Re: CAT Box Project (interest gauge)

From: Clay Cowgill <clayc_at_diamondmm.com>
Date: Fri Jan 09 1998 - 17:52:34 EST

>2. A complete redesign. Not just a catbox "improvement". Here I think
> that an ICE would be the best. Anyone willing to help architect this?
> It could go as far as being customized to ICE the favorite game
> processors -- 6502, 6809, Z80, etc.

Werner Sharp and I were batting this one around a bit about a year ago.
Our plan was to just get one of those off-the-shelf ISA I/O cards
(something like two 8255's on a card) that gave 48 lines of I/O. We were
then going to just make a little "processor pod" that plugged into the CPU
socket and bit-banged the processor lines using the MAME CPU emulation
cores. The timing wouldn't be right, but you could do quite a bit with it.
(As far as exercising stuff on the board.)

I was thinking how cool it would be to have a picture of (say...) an
Asteroids board on the screen, double click on a RAM chip (running an
automatic test) and color "good paths/chips" green and "probably bad
paths/chips" red. ;-)

(It's kind-of a non-realtime, PC version of a Fluke 9010A, but you could
build in more "intiution" about how certain boards worked. So, on a board
with multiple banks of RAM that constantly had a bad bit "D3" the software
would be smart enough to try a couple things to see if "D3" worked to other
locations (ruling out the CPU), and work back through some rules to suggest
shorted/open traces of bad buffers.)

You could also do the "learn" mode stuff like the Fluke 9010A-- plug in a
known good board and turn loose a "profiler" that finds all r/w registers,
ROM (with checksums), etc. Next time you have a "bad" board you plug it in
and hit "auto test" and it does all the basics for you.

Non-trivial project. ;-)

-Clay

Clayton N. Cowgill Engineering Manager
_______________________________________________________________________
/\ Diamond Multimedia Systems, Inc. clay@supra.com
\/ Communications Division http://www.supra.com/
Received on Fri Jan 9 15:45:02 1998

This archive was generated by hypermail 2.1.8 : Fri Aug 01 2003 - 00:31:05 EDT