Third session

Smithsonian Oral History

In 1972, Joseph Desch, Robert Mumma and Don Eckdahl of NCR were approached by the Smithsonian Institution and asked to donate their recollections to its History of Computing Project, now part of the National Museum of American History. Desch and Mumma were interviewed on the 17th and 18th of January, 1973 by Henry Tropp.

This section of the Desch/Mumma interview runs from page 87 to 166 of the first day transcript

[Page 87]


HT: .. You were going to correct that statement you made earlier about the binary multiplier, Bob.

RM: Yes. That – it was the co-invention between Joseph R. Desch and myself, Robert Mumma. And the Patent number is 2,404,697 and it – the application was made March 21, 1942.

HT: Aha. And I see the date of issue is July 23, ’46.

RM: 1946.

HT: Mhm. You want to describe that in a little more detail now that you’ve got the patent in front of you?

RM: Well, it was basically the same machine in a binary sense that – as the second multiplier model that I had built.

HT: Mhm.

RM: However, this device – the only readout we had was to see the glow in the — in the gas tube.

HT: Mhm.

RM: We had no – at least the patent itself didn’t cover any read-out in the decimal system.

HT: Mhm:

RM: But, you could determine by knowing the – what power of two each tube represented, you could

[Page 88]

identify the product or the sum

HT: Mhm.

RM: by looking at the glow of the gas in the tube.

HT: But you would read it as a binary number which you could then convert to a — a decimal.

RM: Convert to – mainly to a decimal number.

HT: Decimal number. You know —

RM: Go ahead.

HT: I’m sorry. Go on.

RM: I was through.

HT: Oh. What I was going to say let’s — let’s consider the people who are going to be looking at, or listening to this tape or reading the transcript, and let’s describe in — in very general terms, between the two of you, how one would do an addition or a multiplication or a subtraction on each of these early prototypes. How you would get a number in and how you would read it out. In very simple terms, not technical terms. What kinds of settings you would need and so on.

JD: Well, on the first model that Mr. Mumma did after — after being stimulated by my book over there, .. I’ve already explained that.

HT: Yeah, yeah.

[Page 89]

JD: That one is on the record.

HT: Right.

JD: Early record.

HT: Right.

JD: Now the second one that he built, which was .. an NDRC project, with miniature tubes,

HT: Mhm.

JD: was also based on the same principle as that bigger machine. And I made mention of that. These are all additions and subtractions.

HT: Right, right.

JD: Because all we do is reverse the order of impulsing.

HT: Mhm.

RM: In the third calculator model which included addition, subtraction and multiplication, subtraction was initiated by depressing the subtract key which energized a bank of relays for each denominational order in the accumulator. These relays interchanged the connections between thyratrons so that each conducting thyratron conditioned the thyratron behind it to accept an input pulse causing the counting ring to run backwards. The transfer pulse to the next higher order was also reconnected to cause the “nine” thyratron instead

[Page 90]

of the “zero” thyratron to generate it.

HT: Mhm.

RM: [for subtraction] The ring would run backwards.

HT: Okay. Suppose I wanted to add three ..

RM: Okay.

HT: say two digit numbers in — consecutive order.

RM: Okay.

HT: What would be the stages that I would go through in the second machine with the miniature

Well –

HT: thyratrons?

RM: Well, first you would clear the machine, causing, all the zero tubes in the accumulator to conduct.

HT: Okay.

RM: Then you would depress the add key and then the key representing the number “three” and press the motor bar to start the machine. The three pulses would be fed into the units order of the accumulator.

HT: Okay.

RM: Now you would have the number three tube conducting and they would remain conducting. Then, .. you would set the number five, for instance, in the units order of the impulse generator, press the

[Page 91]

motor bar. Five more pulses would be fed into the machine, and now conduction moves from the number three tube to the number eight tube

HT: Okay.

RM: and the indicator which indicates that you now have an eight representing tube conducting and you had an indicator that was looking for one voltage difference from all the other voltages. The conducting tube had a different voltage. The indicator would stop at that point and represent the digit eight.

HT: Mhm.

RM: That’s the simple …

HT: And then when I put the third one in, it would just continue the operation and we could do this for any finite number of digits.

RM: Any — any number.

HT: And then when we got the final result, then you would clear the machine again for the next operation.

RM: Yes. Now if you had multi-digit numbers you would send counted pulse groups to each of the selected denominational orders of the accumulator, corresponding to the denominational orders of each of the digits in the multi-digit numbers. The counted

[Page 92]

HT: Okay. It would get you into the second bank.

RM: Yes, and come back up to three, yes.

JD: You know.. .. something dawned on me just now.. which can just take a second.

HT: Sure.

JD: Some of these large-scale .. digital computers – the original ones at least – were working on a string of impulses very much like what he just now talked about.

HT: That’s right.

JD: And the reason that it appeared so fast is that they — they were running too, close to a megacycle. But the early ones were — were much less than a megacycle.

HT: Yeah. Well, I know the first EDSAC mach- the. EDSAC machine that Maurice built was about a half a megacycle,

JD: Yes.

HT: which had a mercury delay line memory. And really the first stored program computer that was built that was operational, and he scaled it down so that, you know, it would work, and it was about half a megacycle. As you say, the — the speed was such that it, you know, it was hard to tell

[Page 94]

what was going on internally except I guess you could listen and tell where things were.

JD: Down at the Bureau of Standards, and I forget his name, the fellow that ran that down there – he was a friend of mine, too. I can’t think of – I can’t recall it now. Oh yes, it was Dr. Alexander. He had a radio set sitting across his table and we — we could listen to what was going on by the radiation from the computer, and it was around a megacycle. But, as it shifted, it just played a musical piece for us over there, on a radio.

HT: Aha. Was this the SEAC machine that was built for the Bureau or was this another one?

JD: Yes, it was SEAC, built by the Bureau of Standards by Dr. Alexander. I used to listen to the musical tone coming out of that computer. [Laugh]

HT: Yes. They could pretty well tell. Now, how about the binary machine that you just discussed. Even though a prototype was not built, .. would you care to describe that operation? Would you have to manually convert the number from decimal to binary before you put it on or could you conceive of it as going on in decimal and being converted to a binary mode?

[Page 95]

RM: In the — in the patent, as written up, let me check it – my memory is that .. we .. set the number up as a two to the zero power, two to the first power, two to the square, two to the cube and so forth. And – so that – you didn’t have to press the keys all at one time. But, if you were going to enter the number three, for instance, you’d hit – have to hit two keys as,

HT: Right.

RM: two to the zero and two to the first power keys,

HT: Mhm.

RM: and one pulse in each line would work the accumulator. And, there you’d have the two tubes conduing which represented three. Then if you would, … for instance, we’d close the two to the first power key again and add a two into this device, you wouldn’t disturb the two to the zero power tube in the accumulator but you would .. .. in effect extinguish the tube that was conducting representing the two to the first power and cause two to the second power tube to conduct. So now you’d have two squared and you’d have two to the zero power. Five would now be represented in the accumulator.

[Page 96]

HT: Right.

RM: much more simply. It became apparent to me at that time, that in business applications, till you add the decimal to binary decoding device to go from a decimal keyboard to a binary entry and then at the output from a binary to a decimal output that you add so much complication and cost that you wiped out the benefit of the reduction in number of tubes. It just did not really pan out.

HT: Mhm.

JD: Now, what we finally did then in our decimal systems, instead of having a bank of tubes for each denonational order of the impulse generator, we used a row of control tubes and a multibank keyboard – to control a single bank of pulse generating tubes. This pulse generator was then switched from bank to bank of the accumulator in sequential order at electronic speeds. And so with twenty-five tubes we did what we did in the original model with maybe seventy-five tubes. This circuit is described in U.S. Patent No. 2,398,150

HT: Mhm.

RM: And tricks like this were done in order to conserve tubes.

[Page 98]

HT: Mhm.

JD: .. May I add something here? It’s pretty hard to explain to someone like you right now, in a very short time, of the gyrations that we went through. Remember, we were in a brand new field, and we–we, we didn’t know which way to go, and we never knew when we had the optimum.

HT: Mhm.

JD: So we were doing like — we were doing in the gas tube field what MIT was doing in their field.

HT: Mhm.

JD: We – as soon as we’d get a new idea, we’d go and build a model of it on a breadboard or like a paper patent to get it under cover. And there was an awful lot of ’em in – in these – the patent books that we got. These are just – is this all of ’em or is this — ?

RM: These were all of the patents I am involved in right here.

HT: Mhm.

JD: And some of them were originated just because we got the idea

HT: Mhm.

JD: and we wanted to try it out to see if it worked.

[Page 99]

We didn’t know if we were going to use it or not. The decision to use it or not to use it would be the time when we would put a new system together.

HT: Mhm.

JD: Then we would have available to us the bits and pieces that we’d already covered. Now, .. MIT was doing the very same thing.

HT: Aha.

JD: But, we — we took a little more urgent approach because .. we were — we were a little more definite in what we wanted to do than .. MIT, I think.

HT: Right. .. Well, the university environment is very different than a commercial operation, where you already have a .. continuous thing going and patents are just a part of your environment.

JD: That’s right.

HT: Speaking of environment, and this is one of the questions I wanted to get to .. was the — the kind of environment that you were operating in. You mentioned your personal input of this one key paper of Wynn-Williams’ .. on the counting. .. circuits. .. What other kind of literature or contact or outside stimulus were you getting in. I terms of ideas?

[Page 100]

JD: None.

HT: None?

RM: Of course, The Eccles

JD: No contact — no contact with other people doing similar work except at MIT.

RM: Of course the Eccles-Jordan trigger pair was in the literature at that time.

JD: Well, yes.

HT: Yes, yes. Yeah.

JD: Anything that was published

HT: Right.

JD: was available to us.

HT: Yes. And I guess we might talk about some of the published literature that people often allude to and, in most cases, you discover that they don’t read it until much, much later. And, we can go all the way back to the .. work of Babbage, and Boole’s work on binary arithmetic and .. see whether that was part of your intellectual input.

JD: No. At one point I was acquainted with the Babbage machine. In fact, Aiken had a — a model ofit.

HT: Yeah, but this was -would have been much later.

JD and RM: Oh yes .

HT: How about — how about Claude Shannon’s work on —

[Page 101]

on switching theory at Bell Labs?

JD: No — I knew Claude but again, it was after the fact.

HT: Aha.

JD: I got acquainted with him later, on committees, and that was all. But I had no input. We were operating completely alone except for IB– MIT.

HT: Mhm.

RM: I suppose, part of it, is why they look so much like mechanical machines, I mean we were simply adapting mechanics to electronics. That’s all we were doing really.

HT: Well, in this, I think, true of all the first machines. I mean they – you can – you can find them in the environment of a person who did it in terms of what he was familiar with. And, as I look at — at ENIAC, it’s clear that John Mauchly had a lot of experience with Marchants and Fridens, the mechanical calculating devices that astronomers and mathematicians and statisticians used in that time period. And the influence is — is very obvious in — in terms of the kind of work they were doing, because it was an attempt to mechanize the same kind of thing and — and get the speeds

[Page 102]

up and get the output –

JD: Using a different technology.

HT: Sure. Sure. But – .. .. let me throw a philosophical question out, because we’ve really stayed kind of technical and, and you know, close to the things you’d done at the trial. But, in terms of the new technology, .. let’s see if we can’t get back into that late thirties and early forties period. .. How did you see the state of the art in that time? Or, how do you define even the concept of what people often call the state of the art? There are — there are many definitions.

JD: .. Could you — could you narrow your question a little bit? What you mean by the state of the art?

HT: Okay. .. Let me — let me recount –

JD: I’ll tell you what I had in mind.

HT: Okay.

JD: The componentry used was — was standard state of the art. It was simply new combinations of — of — of available components with the exception, in our one case, of the little gas tube. But the, rest of it was off-the-shelf items. It was combinations and circuitry.

HT: So it’s primarily what you’d call off-the-shelf

[Page 103]

items. It was combinations and circuitry.

HT: So it’s primarily what you’d call off-the-shelf items but used in a totally new way.

JD: That’s right.

HT:In a way that

RM: New combination.

HT: maybe didn’t violate any fundamental laws,

JD: Well, you gotta –

HT: but in a way that nobody had tried before and you didn’t really know if it — if it would work but nobody could convince you that it wouldn’t.

JD: No.

HT: It’s that kind of philosophy and that.. kind of definition is one that Bob Everett articulated to me in — in terms of his definition or – of state of the art as opposed to straight off-the-shelf, it’s there, you get it, and you put it in and use it. It really wasn’t that first, it was reallyI mean the second – it was really more of the , first, wasn’t it. Tha- that you were probing new ways of using the technology?

JD: That’s right. Well, probing new ways, .. probing. new circuits, basically.

HT: Aha.

[Page 104]

JD: Really, if you really get down to it, we were dealing in circuitry

HT: Right.

JD: more than any other thing at that time, and we had to use components that were available except in the case of that one little tube.

HT: Yeah. Well, the one little tube then you started off from the standpoint of size, characteristics and, as you said, empirical testing until you could finally get something that would work to your own satisfaction, so this was a, just a very hard, laborious approach to solving a problem.

JD: Well, it wasn’t that we didn’t do any computing, computations and design, in our designs, our circuitry, that is. We.. .. our records will show that we — we did an awful lot of – besides empirical work – we did an awful lot of computations to determine optimum condition circuitry additions, but not all. We didn’t know all the factors that would – were influencing this all the time. We were surprised many times with — with some simple, little thing that was preventing us from accomplishing our …

HT: Do you have any comments on that subject, Bob?

[Page 105]

RM: No, I think Joe covered it pretty well. We used standard components. The resistors in those days would drift, when they’d get warm they would change resistance, and we didn’t have as good equipment to work with in those days as we have today, so it made it much tougher when you were trying to hold tolerances tightly, with this device, you see, every tube circuit had to be like another one so

HT: Mhm.

RM: we did an awful lot of calculation to get the optimum with the tolerances that we had to work with.

HT: I guess that was, in terms of a technical question, that – that’s one of the problems of getting tubes that are made one at a time uniform

RM: Yeah, that’s right. .

HT: so that once you design the circuit, a single cicuit that was going to be repeated many times throughout the design you knew that each one of ’em had the same characteristics and one of ’em wasn’t going to go way off on you.

RM: Of course, we had tube checkers that we made for these thyratrons. We knew what the firing potential

[Page 106]

would be in everyone of them and we had tolerances for that. And the deionization time, we also had that checked out so we knew what that was to be. And as long as they matched, why we could use the tube.

HT: In terms of .. some of these devices that we started off with in the — the photographs that were delivered to places like Aberdeen and the University of Chicago, what kinds of reports were you getting back on the reliability of the tubes and their life expectancy and so on? Were you getting data back in terms of the application?

JD: Not that I know of.

RM: I think I went to Chicago once to maintain the equipment – that’s the only time I remember.

JD: Did you?

RM: Yeah, and I forget now what I did, but they brought it out and I worked on it in a room separately and then handed it back to them.

HT: So if if — if it broke down, you were really the only ones then that could get it going again?

RM: That’s — that’s true.

HT: You didn’t stay there and train people and supply them with spare parts.

[Page 107]

JD: We were meticulous about the fact that we wouldn’t deliver anything unless it was in the very best order, even it it took a little longer. And I was surprised about this — this communication [device], for instance, where we used a seven megacycle carrier and a forty mega– forty megacycle .. a 40,000 cycle modulation frequency. That thing just worked for years and years and years down in Washington.

RM: Good, I didn’t realize that.

JD: Oh, yeah;

RM: I don’t really know.

JD: and I just couldn’t believe it.

HT: Aha.

JD: And it had experimental tubes in.Those were those ten-section tubes.

RM: Mhm.

JD: that were in there, you know.

RM: Little storage tubes, yea.

JD: Storage tubes; and we — we even built. ten-section counting tubes, but we never could — could do as well as we — we wanted to do.

HT : Mhm.

RM: That’s right. I think we went to the five-section

[Page 108]

counting tube or something, because it was pretty hard to get this differential between sections in a ten section tube. We would use the glow in one section to condition the next section making it responsive to the next pulse on the input line.

HT: Mhm.

RM: The part similar to the glow.

HT: [Laugh]

RM: We had to keep it going forward instead of backwards

HT: [Laughing] So the glow was uniform.

RM: Yes. These were the difficulties with gas tubes. We ended up with a configuration to produce the desired result.

JD: In the case of storage tubes, it was much easier – to do.

RM: You just fired the section of the tube at the intersection of two wires.

JD: Because you just hit the grid that you wanted to hit, there was only one of ’em going to glow.

HT: Mhm.

JD: And.. you didn’t have to count up to it.

RM: And anyone you would cause to glow would extinguish any previously glowing section. You never would have more than one section glowing at a time.

[Page 109]

HT: Mhm.

RM: So that was no problem, with the storage.

JD: The Navy wanted 40 of those real quick. And –

HT: I thought –

JD: the last I heard they were going to have Philco build ’em, because the Navy wouldn’t let me build them.

Leave page 109 — text struck by Desch included here

[Page 109a]

HT: That–that began-

JD: They just took us–they just took us over. That began when?

RM: Forty-two, I suppose. Wasn’t it? Forty-one? I think it was as early as ’41 because I was off that-

JD: No, it had to be ’42.

RM: Might have been ’42, yeah.

JD: . . . because, yeah, it had to be after December.

RM: After December 7th, yeah.

JD: Yeah, and I don’t think-

RM: That was–that was December 7th, ’41. I was on Indiana sea work before that but we hadn’t gone to–we hadn’t gone to the Navy yet.

JD: The Navy is in ’42.

RM: Yeah, that was in ’42, yeah.

HT: Well, once you got into this top secret work for the Navy, then, I take it, you dropped all these other projects and relationships.

[Page 109b]

JD: Well, we were forced to.

HI: and went underground essentially?

RM: Yeah.

HI: And that involved…the two of you and your whole team that was here or-

[See the names of the team here]

JD: Yeah, our whole team of 20 people.

RM: Yeah.

HI: Twenty people?

JD: [Chuckle].

RM: Which expanded to 800 people before we got through with it.

JD: Eleven hundred.

RM: Eleven hundred, I thought these were–[Laughter]

HI: I’m going to ask a couple of questions in that area that, of course, you know, if don’t want to you’re not going to answer…lt’s pretty well documented in the literature about a- an electronic cryptanalytic…machine that was built in England and completed, I guess, about 1943…Some of the names of the people that I know who were on that project are M. H. A. Newman, a gentleman named Flowers,.. I. J. Good, I think Turing was on that … project. Were you in any way connected with that work, or were – was yours totally independent?

JD: It was a one-way street. The — the British came over and visited me and looked at everything I was doing, but I never could see anything.

HI: Aha.

JD: they were doing. And when you mention Turing; he was one of ’em that came over here very frequently.

HT: Mhm. Turing made a number of visits here. Now, again we have to stay- you know the areas we have to stay away from…much better than I do, and I’m interested in links and ultimate impact rather than in detail, so we’re not going to discuss any details of what you did. And-

JD: I had very little contact with the — with the technical people in England. At one time they were going to send Wynn Williams over and they didn’t; but Turing was over quite a bit .. , but high level people came over frequently. Admirals and even houses of … Members of Parliament and Lords of Admiralty and, Lord, I always ended up with them at my house. [Laughing]

[Page 109c]

HT: I’m interested in the names. But before we get into some of those others-

JD: I can’t give you the names of those men.

HT: But before we get into any of those others, .. I would like to talk about Alan Turing, and let me give you my interest and background and then you can decide how much, if anything, you want to say on it…I mean Turing has been dead now for quite a number of years, .. more than a dozen or so … He was involved with the Institute for Advanced Study in the thirties when he was doing his doctorate in logic under Church and there met Von Neumann, who had asked him to stay, but the war was breaking out in England, so he went back, I guess, in ’37, or ’38, with the imminence of war. And, of course, Turing’s name iso-is very well-known today in terms of the Turing machine and the principles that he enunciates in a paper that he wrote while he was at Princeton; essentially his doctoral thesis. But there seems to be a, you know, a total gap in the period between when he left Princeton and later when he is involved in building a machine in England in the post-war period. And, again, his trips to the United States are known about, but I’d appreciate any illumination, any comments you want to make, any conversations you recall, – with Turing, without getting into technical and secret information.

JD: Afraid I am not allowed to do that.


JD: That-that’s one of the things that I-I’m quite sure the Navy would clamp down on me if I-

HT: OK. No, as I say, I’m–I’m–I’m sensitive to these things. I – what I’m really looking for are personal comments in terms of your conversations about Turing and his level of interest in electronic computation which is outside of this area.

JD: I can’t–I can’t –I can’t talk about it.


JD: Really. Because-I’ll just get-keep getting in deeper and deeper ifI start. [Laugh]


JD: But the British did send representatives over here repeatedly and they weren’t all technical.

HT: Right.

JD: There were just a few technical people.

HT: Right

[Page 109d]

JD: Because the British didn’t believe that I would be successful, and so reported, I found out after the war. I saw the reports.

HT: Well, I’m going to assume that you-

JD: The Navy–the Navy wouldn’t show me the reports during the war.

HT: [Laughter]

JD: Because they were afraid they were going to destroy my morale.

HT: [Laughter]

JD: [Laughing] After the war, they showed me the reports and they weren’t very complimentary.

HT: Well, I take it the reason they showed you the reports was because you were successful and we’ll just drop the subject right there. [Laughing]

JD: [Chuckling] The British must have had to swallow hard because we had to build 100 machines for them.

RM: Of course, they didn’t know how we were going to do it, brute force. They never dreamed they might do like we did.

HT: But, you had an actual manufacturing plant when you say 800 people, so-

JD: I had a brick building over there.

RM: Eleven hundred people.

JD: Yeah, well, you see, I had–I had 600 Waves and I had 100 officers and enlisted men and the balance were all civilians. And the whole total sum of the whole thing was 1100 and they all reported to me.

RM: And we manufactured devices and shipped them out. We had ..

JD: We shipped an awful lot.

RM: railroad cars backed up there and had, of course, many yards all around.

HT: Is there any identification phrase, number, code name, anything, that’s not classified that you could tell me, that say, 10 years from now if somebody wanted to see whether it–was declassiliable, could be used as a -a identification.

JD: Can’t give you ..

HT: OK. I keep pushing, as you know, and I don’t want to get into anything classified and I know that you’ll make sure not to divulge anything.

[Page 109e]

JD: No, the code name is classified.

HT: Code name is also classified. I think one of the stories that is not classified but makes a marvelous story is the burying of material someplace in this parking lot here in Dayton, Ohio,

JD and RM: [Laughing]

HT: which I’d like to have on this tape if you don’t mind telling the story. Both of you tell the story. I think that’s a marvelous story.

JD: Well, it is true that there’s-that there’s–there’s some–some material that is sensitive that is buried pretty deep.

RM: Tools and dies, though?

HT: Aha.

JD: that–that could–could lead someone who is getting very snoopy about it, it could lead them right into the thing, so they’re–they’re buried and by this time, they’re rusted out.

HT: Oh, I’m sure.

RM: They must be.

HT: I’m sure and they would-somebody would have to dig up the whole area. [Laugh] I gather the way it was-excuse me-the way it was paved over nobody would ever locate where it was.

JD: No.

RM: That’s true.

JD: Well, anyway, the Navy did a different-did it a different way when–when they received a lot of classified material from us. They didn’t know what to do with it either, and they just took it out in the Chesapeake Bay and dumped it overboard.

HT: [Laughter]

JD: [Laughter] In deep water.

HT: The early–early polluters of Chesapeake Bay.

JD: Well, that’s how they got rid of some of their stuff.

Return to Page 109: text struck by Desch ends

HT: Let me turn this off and give us a chance to stretch a bit.

[Recorder off — end struck out section]

HT: Now we were — we were talking about Vannevar Bush and you made a marvelous statement about Bush as a visionary in this whole area and I’d like it in your words, without getting in, again, to sensitive toids.

JD: Yes. .. The point I was trying to make was that .. that there is a man that I don’t believe that he’s received the acknowledgment for his visionary work in computers .. that probably hasn’t been. contacted or interrogated and that is Vannevar Bush, whom I am sure has a voluminous amount of of conceptual work .. that is the forerunner of much of the — much of the computational work that been done in the last 30 years.

[Page 110]

HT: Mhm. I’m glad – you know one of the reasons I wanted that on tape is that I recently read his book Pieces of the Action, published three or four years ago. And he took a rather modest position in this area and he said, for example, he would recount an early magazine article, a popularized article in a, say, Saturday Evening Post. That may not be an exact source, but that kind of an article, in which he would talk about things in general and the wave of things to come. And he said people often took those articles and then gave him credit for originating these concepts or .. anticipating something by 20 years. And he said he didn’t feel that he deserved that credit. Now I have the feeling that it’s just the opposite.

JD: [Laugh]

HT: That’s why I wanted to get you on that subject,

JD: Yes.

HT: and I wonder if you care to expand on that at all.

JD: I don’t care to expand on it because I promised Dr. Bush that I wouldn’t reveal any of the infomation that I got from him. I was privileged to see a lot of it and I have never violated my trust,

[Page 111]

and I have never used any of the information in any of our work, ..

HT: Mhm.

JD: because I was honoring my pledge to him that I wouldn’t do it. But I, he — he was certainly extremely knowledgable and was very far-sighted,

HT: Mhm.

JD: and much of it has come true.

HT: Yeah, well, in — in the case of the work that is known the name Vannevar Bush is very, very big and, as I say, all additional

JD: [Laugh] Like the atomic bomb, huh?

HT: Well, I’m just going back to the differential analyzer.

JD: Yes.

HT: .. Even farther back, there was a picture in the IBM history wall of a profile tracer that he built which was an analog device in, oh, I don’t know, the early decades of this century and he’s, in a sense, the — the analog computer .. became a reality and a useable device for computation and the solution of differential equations thanks to his work.

JD: That’s right.

[Page 112]

HT: And that one thing alone would have made anybody’s reputation and I say that Vannevar Bush’s status .. as one of the major scientific figures of this century is one that nobody would question. And. all you’ve done is add a whole new dimension to it.

JD: I agree.

HT: Okay. I’ll turn this off now. [Recorder off]

HT: .. We’ve returned from our luncheon break and I thought we might start by talking about the Electronic .. Research Laboratory as it evolved from various places in this building that we’re in, Building 10, to it current place on the socalled second floor of the one-story building in what’s that, Building 26? Is that the right number?

JD: …. Buil-

HT: Where is it now, is it in Building 26?

RM: No.

JD: No, that’s Building

RM: 20.

JD: 20.

HT: Building 20. Sorry. So you might — you might

[Page 113]

say something about when it first started and what you were doing and – oh, kind of trace it in an evolutionary way, and, after we’ve done that, I’d like to break off and get some of your anecdotes and stories that we talked about at lunch.

JD: Well, .. I’m now examining a — a notebook of Joseph R. Desch, which was started February 14, 1946 and it’s known as “Book Number 1” andthat I noted on it – but it also bears the number “Book Number 69-7” which is another category given to it.

On — on page 100, in fact on page 98, 99, 100, and 101, are diagrams, diagrammatic sketches . of the pumping station that was used to make .. thyratron tubes. And page 98 and 99 is a diagrammatic sketch; and then on page 100 are two photographs, both of which largely depict — depict the.. pumping station as it exists today with the exception – and also page 90 – page 101 is more of a frontal.. .. picture .. truly representative of what exists today – with the exception that there is missing on those referenced photographs an addition that shows up on page,

[Page 114]

on – it shows up in Book Number 2 with the Electrophysics Laboratory .. which was started in September 17th, 1948 known as “Book Number 6298” which shows an additional device that was added to that pumping station, that – which is really an oil manometer, an oil McLeod gauge which is a – had its own oil purifying system to enable the oil to be used in the arms of the gauge to maintain high purity in — in those arms. And that — that was posted on this book on June 1, 1949. ..

This is made to supplement any previous photographs that have been made. .. ..

Page 88 and 89 give a description of this oil manometer. And [pause] and subsequent pages in this book give a description of it performance. Notice that another, on another page, 101, is another photograph of the same manometer, .. a water cooled, self purifying manometer which was posted on July 29, 1949. And, and countersigned by Mr. Evelyn Vaughn Thompson and Harry Williams, and these are additions to the original ones on Book Number 1.

HT: I was just pulling out here the very next book,

[Page 115]

Book Number 3, to see if there were any further additions in terms of photographs, .. but there doesn’t seem to be. This is also labeled “The Electrophysics Laboratory” and its dates run from July 29 of ’49 to November 23 of the same year, and it’s a continuation of Number 2 where you must have used up the last page. So, really, Book Number 3 is a, the first entry is a continuation of the last entry in Book Number 2 .. which talks about degasing the high vacuum manifold. Is that — is that right?

JD: I’d like to see that.

HT: Yeah, HV. Maybe I misinterpreted that.

RM: High vacuum.

HT: High v- high vacuum. I’m sorry. Thank you, Bob. [Laugh]

JD: [Reading] … 1949. ,Last entry in this book November 23, 1949, Book Number 6299. This seems to continue from Lab Notebook Number 2, the same date.

HT: Right.

JD: Now, this, there is nothing in here that I know of that describes the — the pumping station put I it is full of information regarding the design

[Page 116]

of tubes that I had in mind .. in an effort to reduce the complexity of a counting station.

HT: Mhm.

JD: And a study of all of this information

RM: There were no photographs there, were there?

JD: will show that these counters, which are patented and which appear in a group of patents to Joseph R. Desch, .. contain no resistors or — or — or — or condensers in the — in the — in the counting stages, because a — a different principle was used to prime the next stage. And that was the intent. It was the intent to describe the –

RM: Weren’t they cold cathode tubes, Joe, pretty much?

JD: These were all cold cathode tubes.

RM: Yeah. ‘Cause you were pushing the cold cathode to get rid of that filament power problem.

JD: Yeah.

HT: Wihile you were going through that, I just opened this.. Notebook Number 1 of L.A. deRosa, .. and it’s dated from December I, 1938 to June 14, 1939. And before I talk about what’s in some of these notebooks, you might say something about DeRosa. You indicated his .. subsequent rise.

JD: deRosa, deRosa as I recall, and I – I’m not sure —

[Page 117]

HT: Well, you might start with how he joined you and . — and development as you remember.

JD: Well, I can’t recall where I got him. I don’t recall where he came from. I do know that he is an extremely capable – or was an extremely capable engineer and I assigned him the job of trying to develop a different kind of counting ring. I, first, assigned him the one that I assigned to Mr. Mumma,

HT: Mhm.

JD: and he was not enamored of it and decided to invent his own. so I let him have full freedom to do so. So he — he developed what he called the conjugate pair type arrangement which had only 10 vacuum tubes in it. And they operated in pairs across the ring. And they — they just went around the ring in pairs, because they acted as trigger pairs across the ring,

HT: Mhm

JD: and we never were able to get the speed up to the point where – we used it once and that was in th~ ~ that was in the work we did for the Army in the land mine controller. I think the receiver in those — in those, .. .. units out in the field


[Louis deRosa’s list of patents includes:

US Pat 2315249: Pseudo-extension of frequency bands, 1941 Dayton

US Pat 2671896: Random Impulse System, 1942

US Pat 2774965: Radio Detection System, 1943

US Pat 2406813: Intelligence Transmission System

and at least 10 others.

[Search at Google Patents.]

[Page 118]

used conjugate pair counters. I’m quite certain of it.

HT: For the decoding of the signal

JD: For the decoding of the signal, yes.

HT: to determine whether or not it was going to be triggered or –

JD: Yes. Well, it was a question there of power

HT: Mhm.

JD: and so forth, and he used vacuum tubes in the cojugate pair. I’m quite sure that’s the way he did it and I – an examination of the patent, I think, would show it.

HT: Mhm.

JD: Because he did it.

HT: .. The reason – one of the reasons I mentioned this book was because on page 2 it looks like he had a diagram which he calls “the mechanoelectric converter” which looks very much like the blueprints you showed me in that other .. the’ 45 document

JD: Hmm.

RM: Of course, these are high — these are all high vacuum tubes.

HT: These are all vacuum tubes? These are not the thyratron .. counters?

[Page 119]

RM: I believe. I may be wrong.

JD: These must be the key switches.

HT: Mhm.

JD: And then they go through an amplifier of some kind to give a sensible output here on this line right here.

RM: Yes. This is an impulse generator and this is what he –

JD: Yes.

RM: held close those switches to get the operation he wanted.

RM: Yes, you see, that was done before I even came here yet. He’d already gone that far.

HT: That’s labeled December 14, ’38.

JD: Yes, that was before Mumma.

RM: Before those thyratrons, yeah.

HT: Well, you wanted to mention something about his subsequent career.

JD: Well, he – he left here to go to IT&T

HT: Mhm.

JD: and he stayed there — he, he was a – an antenna expert in his own — Well, I first of all not only used him to try to develop new and different ways to build counters, but there happened to be a

[Page 120]

program come up that the company needed me badly to put on a show in the auditorium for the Hundred Point Club, which is a yearly event in the Sales Division

HT: Mhm.

JD: for those people who make their objectives in sales And so I worked for three, four or five months on that one and what needed a lot of attention was the sound system in that auditorium was practically obsolete. I mean you couldn’t hear anything. So we built a whole new sound system, which is in today,

HT: Mhm.

JD: and it was flat from 20 to 20,000 cycles. and DeRosa built it and put in the woofers and tweeters and all the — all the power amplifiers and everything. And we developed a very powerful system including i dual turntables, we could run it for a program that had music associated with it. You were here at that time.

RM: Yeah.

JD: And, we had stacks of records, all marked with tape so that we could change the music back and forth. It’s just like the disc jockey of today

[Page 121]

HT: Right, right.

JD: only we did that back in ’39, something like that, ’40 I guess. And DeRosa did that whole thing because he was a – an expert in sound.

RM: He wasn’t only.. just let — See if I can trigger something else right. Remember how he was working while he was here on this pseudo-bass device that

JD: Yes.

RM: he could generate bass with the audio system even though there was no bass actually there, by the proper choice of harmonics he would give you the feeling that there was bass there.

HT: Aha.

RM: And then it got so much talked about. Remember when Colonel Deeds invited us all out to his home

JD: Yes.

RM: and brought DeRosa out there in order to probe to see what was in this new audio amplifier?

HT: He was kind of using the overtones to – to give you a beat effect of the bass.

RM: .. The higher frequency harmonics were there of the bass note but the bass tone itself wasn’t there, you see.

HT: [Laugh]

[Page 122]

RM: So, he could – didn’t have to have the wider frequency response and still make it sound like the bass was there.

HT: [Chuckle]

JD: He certainly was an expert in – in – in audio.

HT: Aha.

JD: And also, as it turned out, he was an expert in antenna systems, because that’s what they used him for in IT&T.

HT: Aha.

JD: And, I, during the war, I met Louie very frequently, because I traveled a lot during the war, and I’d meet him and he was working for IT&T and he was working on field radio systems with the parabolic antennas and all that and mobile devices

HT: Mhm.

JD: for field use. And I saw Louie written up in different places and his picture and so forth, because he was a very, extremely, capable man. And I don’t think that really I had a big enough challenge here for him to satisfy him

HT: Mhm.

JD: and that’s the reason he wanted — he left 1n very good spirits and always – always was very amiable,

[Page 123]

HT: Mhm.

JD: but he needed a terrific technical challenge. And if what I assigned him to didn’t have the proper level of technical requirements, .. or, what I might say, research unknown, Louie wasn’t happy.

HT: Mhm.

JD: So that was the story of Louie who accomplished nothing while he was here of any permanent nature. He hit a little here, he hit a little there and, but a very high caliber man – very high caliber man. .

RM: .. Y’ see I was more application oriented and he was more theoretically oriented and that was the big difference

JD: Mhm. That’s right.

RM: between the two of us.

HT: Mhm. You say he eventually became the scientific advisor to the President; and when was that?

JD: Nixon, you bet. Now, why he left, I don’t know. But I noticed that when Nixon entered the White House for the first time, that is, the first four year term, .. Louie DeRosa’s name appeared in the press as a technical advisor to the White House, and, now I can’t prove that, it’s just-

[Page 124]

something that I, seem to tuck away, because he didn’t work for me anymore. But-

HT: Mhm.

JD: and I don’t – I don’t see him anymore, but it was one of those things that made me very happy. I was quite proud of him

HT: Mhm.

JD: to think that he could make the grade and get up there.

HT: Did you say he was no longer alive?

RM: Yes, he’s dead. He’s been dead a couple of years –

JD: That shocks me because I didn’t know that.

RM: Yes, I saw it in some type of a magazine, Joe.

JD: Is that so?

RM: He died more recently than Frank Bucher, and Frank Bucher died several years ago.

JD: Yeah, Frank Bucher worked for me, too.

RM: Frank was at Bell Laboratory.

JD: Yes, he went to Bell Labs.

RM: I thought that DeRosa carne from Bell Labs here.

JD: I don’t recall. I can’t recall where he came from I’d have to have those resumes –

RM: Yes, I do not remember, he was here when I came, I know that, because DeRosa and Killheffer were both here.

[Page 125]

JD: I had four or five men here when I got Bob. Bob came from Frigidaire, and – and picked up this concept that I had written up earlier, and with modifications and additions and so forth, he essentially followed that pattern. And, of course, this was the subject of a violent discussion in the court room

JD & HT: [Chuckle]

JD: when the lawyers tried to prove that – that he was the total inventor and I had nothing to do with it, see.

RM: But it was a joint patent.

JD: And so I don’t know how it ended. Did it end that there was joint interference?

RM: Yes, the first model was a joint one. On the one that had the calculator, that I was sole inventor on that one, but the first one that we went into interference with IBM, why we were joint – you were joint on that.

JD: Yeah, but there was another one. They – they switched that thing three times – or twice.

RM: Well, they split the case – I mean, they broke it up into sections, you see.

JD: Yeah. Well, anyway, I – I noticed that the – that the court .. made one statement in its decision

[Page 126]

and in their opinion they had serious doubts that I was even an inventor. [Laugh]

HT: [Laugh]

JD: And they were using my notebook all the time.

HT: [Laugh]

RM: Well, of course, I was more closely associated with it and I knew the details.

JD: Well, we communicated daily.

HT: Mhm.

JD: I mean I had not over 20 people, and I got to talk and sit down with all of them every day,

HT: Mhm.

JD: with no communication problem at all. I knew exactly what everybody was doing every day that I was here. And so the lawyers tried to make something out about this – my lack of communication and that was a lot of baloney because if it wasn’t in that book, I was telling them what I thought. And while I didn’t order anybody to do anythingI never did because it was a semi-research thing . and I knew you don’t order people to do research. .. Nevertheless, you guide them, even express an opinion and, in the end, they’ve got to have something that works.

[Page 127]

HT: Right.

JD: So the ultimate test

HT : Yes.

JD: is not whether they followed your orders but whether or not they made it work.

HT: Ahm. Well, back on the pumping station, .. .. maybe we ought to trace that through. Because you showed the diagrams, approximately 1948, ’49, as it existed if we were to walk over to Building 20. today, but you were building these small thyratron tubes .. almost immediately when you built that first –

RM: 19– 1940.

HT: Right, that first prototype,

JD: Sure.

HT: which means that you had to have some kind of an operation like that pumping station.

JD: Yes. There is a diagram there of the very experimental station.

HT: Mhm.

RM: Have you got your number — Was that in your number one, early notebooks, or not?

JD: Ah – ..

RM: Well, I don’t know.