5 ms·
I found this story interesting, but unsurprising knowing the author. I'm amazed he got that thing to work - but if there were anyone who could do it, it's him.
by cr0sh 7y ago
I found this story interesting, but unsurprising knowing the author. I'm amazed he got that thing to work - but if there were anyone who could do it, it's him.
Unfortunately, I didn't find an answer to something I suspect, but I don't know for certain? Maybe someone here could shed light on it?
He mentions that the AGC was the first computer to use integrated circuits, in the form of 5600 NOR gates, built into custom "logic modules", each set then wired together using a backplane or something like that. But take a look at those ICs - are they thru-hole, or SMT?
I seem to recall seeing pictures of the AGC - of other parts of it - and that it used a form of SMT in its construction; does anyone know if this is true? If so, would that also make it the first time SMT was used in a computer? Or were there other prior examples.
It seems also curious that if it was SMT being used - that such a thing was chosen for this critical piece of hardware, mounted on a rocket undergoing tremendous strain and vibration. I would think in that case DIP would be the better solution, ideally with extended pins for wire-wrap, then everything embedded in conformal coating after test - but while the DIP vs SMT question is up in the air, I've never seen any pictures that suggest wire wrapping was used, and nothing that showed anything like conformal coating.
So - can anyone shed any light on this?
- kens 7y agoTo answer some of your questions. The AGC was a very early integrated circuit computer, but probably not the first. The ICs were surface-mount flat packs onto boards in the modules. They were welded, not soldered. The modules plugged into the backplane connectors, which were wire-wrapped. The modules and backplane were encapsulated in plastic (epoxy or polyurethane) for flight. (This AGC was for ground testing, so it was not encapsulated.) The non-logic circuitry was built with cordwood construction, i.e. the components were inserted into holes in the module. Point-to-point wires were welded onto the components on either side. These were also encapsulated in plastic. I don't know if DIP integrated circuits would have been more reliable, but the DIP wasn't invented until 1964, too late for the AGC. The Saturn V rocket contained a Launch Vehicle Digital Computer (LVDC), which was completely different and built by IBM. It also used surface-mount modules, but IBM's hybrid modules, not integrated circuits. So the AGC wasn't the only surface mount computer at the time. (Does anyone have a good history of surface mount?) > if there were anyone who could do it, it's him Honestly, Mike Stewart is the AGC super-expert who made the restoration possible. He knows way, way more about the AGC than I do.
- nas 7y agoKen, I love watching you and Mike do your stuff (Marc too but usually he is behind the camera). Thank you for taking the time to explain things.