Odd memory error in PDP-11/04

[Ethan Dicks]

Hi, All,

I've finally tracked down the simplest problems in my PDP-11/04 that's
been sitting unused for many years (the one that we formerly used as a
hardware test platform for Unibus COMBOARDs 30 years ago).  The
primary faults were a half-bad 7474 in the console (the flip-flop
attached to the Run LED was not toggling when it should) _and_ an
apparently bad DL11-W that isn't passing grant - for now replaced with
a fully-functional module, but to be debugged later).

What's happening now, since I can finally enter octal at the console,
is when I change one location, 000000 for example, it echoes across
multiple locations...

To whit:

I use the console to fill 000000 to 000040 with zeros, and verify they
are all zeros.  I deposit 177777 in 000000 and I get back the
following...

@L 1000
@D 177777
@L 1000
@E 001000 177777
@E 001002 004000
@E 001004 004000
@E 001006 004000
@E 001010 004000
@E 001012 004000
@E 001014 004000
@E 001016 004000
@E 001020 177777
@E 001022 004000
@E 001024 004000
@E 001026 004000
@E 001030 004000
@E 001032 004000
@E 001034 004000
@E 001036 004000
@E 001040 004000
@E 001042 177773
@E 001044 177777

I replace the 000000 at 000000 and get all zeros.

So I appear to have two problems:

1) Depositing any value is echoed 000020 later.

2) Setting D10 in location 000000 results in D10 set in all the locations

I have few spares for this machine.  Lots of spares for my 11/34
(which I will want to test at some point soon), but this box (BA11-L)
has a DD11DK not a DD11-PK, so I can't just upgrade in place.

Does this sound like a dodgy CPU, dodgy RAM or both?

I have this minimally loaded...

M7263 PDP-11/04 processor
M7847 16K MOS RAM (half loaded)
M9312 with console ROM and papertape boot ROM (I have more ROMs available)
M7856 - DL11-W strapped to defaults as a console/RTC
many dual-height grant cards
M9302 terminator
UA-11 debugging board

Oh... and I see while typing that it just started halting immediately
after reset... so something just broke while it was powered on and
running the console ODT.  :-(  I guess I'm back to low-level hardware
debugging again.

But in the meantime, any CPU/RAM symptom suggestions?  I have the
prints.  I'm just looking for any "oh, yeah!  That happened to me!"
with "...and I fixed it by testing X and replacing the bad ones".

All this so I can make a test bed for my M9313 boards and get them
working again to fix the DWBUA on my VAX 8300...

-ethan