On Mon, Sep 13, 2004 at 01:11:49PM -0700, vrs wrote:
> I have been looking at the R210 module when thinking about feasibility...
>
> I think it would be interesting to make one with an edge connector (so it
> could be tested in a real straight-8), then "smash" it into the area
> formerly taken by the socket. In principle, you could do that to the whole
> CPU, then decide if it was worth the trouble to re-arrange the components
> (or not) after you were fairly sure of the correctness of your
> implementation.
One could take a single and a double-height proto card, solder on an
appropriate sized .1"-spacing socket, then make your proto SMT designs
on a board with pins and test whatever particular module design was
under review at the time.
Saves on edge connectors for everything, especially if you want to go
the proper gold-finger route.
OTOH, doing all the protos with their own edge fingers means that you
could, theoretically, use these as replacement parts for the real thing,
but in practice, the real ones are easy enough to diagnose and fix that
it would probably only matter if you were _missing_ cards, not had a
box of broken ones.
Long ago (1984?) I made an M-series tester that I never completed. I
finished the hardware (a 44-pin RadioShack proto card with a 6821, a
power switch, and a single DEC backplane socket), but never wrote the
VIC-20 software to test any cards (what I later did instead was to
make an extension clip for a commercial IC tester and now I test
many kinds of M-series cards on the table-top, with the IC tester
powering on all (3 to 4) ICs on the card, but only testing one chip
at a time... it works for any card that doesn't have internal
interconnects between the chips, so not the M220 or M706/M707, etc.,
but it does test M111, M113, M117...). One observation to share for
testing M-series cards - I have seen more 7474 and 7440 failures than
anything else; in fact, probably 90% of the time, a dead -8/L is one
chip or the other has failed on some card. Rarely it's a 7400 or 7404.
As I have a quantity (enough to fill several racks) of R-series stuff,
I've always thought about what it would take to do something similar
for those modules... even if I'm using an O-scope to check transition
levels, flip-flopping, etc., it would be nice to have a software-
controlled socket that knew what lines to toggle for me so I could
just watch the scope. Has anyone here ever made any automated or
semi-automated debugging equipment for pre-TTL logic?
In terms of the PDP-8/S-on-a-board project... I have a real PDP-8/S
that still needs to be fixed (two biggest problems are broken lamps
and a panel-lock switch that is sensed as always locked - i.e., the
switch tests out, so it must be a transistor or related component
that monitors it). Once I delve into that machine enough to get it
working, I'd probably a lot more interested in details of a modern
transistor-based replica. For now, I'm just trying to get my head
five years further into the past than it is now (I'm deficient in
wide swaths of pre-TTL gear, but we happen to have multiple copies
of "The Art of Electronics" down here, and I'm better than I used
to be).
-ethan
--
Ethan Dicks, A-130-S Current South Pole Weather at 13-Sep-2004 22:20 Z
South Pole Station
PSC 468 Box 400 Temp -43.3 F (-41.8 C) Windchill -76.5 F (-60.3 C)
APO AP 96598 Wind 14.7 kts Grid 336 Barometer 674.6 mb (10832. ft)
Ethan.Dicks_at_amanda.spole.gov http://penguincentral.com/penguincentral.html
Received on Mon Sep 13 2004 - 17:58:12 BST