Apollo Guidance Computer History Project
November 30, 2001
Eldon Hall's Introduction
ELDON HALL: I started from Harvard too, but, when I was in the Harvard physics department, the word digital computer never came up. I didn't know what a decimal and binary system was. I didn't know what hexadecimal was, as none of that stuff was ever mentioned at Harvard, in the physics department. So when I got to the Draper Lab, in 1952, I was working for Hal Laning, who was over his head programming pipes and things with analog computers. So I began to be interested in it fairly early. And by the middle fifties, when the lab got involved with Polaris, I had run some experiments about how to make certain type of logic models, so I got grabbed from a pile of people and headed up the Polaris computer development.
And since that went fairly successful I inherited the Apollo responsibilities when that came. When NASA awarded the Instrumentation Lab a contract for the Apollo guidance system, they planned to follow in the footsteps of the Polaris program, which they did, and they planned to set up a group of support contractors to manufacture the system, and we would have the design responsibility.
And, as I mentioned, I inherited the responsibility for managing the guidance computer, since I'd spent several years on the Polaris program and essentially knew the type of job that we had on Apollo. So to follow this plan after we received the whole project, we put together a documentation to support a bidders conference in early 1962. Alonso and Hanley, here, and I were the nucleus of the evaluation team for the technical proposal.
Fortunately, for me at least, with Jack Poundstone's understanding and experience with MIT on the Polaris program, Raytheon put together a winning proposal. We had developed by that point a very good working relationship with them, and, with Jack's influence, maintained that throughout the Apollo program, even though he "evaporated". I think his heritage stayed with us, fortunately.
ED BLONDIN: He makes a pretty big ghost.
ELDON HALL: Right. So you felt his influence, too.
ELDON HALL: Also, somewhat related to the overall development plan, there were turf battles--two of them, actually, that had surfaced within the lab. One dealt with the responsibility for test equipment, and the second with the responsibility for the guidance system displays and controls. In both cases logic prevailed, and the computer portions of those were split off and became the responsibility of the computer design team.
I turned the computer test equipment design over to Raytheon knowing that they would follow the earlier work of the computer design team. Then the computer part of the systems' displays and controls would be more complicated--astronauts and human factors types within the lab and the space contractors would divide requirements, such as color of light, the size of the light, the shape of the numerics, and the placement and action of the keys, and all that kind of foolishness that human factors has to work out.
Again, Raytheon took a very responsible role in the mechanical design of what eventually became known as the DSKY. This arrangement set the baseline of the design and production responsibility, and I was fairly comfortable with the arrangement. And a lot of this was occurring in 1962, which was a very busy year. In addition to the bidders conference, I got Dave Hanley going on integrated circuits, because I felt the core transistor type of logic that Ray Alonso and Al Hopkins were pursuing, the things that everybody else has mentioned, was not the way to go. So I challenged Dave to try to get the ICs--this is early '62, and I placed a bet with him.
And then, in addition in 1962, many other things were involved Duggan had to work out the size and shape of the things to go in the spacecraft, there wasn't much room to fit all this "garbage". And so all of that kept things pretty busy in '62. And then, near the end, I had the responsibility of facing Charles Frick and trying to convince him to change this contract and use integrated circuits, because NASA had just finished negotiating the contract with Raytheon and in fact getting the full go ahead, to go into integrated circuits, upset them too. So I had to face Charles Frick, and apparently was able to convince him that this was the way to go, with integrated circuits, and we got the approval, in December of '62, to go with integrated circuits.
And even though I was comfortable, various people within NASA had some concerns, and they brought Bellcomm on board, to search out potential problems. Bellcomm raised several issues, but most important was the questionable reliability of ICs and the capability of the MIT design team to complete the design on schedule. This led to computer design studies for a back-up development. The back-up design faded away when it became apparent that the MIT design was progressing. So I won't go into the details of that backup, but it got kind of messy at times, from my point of view.
Then, another NASA concern was the contractual arrangement with Raytheon, as a subcontractor to MIT, and Jack mentioned this. The introduction of the ICs into the computer design and adding test equipment to the Raytheon contract, and a multiplicity of other design changes, was resulting in contractual problems. And as Jack mentioned, he was a little too soft in the way he handled this, from the production end of the contract.
So NASA was questioning MIT's capabilities to control Raytheon's contract, so they changed the arrangement and put Raytheon as a sub to A.C. Sparkplug, who would be the prime, reporting to NASA. This change in contractual responsibilities worked okay even though Draper Lab was extremely nervous for a long time, about that change. We felt we should have a more responsible role than production.
Then, in 1963, there was major progress in various areas. One was getting the AGC-4 up and running--that was the first real Apollo computer in integrated circuits. Dave had built one previously, and that was sort of transferred into the AGC-4. Also, Raytheon built their first version, which was called AGC-4B, which was essentially a copy with integrated circuits, and then, by the end of '63 they were cranking out modules for AGC-5, which was the things that you talked about with the TL 47 cans and so forth, and would be the first flight computer which was up and running in '64.
And then that was followed with a set of what was called implementation meetings, which Cline was very familiar with, which put the auto-pilot in the computer and lots of other responsibilities for the computer, which sort of forced us to go to the Block 2 design.
JACK POUNDSTONE: Thank goodness.
ELDON HALL: And a serious expansion in logic required that we went to the flat packs and multi-layered boards and so forth. I don't want to say much about my future after Apollo because it sort of went downhill and I wasn't too happy with it.
__: I put the water in the mould?
DAVE BATES: Remember, we were looking for water in the mould
ELDON HALL: You were, but I wasn't.
DAVE BATES: You were part of the evaluation team.
JACK POUNDSTONE: But if you had gotten fired like me, you'd been a lot better off.
ELDON HALL: I would have been much better off, that's right. But I had fun. I wouldn't want all those manager responsibilities.
DAVE HANLEY: Were still working together.
ELDON HALL: We're still working together, but you were giving me trouble too. That's my history. I retired in 1988.
Home: Conference 3