PDA

View Full Version : Very interesting storage device



Homunculus
November 2nd, 2012, 12:38 AM
This is a part of my Kienzle 6000 minicomputer.
The computer has two of these modules, connected to the part "MPS Steuerung" (MPS Control)
These are called "Speicherplatte", which means storage plate, but the computer has normal core memory.

10882

It doesn't look too interesting on the outside:

10879

But in the inside there is only some U shaped ferrites:
10881

10880

Each ferrite has a spring on the other side, and the removable top holds an other piece of ferrite over every ferrite, so when assembled, they look like an O.

10883

Core memory store one bit in every ferrite core, but if this uses the same principle, then the capacity is way too low for even 1970 standards.
Anyone has seen something like this?

xprt
November 2nd, 2012, 09:52 AM
Maybe not a storage device, but some sort of switching matrix?

Chuck(G)
November 2nd, 2012, 10:12 AM
A read-only memory, perhaps? Could it be that a large punched card could be placed between the top and bottom parts?

Homunculus
November 2nd, 2012, 10:25 AM
I dont think so, its actually very hard to get out of the machine. You have to remove the side of the computer, and there is two ribbon cables in the front of these things. It has 104 of these U shaped ferrite cores, too low capacity for a practical ROM.

Chuck(G)
November 2nd, 2012, 10:45 AM
I dont think so, its actually very hard to get out of the machine. You have to remove the side of the computer, and there is two ribbon cables in the front of these things. It has 104 of these U shaped ferrite cores, too low capacity for a practical ROM.

They could be nothing more than transformers for driving the core memory--does this card connect to the core memory at all? Does the number of these (104) correspond to the number of X and Y lines on the core memory?

Homunculus
November 2nd, 2012, 10:57 AM
No, its far from the memory, of course the backplane connects everything but i think its has nothing to do with the core.
I just guessed that its some kind of storage, because of the "Storage Plate side A" print on the PCB.
The core memory is a big fat card, it has the driving circuit on the other side.

10886

Chuck(G)
November 2nd, 2012, 11:14 AM
Is the wiring (the rat's nest) the same on each core? If not, it could well be a hard-wired ROM. That could explain the arrangement--for changing of memory contents.

MikeS
November 2nd, 2012, 11:29 AM
Core memory store one bit in every ferrite core, but if this uses the same principle, then the capacity is way too low for even 1970 standards.
Anyone has seen something like this?Well, the name would suggest that it's a solid state disk; there were multi-level ferrite devices capable of storing octal and decimal values, maybe that's what they are?

How does it connect to the backplane, and how many connections? The handle suggests that it just pulls out, but the green parts look like 'sideways' connectors.

Homunculus
November 2nd, 2012, 11:36 AM
You're right, i just googled "core ROM", looks like its a core rope memory (http://www.computerculture.org/2012/10/dimond-rings-and-read-only-ropes/).
Hard to imagine programming it, must be a real pain.

Homunculus
November 2nd, 2012, 11:43 AM
The connection is only a flat PCB end, the green parts aren't normally exposed, i just removed the top.

MikeS
November 2nd, 2012, 12:01 PM
You're right, i just googled "core ROM", looks like its a core rope memory (http://www.computerculture.org/2012/10/dimond-rings-and-read-only-ropes/).Chuck had it as usual; that's what it looks like all right, your computer's 'BIOS' ;-)
I hadn't seen that type before, just the traditional 'ropes' where they got their name:
http://www.computerculture.org/wp-content/uploads/2012/10/rope_ferrit_core_memory.jpeg

Hard to imagine programming it, must be a real pain.I think that was the advantage of the two-piece cores like yours, that they could be 'programmed' by machine instead of by hand like the 'ropes'. Any idea what the green parts do? Are they just transitions to the edge connectors or are they maybe involved in the actual program?

Interesting machine for sure!

Homunculus
November 2nd, 2012, 12:53 PM
Any idea what the green parts do? Are they just transitions to the edge connectors or are they maybe involved in the actual program?

They are just connections with the other side and the edge.

If its really a ROM, then I have a chance to reanimate it, because it means I have some kind of software.

commodorejohn
November 2nd, 2012, 01:09 PM
They are just connections with the other side and the edge.

If its really a ROM, then I have a chance to reanimate it, because it means I have some kind of software.
If I'm reading about this right, then each of those pairs of posts is one bit, so that's 8x13 or 13x8 per module...probably only a simple loader program, then...?

Homunculus
November 2nd, 2012, 01:56 PM
I don't know, there are multiple windings on each core, normal core memory has only 3 wires, so maybe it has higher capacity.

xprt
November 2nd, 2012, 02:29 PM
I don't know, there are multiple windings on each core, normal core memory has only 3 wires, so maybe it has higher capacity.

From the article you linked to:

The basic idea was to create a transformer with a multi-turn sense winding for each bit, and a wire for each word in the memory

It looks like each core is one bit, so you have 13 bytes total.