[vvuk]

Yep, I've got a second DG5-2 coming (with TVO even) so hopefully that will be all I need to get it up and running.  

I took the TVO option board off and took a close look at the main DG5-2.  Zero visible damage or even anything that looked off -- pretty sure it's just been enclosed in the cabinet most of its life.  Virtually no dust.  Two bodge wires that are on every image I've seen, both properly connected still.  I was hoping to find something that looked like a small microcontroller or other bus arbiter; all I could find that had meaningful part numbers was a PLL clock generator, and a pile of 16-bit bus transcievers, all of which looked visibly fine.  There are two large custom chips clearly labeled "VOC 1b"; both BGA, both still solidly soldered.  But since register access on both is failing identically, the problem is definitely somewhere earlier.

irsaudit is quite happy with everything else, so it's either the DG5-2 or it's the backplane, which would be.. not great. It *could* also be the KTOWN, and the ARM controller that (I believe) lives there and manages the overall IR system.  But given that the other 5 cards (GE + 4 RM) pass all tests, all signs point to the DG5-2.

A little side interesting bit -- trying the board with TVO removed, irsaudit knew what I did and complained, but given it sounds like folks move the option boards over without problem it's not an issue in practice:

Code:

TRCE                DG NIC #: 0000.0059.c878 (family: 0b)
TRCE                    Serial #: ......
TRCE                    Part #: 030-1242-001
INFO                DG NIC reports DG w/ option board but no option board detected
INFO                NIC part number field must be one of 030-1055-XXX (2 VOCs),
INFO                  030-1087-XXX (8 VOCs),030-1524-XXX OR030-1242-XXX (2 VOCs
INFO                  with GVO/DDO2/DPLEX/TVO).
DIAG                Fix by correctly reprogramming the DG NIC.
TRCE                    rev_code: H
TRCE                    name: DG5-2

There's a little 2-pin header labelled "NIC" in the corner of the board too; sometime in the next week or two I may stick a logic analyzer on it and see what I can see.  I have a working theory it's i2c just due to VOC & other major chips manufactured by Philips Semi, where I2C was big in the 90s.  (Though you'd need to pull a common ground from somewhere else; it's a little weird to not just put a ground pin in the same location)

I've got some friends-of-friends who may have direct knowledge of the hardware.  I'll ask and see if I can get a bit more detail about how some of the comms here works.  Probably won't help me with this board, but may be useful info to someone down the line.