you could drive some of the-- with some of the pins, you drive the rows, and then you sense which key is depressed. So I got a wonderful introduction to the whole system by doing that keep peripheral interface adapter. Diamond: So on PIA, what did you do? How did that project start, and what were your contributions to it? And take us through the-- Mensch: OK, so at Motorola, they had a-- like that applications department. They specify things. The marketing specified what they thought the industry wanted, and so as a result then, you would take that spec that was created in combination with the design group, but mostly dominated by the marketing group. So in some cases, that's a good thing. Some cases, that's a bad thing. The 6800 microprocessor was thrown together by a lot of contributors, so therefore the design group didn't really have much of the say about what went into it. And if a sales guy came back and said, hey, we need this instruction, so they added another instruction, whether-- whatever. So it was kind of a random thing, but it was built on the specification of the PDP 8, the 6800. The 6502 came later and was modeled on the PDP 11, better architecture. And now going back to my contribution on the PIA, somebody picked the numbers. I didn't pick the numbers. They said it was supposed to be 135 mils by 135 mils. So that was my challenge. So then I think-- Diamond: And what was the feature size? Mensch: Oh, it was 0.8 micron-- 800 nanometers? So it was 0.8 mic-- that means 4/10 of a micro-- I mean, 0.4 mils-- 0.4 mil? I got the number system right. The line width, metal line width, was 0.4. The spacing was 0.4. So it was 0.8 mil centers, so that was large compared to the future, where we are now. CHM Ref: X7273.2015 © 2014 Computer History Museum Page 10 of 54 Oral History of William David “Bill” Mensch, Jr. So we knew what those characteristics are, and so how they came up with 135 by 135, I don't know. Diamond: Maybe that was the cavity or something. Mensch: I think sales said we needed that to make money, and so-- no, it wasn't the cavity because the same package accepted to 217 by 210 mill, 6800. Diamond: So you got the outline of the die, the size of the die. Mensch: Size of the die. Diamond: And they tell you-- Mensch: Basically-- Diamond: Do they give you a pin out? Mensch: I don't know if I was given a pin out, or I chose the pin out, right now. But when you design a chip, on one side is the data bus because you're going to latch in information from the microprocessor. Then you're going to have I/O, so across the chip, you have a data path. That data path kind of tells you where you want to put things, so I don't know if I actually defined the pin out. I could have because I knew I wanted the flexibility to define where things were. Actually, maybe I did do the pin out because I got a patent on it. It was the first patent for the layout of an IC. Lawyers have called me from time to time, not recently, but they say how'd you get that patent? Well, my patent lawyer, Charley Hoffman, he put that together. But he thought it was worthy of a patent, and so I got a patent. The patent is that I place things that, if you changed it, it would be a bigger chip. I convinced the patent office that it was unique, and if anybody copied it they were violating our patent. So a patent was created-- the patent office was created in 1800-- well, 1800 is when Thomas Jefferson was president. He was the first head of the patent office, and I do have a patent-- the very first patent is on potash in my office. Charlie gave me that copy, but he made it look real. So going back to that is what a patent is about. It's about teaching someone what you know and what you've invented, so you get the monopoly. That's the word that's on the patent. You get a monopoly if CHM Ref: X7273.2015 © 2014 Computer History Museum Page 11 of 54 Oral History of William David “Bill” Mensch, Jr. you're willing to teach someone. So I was teaching them how to do a Peripheral Interface Adapter, so I got a patent on it. So that device, I chose every transistor on there. I taught the mask designer how to design integrated circuits, and when he said he couldn't do this one part, I stayed up-- after a party, I stayed up on a Friday night and finished the chip. He didn't have the advantage of knowing what he could do with the actual transistors and the resistors to make it work, so I figured it out. And I took it in on Monday and say I think we're done with the chip. Now, we just need to check it. Diamond: So what tools did you use to design this? Mensch: OK, so back in those days, there was no LVS, no layout versus schematic, so you had to check those things. And it was with Rubylith, so you actually had a flat bed with a knife on it and cut it out, and then somebody had to peel off the Rubylith from the places where you want a light to shine through. So those patterns were driven by a Calma-- it was the original Calma I system-- could drive a plotter, or it could drive a Rubylith cutter on a flatbed. So the simulation of it-- I was one that liked computers. So I was given, by a friend of my aunt back when I was at Temple-- he was an IBM supervisor, and when my aunt-- my father had passed away, and so my aunt said maybe this friend of mine, George Rogers was his name, maybe he knows what you could do after you graduate from Temple. And his one comment that-- I don't remember anything else. I don't even remember hello or goodbye or anything-- get into computers. So that's why, every chance I got, I used a computer because I had one piece of advice, get into computers. So that's what I did. So that means that, when I'm designing an integrated circuit, I want to use simulation because I knew the computer could help me figure out whether it was going to work or not. So I had to prove out that the PC control-- or the wafer-- the process control monitor matched up with the actual characteristics in the simulation program, so I knew whether that was accurate or not. We had to make some changes to model because of things I found. And I found that narrow transistors had a higher voltage threshold then a short one, and these are things that the memory product guys didn't use. And so they had to change their design because of what I found on the process control monitor. I put very narrow transistors, very wide transistors, very large transistors, and very short transistors, so I knew the characteristics and what the actual sizing might have an effect on. So I did that, and then I matched it up. So then I simulated everything. So I simulated everything, and I knew it was going to work. CHM Ref: X7273.2015 © 2014 Computer History Museum Page 12 of 54 Oral History of William David “Bill” Mensch, Jr. Diamond: What software did you use to do the simulation? Mensch: Motorola-- there was an engineer there. His name was Frank Jenkins. He, I think, graduated from Stanford, but I'm not sure if it's Stanford or Berkeley. One of those two. And so Frank wrote a program called TIME, T-I-M-E, stood for Transient, something, integrated, whatever, microelectronics or something like that, circuit analysis. So when he went to Motorola from Stanford or Berkeley, he wrote MTIME, so he just put the M front of it. And so that MTIME-- Frank and I became friends. I was taking the data in the lab. He was writing the software, then I'd tell him when the model had to change. We only had like less than 10 characteristics that we plugged into that model. It was very empirical. So I got to know everything about the transistors, what worked, what didn't work. Diamond: How complex was this chip? How many transistors or gates or what's were there? Mensch: In the PIA? I'd say 1,000 or 1,500, not that many. We still supply it, by the way. We're still getting orders for a compatible version with the original design. Diamond: And how many I/O lines did it have? Mensch: OK, it had 20 I/O lines. It had eight individually programmable ports on one write, so when you write with an 8-bit processor, you store your accumulator into a register. We had bi-directional control, so we had a data direction register and a data register. Intel I don't think ever had that, so we actually were selling PIAs with 8080 processors from Intel because they had eight bits out and eight bits in, but not individually selectable. So that really benefits if you only have a couple of bits, but you have more outputs or inputs or whatever so you can mix and match. NCR bought-- I don't know. They were a big purchaser of PIAs. Motorola would force them to buy a 6800. They were buying Intel processors but buying my PIAs, so Intel marketing thought, hey, for every 10 PIAs, you've got to buy one microprocessor. So I don't know what they did with the 6800s, but I think they tried to find a home for them, sell them or something. But they used my PIA with the Intel processor. So the other two-- so there's two handshake pins on a port, so the handshake-- on two different ports. So now we have two 8-bit ports that would be 16 pins, then two handshake pins per port. So when you wrote CHM Ref: X7273.2015 © 2014 Computer History Museum Page 13 of 54 Oral History of William David “Bill” Mensch, Jr. to the output, let's say, when you wrote to the output, you wanted to tell the rest of the world I wrote to the output buffer. Now, it's your turn to take it, so it's handshaking. And so then when that peripheral took the value of the pins, it would tell you that it read the pins, so then you can go back and forth. It's interrupt-driven or polling. You can see that this transfer had occurred, so that's basically a PIA. The other side of it is the interface to the microprocessor, eight data pins and register addressing, and then you have read, write, and phase two, and power supply. That's the 40-pin part. Diamond: What other chips did you do at Motorola? Mensch: Well, my patents-- I designed the buffers that were used on all of the parts. Diamond: I/O buffers? Mensch: I/O buffers. Clock generators-- I designed all the clock generators. And so the individual circuits- because of my work with the standard cell library, I knew how to pick the transistor sizing because you look at the loading that you're going to have on a data path or a wire, you can calculate the capacitive loading. If there's resistance in it, you can put that in the model, simulating it. So all of that, I knew, and that's what I-- all of the chips at Motorola used some of my circuits. Diamond: So you were kind of the analog guy in the digital group? You understood the dynamic issues with the cells. Mensch: Yeah. In fact, much later after when I started my company, to survive, I taught Xerox how to design microprocessors, and that was a big part of what they were missing. They didn't know how to really put a microprocessor together. I showed them how to do that. Because you have to use some basic physics-- what capacitive loading do you have on it? And if you don't know what the capacitive loading is, you don't know what transistor sizes will be required to switch in a certain timeframe. So, yeah, I knew the characteristics of the transistors. Diamond: So after the PIA, what did you do next? Mensch: Well, my next chip was to design a peripheral-- I mean a programmable timer unit. But it was in the middle of that that Chuck came a ‘calling and said, Bill, do you really want to go to Austin, Texas? No, CHM Ref: X7273.2015 © 2014 Computer History Museum Page 14 of 54 Oral History of William David “Bill” Mensch, Jr. I don't really want to go to Texas. So he says, hey, how about if we get a group together, and we design the microprocessor we really want to do. And you know that the 6800 isn't the right one. It's too big. Right, Chuck, I know it's too big. So Chuck said who do we need to be successful? Who can we take out of Motorola to be successful? I'm going to look for some place that we can land, and we'll design the microprocessor. And that's what started my thinking “Motorola isn't going to be the place I'm going to retire at” But another thing happened. At Motorola, I made a bet with the Operations Manager of the MOS group, Otis Wilkins. Otis ended up at GTE heading up their semiconductor operations later, but Otis said-- was it Otis? Maybe I have the wrong guy. But anyhow, we had a party for releasing the microprocessor and the PIA at the same time to manufacturing, get the samples first. And so what I said, I'll bet-- now, this is after a glass of wine or two or three at a favorite restaurant. Everybody's celebrating. I'm saying my chip will work the first time, and he says that's an easy bet. No chip has ever worked the first time. So what's the bet? And I said how about dinner? What do you think? You and your wife, me and my wife, dinner. He says, OK, where? I said, well, for me, the Wharf in San Francisco. We're in Phoenix. He says, well, I know a restaurant in Scottsdale with a $100 bottle of wine. I go, OK, you're on. My chip came out and worked. Well, why did it work? We didn't have LVS [layout vs. schematic]. What I did was I checked it five times. You make a bet, you better back it up. I wasn't wanting to buy him a $100 bottle of wine anyway. Well, he never paid off. So when somebody doesn't pay off, then you go is this the right place for me? And then another thing was said to me in a meeting with AMI, which was our second source. They said if Bill's chip works the first time, we're going to give him the “first” engineering bonus. Whoa! And John Buchanan, he's passed away now, John's like what? What about me? He was the guy that was in charge of the project leader of the 6800 in this meeting, so he wasn't happy with that because he thought probably that he deserves a bonus just because he held the team together to get the project done. Chip came out. Worked the first time. Never paid off on the bonus. CHM Ref: X7273.2015 © 2014 Computer History Museum Page 15 of 54 Oral History of William David “Bill” Mensch, Jr. Diamond: So at that point you were skeptical about how successful the 6800 was going to be? Mensch: No, no, I thought that would be successful. No, it was the fact that I needed to leave because, not one, but two promises were broken. That was powerful. So who do you think I meet in the hall at 52nd Street? Jerry Fulton. Bill, I heard you're leaving. I said, yeah, Jerry. I've got to go. Diamond: So then what happened? Mensch: Went to MOS Technology, so now I'm fueled by “we need to make this successful”. Diamond: So what was MOS technology like when you showed up? Mensch: Oh, my gosh. That's a good question. It was run by calculator people out of Texas Instruments. And so the operations guy-- I forget what his name was, but he was a very successful Texas Instruments production manager. So he was like vice president of production. And the people that were designing the calculator chips, they were PMOS metal gate. That's what I think TI had at the time. And so they were very successful in the calculator business, and in fact, the-- oh, almost had his name. His first name was Don. But he came up with a thing called “spot knocking” the masks. So they had perfect masks, but nobody else had perfect masks. So they had a great manufacturing team from a TI background and had a great logic team designing calculators that were really significant. They're scientific and everything, so they know what they were doing. So Don McLaughlin was the vice president of engineering. He was one of the founders. So what happened when we got there, they didn't want us there. They were saying, well, how could John Pavinen and the founding president, how could he hire this group out of the west? They don't know what they're doing, and we know what we can do. Why aren't we doing it? So they weren't happy that John Pavinen had brought in a team to do a microprocessor. CHM Ref: X7273.2015 © 2014 Computer History Museum Page 16 of 54 Oral History of William David “Bill” Mensch, Jr. Now, the actual 65 where’d that come from? I guess it's a lower number than 68. I don't know where the 65 came from. And then of course, you start at 00, then 01, 02, like that, but 00 isn't too interesting. So the first microprocessor was the 6501. Diamond: And how big was the team that came over with you? Mensch: Eight of us, and then we hired one. I think we hired one back there. So we had two mask designers. We hired one mask designer, so we had three mask designers on the project, Harry Bawcum and Mike Janes, J-A-N-E-S I think. And then Terry Holt was the operations guy. And then we had Chuck Peddle was the marketing, and he was the ringleader. He was the business guy. He was the guy that said, if we do this, we can create a good business opportunity. So Chuck was the marketing guy, and he kind of controlled the opportunity, so to speak. He put together the package for MOS to the contract. And so then there's myself, Ray Hirt and then-- how many do I have there? Cid Holt was one that we hired that was back there. So there's a picture in EETimes of the group that designed the microprocessor and the product. Diamond: Did they have an MOS process when you joined? Mensch: No, so we had to bring up an MOS process, which-- going back to Motorola days, which was another reason why I left Motorola, is when you want to make a difference, and somebody stands in the way for no apparent reason, you kind of say what's going on here? It's confusing. So what happened was, because I had the process control monitor responsibility, I decided that I would have a line with the same lattice structure and everything. If you know a little bit about the physics of the lattice, the mobility can change depending on which way it's oriented subtly. And so I had depletion load process and enhancement load process on the same process control monitor, so all you had to do is use the depletion mask so that you could implant for the depletion thresholds for the load device. It was a much faster process, much easier process, less transistors. It was the way to go. Well, Chuck saw what I was doing, so when we went to MOS technology, we specified what we wanted. It's what I wanted at Motoro