Mystery core memory (core plane only)

[Terry Kennedy]

Does anyone recognize this?

I have no real reason to believe this is from a DEC system (and, with a 20 50-bit words or 50 20-bit words, it probably isn't). The RICM has an identical board (here), but they don't identify what system it is from.

Gratuitous close-up glamour shot:

 

[bombnav]

Is there any part numbers or a CAGE code on it anywhere?

 

[Christoffer]

That may be a tall order without any ID numbers. They were often stacked in little frames, and so the entire module with many boards would probably be more telling than a single board.
It looks sorta like a board from the module this guy shows: https://www.youtube.com/watch?v=3rfCDM5olfg but i'll be honest, I havent counted!

Id say with the green fiberglass pcb, it looks more 70'ies-ish than 60'ies-ish. A lot of these boards were also used in lesser known systems like desktop calculators and such.

 

[Chuck(G)]

...or jukeboxes (Seeburg used them in some models), printer buffers....

 

[Terry Kennedy]

That may be a tall order without any ID numbers. They were often stacked in little frames, and so the entire module with many boards would probably be more telling than a single board.

Indeed. But I was hoping someone here (including the RICM, where a matching board exists) might know more.

Id say with the green fiberglass pcb, it looks more 70'ies-ish than 60'ies-ish. A lot of these boards were also used in lesser known systems like desktop calculators and such.

The fact that it appears to have been parted out of a larger 3-dimensional stack argues against a limited-memory system like a calculator or jukebox. The green woven PCB does look somewhat more modern than the era I'd expect this board to be from, based on the size of the cores. If it was IBM, it would have probably have been on the muddy yellow phenolic board they used back then. A bit more research shows this image, described as "Magnetic core memory from 1956 for SEL's ER 56 electronic reservation system" with the same 20 x 50 organization. That seems to be "Standard Elektrik Lorenz", not the Florida SEL that was acquired by Gould.

See page 9 of this PDF for pictures of both a single board and a complete stack.

Also, the core memory circuits are complete with transistors equipped. They differ for the two memory sizes are only slightly. The core planes (matrices) are 20 X 50 = 1000 cores (Figure 10) depending on the capacity, or with 10 X 20 = 200 cores. 35 such levels are merged into a memory block (Figure 11) inserted between the plug - in units into the frame and is connected by plug connections to this central link quickly. A cabinet takes four such partial memories with their entire electronic equipment.

That does make me wonder how these planes made it to SPC and the RICM. The articles I've read say that ITT was a majority shareholder in SEL at the time, so perhaps licensed copies were made here? That might explain the more modern board material and the lower-cost construction, compared to the metal frame of the original. Or maybe they continued with that architecture for years in subsequent models

 

[Dwight Elvey]

Nicolet was one of the last manufactures of core memory. They used 20 bit memories a lot for there various signal accumulators and computers. It doesn't look like a core from a Nicolet 1080 as these were more square than rectangular. Still, they made many different machines. Many were designed to take analog samples of repeated signals, such as NMR, and average them over many samples. 16 bits was not enough and 32 bits was too much. 20 bits was considered a good size for doing math on analog sampled data.
Dwight

 

[Terry Kennedy]

Nicolet was one of the last manufactures of core memory. They used 20 bit memories a lot for there various signal accumulators and computers. It doesn't look like a core from a Nicolet 1080 as these were more square than rectangular.

I probably wasn't clear in my reply. It looks like my board is a variant of the SEL ER-56 core, probably from a later date: