An Apollo Guidance Computer Laid Bare

An Apollo Guidance Computer probably isn’t a machine that’s likely to come the way of most Hackaday readers. The device that played such a vital role in taking astronauts to the Moon and bringing them home again is hardly a common find, even if it is one of the most iconic machines of its type and era.

[Carl Claunch] was approached to assist in the restoration of an AGC, and while he can’t reveal any information about its owner he is at liberty to document his progress. The result is a fascinating in-depth technical examination of the device over multiple blog posts, and is well worth a read for anyone with an interest in the Apollo program. It’s an ongoing progression of blog posts that are probably too numerous to list individually, but include the construction of a substitute for the DSKY control panel as well as looking at the device’s memory and construction. [Carl] then embarks on a series of posts looking at the restoration itself. This is where we see the computer in greatest detail, and learn the most about it.

If you think you might have seen [Carl]’s name here before, you’d be right. One of his past exploits was getting the first version of FORTH running on an IBM mainframe.

40 thoughts on “An Apollo Guidance Computer Laid Bare

    1. “It would be very hard to achieve a full gate by gate replica, in part because the electroluminescent panel on the right side of the DSKY is unobtainable. That obviates the need for 250V 800 Hz power which is what the relays switched to the various EL panel segments, making the internal construction different from the real DSKY.”

      they need to know about fran!

      1. Not really. I actually side with the owner. The Smithsonian already has one, as does The Computer History Museum, and a bunch-o-other places. It’s not like NASA’s ever going to use it again, nor is it a on-of-a-kind artifact. But they still claim ownership. This guy is paying to restore it to operating condition, something the Smithsonian and CHM have not done. Good for him.

        I think it’s more of a fear that if they let one person get away with owning a souvenir, stuff will start disappearing awfully fast. Like the military did with radio gear. Rather than sell all their old R-390s to hams at auction, and make a tidy profit, they crushed them, and all other surplus radios, starting 10 or 15 years ago.

      2. “NASA considers all their stuff to be theirs. Weird, innit?”

        It is when it’s not their stuff.
        I have owned lots of things in the past that are no longer mine. Some of those things I regret ever getting rid of. That does not give me any right to take them back from their current owners.

        If they sold it, it’s not theirs.
        If they gave it away it’s not theirs.
        If they told someone they could have it and they accepted then it is not theirs.
        If they instructed astronauts to abandon it but also told them that any “to be abandoned” goodies that fit in the astronaut’s limited “carry on” could be theirs if they want it… then it does not belong to NASA.
        If the next person to own it after NASA chose to give it away or sell it to someone else.. unless that someone else was NASA it STILL does not revert to NASA.
        If they abandoned it then the rules of ownership should be equal to when anyone else abandons something.

        I don’t understand why some people seem to have a problem understanding this. Part of ownership is having the right to transfer ownership away from yourself. This transfer is a real thing. Once you do it that thing no longer belongs to you, it is someone else’s property, seller’s/gifter’s/abandoner’s remorse be damned!

        1. If you happen to be hiking across the southwestern USA, coincidentally along the debris trail from Columbia, and pick up something un-natural. Keep it to yourself. Though if you do blab it may take NASA years, if ever, to get around to you. Some police department along there that aided in the search has had a couple of pieces they’ve tried for years to get NASA to come and pick up.

    1. The owner may have not gotten in through one would say legal means and he knows it. It’s not like the AGC is something you’d find at a government surplus auction. Chances are it’s a ex-NASA or contractor employee who had access to certain storage areas and availed himself to it and walked right out with it.

      1. Actually, it’s not unreasonable at all that the owner got it though legal means. There were probably more copies running in contractor integration labs all over America than ever saw the inside of a NASA facility.

        Once the project shut down all that “customer provided equipment” was likely abandoned in-place and shoved into storage while someone else figured out what to do with it. A few years after it flew it was all obsolete and nobody needed it and certainly nobody wanted to store it. Eventually it was auctioned off in a lot along with 200 pounds of surplus cables and 10 VT100 terminals..

        I was down at the Rockwell Downey facility about 20 years ago after it had been repurposed into a film studio. Being an engineer, and aware of the history of the place, I had a look around one day and walked into an old lab that had been turned into a storage room and was jammed with racks clearly marked as Apollo service module integration and test equipment.

        They meant nothing to the film crews working there and even less to the management, who wanted to junk the lot of it because it was filling up a room that could be used for storing costumes or something.

        1. Yup. That’s almost certainly the case. (I like the 200 pounds of cables and 10 VT-100s…all on a pallet in a dusty corner of some warehouse in Boston? Long Island?)

          Though now that it’s public, of course NASA will raise a stink about it. Well, maybe they’ll just let it slide, but they don’t have a history of doing that. More likely, they’ll wait until he gets it all working and then make their claim.

        1. If this is the guy that brought the AGC to SpaceFest (and by the story you tell, I’m pretty sure it is), he’s also got the only existent MOL guidance computer.

          And I’ve touched both, and have been giddy ever since.

          1. Yeah, same AGC! Spacefest was the start of this restoration effort — I spent the whole time there mapping out all of the unknown backplane connections and taking high resolution reference pictures. You can see the resulting tool, which we used extensively last week, here: http://apolloguidance.computer/2003100_071/pins It contains all of the backplane wiring info for our specific AGC, and is accurate to the best of my knowledge (minus some pins on the erasable sense amplifier module that are no connect instead of grounded that I need to push the database updates for).

        2. Talk about a lucky find!

          The nerd community is the probably only way we will ever see an operational AGC on this earth. I doubt any museum would be able to fund the restoration to operating condition. And a real live operational AGC (even if it does have a modern DSKY) is many, many times more awesome than an inert metal box behind glass.

    2. The owner bought the AGC from a scrap dealer decades ago. He also dug up a copy of the official document by Eldon Hall (head of the AGC project) scrapping this unit. So everything is legit.

    3. I definitely remember there was a “ground-only” AGC in a space memorabilia auction only a few years ago. It might have been Bonham’s or Sotheby’s. Come to think of it, it might have been a shuttle-era computer. I remember seeing the “ground-only” or “not for spaceflight” labels on it.

  1. Fantastic project, Carl, I have followed your 1430 saga and your FPGA interface for getting the Diablos running during your Alto restoration and this is really a wondrous goal, the best of luck and thanks for bring us along for the ride!

    Would it be possible to settle the exact dimensions of the chassis? Given the original requirement “fits in a cubic foot” we have measurements mentioned of 12″ wide x 24″ long x 6″ high. Yet the ‘utlimate AGC talk’ had the measurements as closer to 12-1/2″ wide. Which is the more accurate, and are there any 3-view drawings of the chassis available?

    1. Mike Stewart is the expert on all things AGC. Ken, Marc and myself just learn at his feet, as we help him with the restoration.

      Thanks to his many years of effort, many documents were uncovered and published to extend what we know about this computer and the related programs.

    1. Thank you Mike, that is totally and completely fabulous! I really appreciate it, and the efforts you, Carl, Ken and others are doing. I’m reading Don Eyles’ SUNBURST AND LUMINARY right now, and to see the actual hardware in such detail actually being power up is just mindblowing. I’m sure I speak for all of us reading here: keep up the good work!

      1. My pleasure! As it happens, this AGC Handbook was given to me for scanning by Don, who received it from George Silver’s widow after George passed away. Don is the only reason we have any schematics for the Tray B modules at all, and without him this project wouldn’t be possible. He’s also the source of over half of the program listings we have, including Aurora, the LM system test program we’ll be using to have the computer perform a self-test. And to anybody else reading — I highly recommend his book. It’s great!

  2. My wife built the LEGO ‘Women of NASA’ kit, so I just had to do a little AGC and DSKY to suit. I tried to make them the same scale, not so easy but I’m happy with it. Thanks again for the actual chassis and DSKY physical specs!

  3. Its amazing to think that someone not only has an AGC but wants to restore it to some kind of operable condition.
    To think that in the mid 60s someone was building a computer that was small enough to go all the way to the moon and back.

    Given that the source code for the AGC is out there, I wonder how hard it would be to build an emulator/simulator for it. (hey, if MAME/MESS can emulate such things as an ANZ Bank Terminal then surely the Apollo Guidance Computer would be something that would fit)

    1. We at the VirtualAGC project maintain an emulator, yaAGC, which has been continually updated as I study the hardware design. I’ve been keeping it as accurate as possible, and it now even simulates hardware bugs caused by transients on certain wires (amusingly the original self-test code will fail if the emulator has correct timer/alarm phasings and doesn’t emulate the bug). The code for yaAGC, as well as the transcribed listings for all of the AGC programs we’ve come across, are in our GitHub repository here: https://github.com/virtualagc/virtualagc

      As for simulator — it’s not really possible in pure software even on modern computers, easily, because the AGC is an asynchronous computer that uses the propagation delays of its NOR gates to time things. You can get the logic to work in an FPGA, though, if you run a clock into every NOR gate to propagate them in locked step, simulating a propagation delay. I did this for the DE0 Nano a while back: https://github.com/virtualagc/agc_simulation

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.