I was on a late evening flight from NYC to Binghamton, NY. Across the aisle from me were two gentlemen who had a travel case with a NASA label. I got to chatting with them and found out they were engineers from IBM's Federal Systems Division. The case contained a computer from one of NASA's sub-orbital flights. The computer had failed during the flight and the engineers were transporting it back to IBM Owego where it had been built.
I found out later that the astronaut was using the computer and when he was finished he powered it off without invoking the shutdown routine. There was a program still running on the computer and it was in the middle of reading from core memory when power was removed. The result was the program was trashed and the computer was useless. When this was determined it was communicated quietly to NASA and nothing more was said. IBM and NASA didn't want the public to think an astronaut could make a mistake.
Jim Campbell
IBM 1960-1994
toggle quoted message
Show quoted text
On 6/18/2022 10:08 AM, James Daldry W4JED wrote:
Hi, Tony
In my "toy room" I have an assembly of several 4 inch square circuit boards, bolted between a pair of 5 inch by 6 inch boards with edge connectors along one of the shorter edges. The assembly is a core memory stack with 4K, 36 bit words. Each "core" is a teeny tiny toroid with 3 wires passing through it - x and y position, and sense. If you sent a pulse through both x and y wires the sense wire would send out a pulse if the core was magnetized in 1 direction, no pulse for the other. The direction of the magnetism determined whether the core was storing a 1 or a 0. Since reading a 1 was destructive, you had to re-write every 1 after you read it. That made it very slow, but you could power the thing off and come back in 10 years and find the information still there.
The Apollo computers used what was core rope memory. It is explained in depth on the (uh, oh, should I say the word?) Wikipedia page on core memory.
73
Jim W4JED
