New subject: DEC M220 as a board-level bit-slice device (was Re: Inside old games machines)

19 Jun 2007

On 6/19/07, Brent Hilpert <hilpert at cs.ubc.ca> wrote:
...
  Before bit-slice chips, the HP 2116 was constructed in
'bit-slice form' from
 SSI: 4 boards plugged into the backplane are identical, each board containing
 4 bits of the 16-bit ALU and main registers. I expect other SSI-era minis were
 done this way. 
The PDP-8/L and PDP-8/i used this technique - six identical M220
"major registers" boards provided 2 bits each of the AC, PC, MA, and
MB registers and the adders.

http://www.bitsavers.org/pdf/dec/modules/mSeries/M220.pdf

...
 From a glance at the schematics, it's about
30-some-odd inputs (various enables and adder/shift signals) and almost as many
outputs
on a dual-height card (72 pins total, including 2 Vcc and 4 GND).
Looking at how the logic breaks down, perhaps it might be
reimplementable with a pair of 22V10s for the lower part of the flow
(combining enables and register bits into the adders), then perhaps a
pair of 16V8s for the middle part (shifts and adds), then piped out
to, perhaps, one 22V10 and then, if required for fanout, a quad buffer
of some type to take the place of the two 7440s - six more modern
parts to replace a stack of 7474s, 7483s, 7460s and 7440s and a 7482,
of which, I think, only the 7474s are easy to find anymore.

I only bring up the M220 at this level because I have a few dead M220s
and will need to resort to component level repair or board-level
replacement to keep my -8/i and -8/Ls running.  From my existing
experience, though, I'd start by ensuring my 7474s and 7440s are
working as expected - I've replaced several in other boards over the
years.  They seem to be the two most common ICs to fail in my PDP-8/L.

-ethan