Apollo Guidance Computer History Project
July 27, 2001
Dan Lickly's introduction
DAN LICKLY: I'm Dan Lickly. Before I go on the exact stuff on Apollo, I'd like to make one other pitch. A number of us worked on the Polaris Guidance Computer and the Polaris system beforehand, which was a very interesting computer, although again, Eldon probably can tell you more about it than we can. But it was, for its days in the 50's, very interesting. And the way it controlled the Polaris missile was a precursor to the way we did all of the control runs on the Apollo.
And furthermore, for a lot of us who have spent a long time in Florida watching launches, and data reduction and all of the things that happened, it gave us a lot of real feel and confidence in what we were doing in Apollo, including some things that never came out. For instance, we saw things on Polaris like squib firings that messed up all of the computers. And we realized on Apollo that we should make the computers as fail safe as possible because things like that just plain happen.
So when we started on Apollo, which I guess was that August 8th date, there was that meeting with Bob Chilton down in Washington. Is that the date, in '61? I was trying to put the times together. That's when we really got in high gear, but some of us were working on it in '61. I don't remember how many of us were down there in Washington, but quite a few of the people who are still here are. It gave us a great deal of confidence, to me, that we could do a lot of the things that Apollo called for.
Now the Apollo guidance computer was quite a different computer, and Eldon and Ray know far more about it than I. I spent the first few years working on the reentry guidance -- the development of the equations, the reentry guidance, reentry autopilot, although the main person -- There are only three, or four of us max. It wasn't a very large group; one for the autopilot, and Ray North, Bard Crawford and myself did the guidance and that kind of navigation.
Ray died in a ski accident about five years ago, and unfortunately, a lot of that is not as well documented as I would like. I'm trying to find old material on it. It never made it. Although Dick Battin invited us to write it in a book, we never could make it elegant enough for him. It was pretty ad hoc in many ways, but it was a very tricky job, very tricky. A lot of people don't realize that even today, coming in from orbit, you can do anything, piece of cake. Whereas coming in from lunar or parabolic speed is basically a very unstable, tricky problem, and you have to be extremely careful on what you do because it's a very, very narrow window, and without much lift, very limited maneuverability that you have.
So we spent quite a while there, and actually, I ran into Alan Klumpp our lunar landing person. He gave us a tour, my son and I, around JPL. He's using Ray's equations now, and development for the descent and part of the Mars that he's working on. I don't have the details, but he was after me to get some more data, which I'm still trying to get some old reports that explain more. He said I've implemented Ray's method, but I don't understand all of it.
And one of the tasks I'd like to do some time is try to get it together and write up something definitive that explains the whys and wherefores and so on, while there are still some people alive and well that remember what's going on. And when I've asked in NASA -- We met with Aaron Cohen month after month during all of those long developments in Texas, but they tell me Rodney Wingrove, who did understand it, is still working out in Ames. And he'd be a great resource for one of my -- If I ever can get motivated in time to be digging into this.
Now as far as programming AGC, I know we did a lot of the programming. We didn't develop it; Ray, Al did and so on, but we spent many years, you'll hear more from the others here, on the actual programming. We have various programming groups, and many, many people.
It's funny. I flip open this and it was handed to me and I start to read AGC assembly code which I haven't seen in a long time, and I find it still reads pretty easily. I didn't expect it. But there I see, and I actually remember what's going on.
I was also in charge of the first Saturn V's for the Draper Lab instrumentation lab, their unmanned SA501 and 502. And we had a nice, small team and did them, and I got to go down to Florida and watch that incredible lift off, which was the most thunderous thing you could ever imagine. And then we flew over to Houston afterwards. It was an amazing sight. And one of them, I think the 501, it actually didn't burn for the -- Which stage, didn't ignite? (See Hugh Blair-Smith's annotation) It was supposed to put us in a translunar injection, and the S4B didn't really do it. But fortunately we had done the program general enough, instead of doing a retro burn to stop us from the translunar and bringing us back, we did both. We first fired out, got it on the way out, and then coming back. I'd have to go back and look at the details, but we were very glad we had done it in an extremely general way so that we could pick up the pieces, and we landed pretty much where we expected.
After the programming, I left the instrumentation lab. I'm not sure when it became Draper. I keep calling it Instrumentation Lab. Fred can fill you in more later. We were in another small company here in Cambridge.
In the last ten years I've been teaching C++ programming. (Laughter) A lot of the people don't like C++. I did teach Fortran and C and some others. I'm not a big fan of it. In the world, a lot of the people are switching to JAVA. In fact, I'll tell you the high school, standard for all high schools in this country as put out by the people in Princeton, switched to C++ about three years ago as a standard. If you take an advanced programming class in high school, you get an AP in computer programming and C++ now. And in 2003 they're going to switch to JAVA. They've already made the announcements. So high school kids who take programming, my son just did it last year, will be in JAVA.
I don't like C++ particularly, but I'm up there teaching C++. Now I'm doing a little more advanced class, which is better than the freshmen. Don't teach freshmen anymore. That's one of my rules. Teaching at University of New Hampshire. All the engineering students were required to take a programming course. They used to take it in the sophomore year, and then they moved it up to the freshmen. And freshmen in the Fall were not ready to take a very difficult programming course. They had too many other adjustments, too many other demands on their time, and it was just a horrible record as far as who could keep up. They'd get behind, and totally lost. Half of them passed the course; it was about par. I thought that was bad, and then I found out the CS majors, a third of them flunked their first course in computer programming, and less than 20% make four years. I don't think it's that bad at MIT.
But there are a lot of people that just are overwhelmed by programming. A lot find it easy, but there is a big body -- That's the trouble of teaching. A third think it's too easy and a third think it's impossible. Very few are in between. My grading would be a third A's and a third fails and a few in between. There are a few courses like that in college that are binary -- You either get the idea or you don't. That's what happens teaching programming.
Dan and others were uncertain about which mission it was where the CMs GN&C had to make up for a defective booster, SA501 or SA502. I know it was SA502, in April 1968, because Steve Copps and I attended that launch and the splashdown party. And thereby hangs a tale:
Steve and I were sent to the Cape to take down numbers from control station screens in the backup control room in the hours before launch. OK, the real reason we were sent was to give us a treat after much laboring in the vineyards of Apollo; it was clearly understood that there was essentially zero chance that anybody would look at our logged numbers, ever. We were told to get there a few days before the mission, check in with the Instrumentation Lab office, and then stay out of the way of the people doing the work until T minus 8 hours. So we got in a couple of days of stealing grapefruits from trees and fishing in the upper St. John river, where it seems to be maybe 5 inches deep with cattle standing around in it. Somewhat more to the point, we explored the VAB from top to bottom, getting vertigo on catwalks and discovering that if you walked down the stairs you could get into all kinds of places that werent supposed to be available, even to us. We stuck our heads inside the interstage space of SA504, for example, noticing the huge number of red tags labeling gear that wasnt flight-ready. At night, the "stack" of booster and spacecraft were brilliantly floodlit, giving them an intense ethereal beauty.
The first brush with history on this trip was turning on the TV news in the motel, and finding President Johnson ("Father Cornflower," Steve called him) announcing that he would try to heal the deeply divided country by not running for re-election. Another brush came later...
We did our all-night stint in the backup control room, and were allowed to go out to the fire escape stairs for the launch itself, at about 7AM. So there was not so much as a pane of glass between us and the most explosively powerful machine we could ever imagine. The deepest bass notes of that noise were well below normal audio range; rather, you felt the compression waves on your chest. Without a clue as to the problems with the second stage, we staggered off to get some sleep.
What happened while we slept was this, as near as I can remember the explanations: one of the five engines in the second stage failed to ignite, and then the control logic that was supposed to adjust the behavior of the other four also malfunctioned, in such a way as to shut down one of the good engines! So the second stage did what it could on three-fifths power, leaving it up to the third stage to achieve orbit, but it naturally didnt have enough poop to perform TLI (trans-lunar injection). The GN&C system was commanded to observe and correct the VG (velocity to be gained) earlier than it would have normally, and so it did, placing the unmanned spacecraft approximately where it should be, with the appropriate velocity for the simulated return from the moon. A black eye for IBMs LVDC, and a triumph for our system.
So the splashdown party should have been a super celebration, with MIT folks as the heroes du jour, right? No, history stepped in again: all anybody could talk about was how Martin Luther King had been assassinated that afternoon, and what catastrophes would result.
site last updated 12-08-2002 by Alexander Brown