29-Feb-2003 The outer edge of each board has two tabs that stick out, and are flared out. It looks like they are meant to be some kind of handle or knob by which you could pull the board out of the card cage. Without that, it would be next to impossible to get a board out when the card cage is full, because the outside edges form a pretty flat surface; nothing to grip. The boards that I've got out so far have all been next to empty slots, so I could pry them out by reaching into the gap, behind the plastic end-piece.
So I'm starting to think about making a board-pulling tool. It would have two sets of closely-spaced prongs, which would slip over the narrow part of those tabs. It will have to be very strong, and may break those tabs if it is not aligned very accurately.
I'm also wondering if this plastic will be getting brittle with age, as many do. If those tabs start breaking off, it could become a major headache...
4-Feb-2003 Took a little siesta there, but I'm back now. Looking at some of the boards, I noticed an interesting thing. Along the edge of the board that faces outward when the board is in place in the card cage, there is a row of small holes, and traces on the circuit card lead up to them. These have to be test-points; you could stick a probe in there to see what's going on while the machine is running. Nice idea.
28-Feb-2001 Looked at those boards a little more, and found more chips with both markings: 2788/3005662 and 2789/3005663. I wonder if the pattern continues?
27-Feb-2001 I bought a handful of Univac circuit board from e-bay, and they arrived the other day. My best guess, from their shape & part numbers, is that all of them are used in my machine. And there is one interesting little discovery: on these boards, some of the chips have both TI part numbers and those "mystery" part numbers. For instance, TI's SN2790 also shows the part number 3005664; those two parts must be compatible. So finding info on either one will help with the other. Since the card-punch uses some of the same chips, and its service manual has schematics, maybe I can figure out something about those chips. (That is, if I can figure out their wacko(TM) notation!)
I'm starting to think about how to make it go. Rebuilding the power supply is going to be a pain. As a shortcut, maybe I'll first try disconnecting the printer, and just getting the CPU to run with the front panel for I/O. That might mean needing fewer distinct voltages, and a lot less current. Of course, even then, this is gonna be a huge job. There are empty card slots, and I don't know what belongs in them. And I don't have any schematics, so it'll take some reverse-engineering.
From what I've seen of the circuit cards, they seem pretty standardized. Most have just a few ICs, with each pin connected to the card-edge connector. So the backplane is going to be a spaghetti bowl of point-to-point wiring. Debugging that is going to be some real fun.
A lot of the boards in the card-cage have IC's on them, but there are also a lot of discrete transistors, and a ton of resistors and small caps. The IC's seem to be of two types: some made by TI with numbers like 2789 and 2790, and some with no recognizable manufacturer code with numbers like 3005662 and 3005667. They all seem to be 14-pin DIPs. (Or maybe 16-pin; I didn't actually count pins.) So far, I have no idea what these are. Maybe RTL or DTL?
There are piles of decomposing black foam rubber lying around inside, which probably once functioned as sound insulation. I noticed a tag on a motor that said "re-lubricate every ten years". This was not meant to be a disposable machine.
My next steps will have to be documenting and cleaning it.