On Sun, Sep 12, 2004 at 09:28:25PM -0700, vrs wrote:
An SMT
transistorized machine would be an interesting project, but hell to
debug I would imagine.
One could perhaps build it out of modules -- make SMT versions of Rxxx
modules, etc., then use those to build up the CPU. Then one could debug at
the module level before tackling the whole thing.
One could, but the interconnection thing would be a killer - I suppose the
cheapest way to do it might be a "fatherboard" arrangement with a nest of
pins for the backplane, and pin receptacles on the modules (though the
more traditional pins-on-the-module-and-sockets-on-the-motherboard would
be about the same cost, probably). In the case of a PDP-8 design, there
are several busses that might have to be run, but you could sort-of cheat
and aggregate the registers (like is done with the 6 x M220 cards in an
-8/i or -8/L) to group things together. I would expect that in addition
to the registers themselves (AC, MD, MA...) whatever memory there was would
fit nicely in the same area.
-ethan
--
Ethan Dicks, A-130-S Current South Pole Weather at 13-Sep-2004 04:30 Z
South Pole Station
PSC 468 Box 400 Temp -53.6 F (-47.6 C) Windchill -83.90 F (-64.40 C)
APO AP 96598 Wind 9.69 kts Grid 339 Barometer 671.6 mb (10945. ft)
Ethan.Dicks(a)amanda.spole.gov
http://penguincentral.com/penguincentral.html