29 May
2012
29 May
'12
2:12 p.m.
Further testing tonight. When the deposit switch is toggled, it does
increment the address. I have verified that the processor can write to
memory, by toggling in a program with a known good control board, then
swapping in the bad one and running the program. I have also verified
(with a logic probe and creative cabling) that the basic switch logic
(consisting of a couple of flip flops and a 7400) does appear to be
working (chips E97-99).
It would seem that the actual switch register/buffer is located on the
data path board, so it's not a problem with that itself.
I think if the data path from the swtiches was the problem, then the
thing would write 000000 or 177777 or something constistent to memory
every time you hit Deposit, it would not leave the memeory unchanged.
What I swould do next is see just what the Unibus is doing when you hit
Deposit. In particualr, Is MSYN plusing low (indicating abus cycle is
occuring). Do C0 and C1 do anything? You can pick these up on the E
conenctor of an SPC slot or on the Unibus terminator so it shouldn't be
too hard to test them.
My tirck for not needing an exteder board in low-ish speed machines
(PDP11s genrally qulaify!), is to pul lthe board I am interested in and
solder wires to the IC pins I am interested in montioring (Just
tack-sodler them on the track side). Put the board back in, make sure the
fee ends of the wires don't short. Then power up, and probe the free
ends. Power down, pul lthe board, resolde the wires to the next set of
points (beased on the results of the first set of tests) and so on.
I gneerally find 2 or 3 iterations lets me find the fault.
-tony