22 Jun
2000
22 Jun
'00
3:33 p.m.
It was the timing problem... the failure rate (soft)
for the alpha
particle thing was so low you'd see power hits and other gremlins
first.
S100 system were prone to bus noise (even with wonderbus) so where a
card was installed could litterally mean fail/flakey/works for the same
card! Bus termination schemes were used to help but the 22 slot bus was
too long and a more modest 18 or 12 slot was always more reliable.
Dram cards were by and large trouble as they were most sensitive to timing
problems. I always ran static for testing and some systems for that
reason.
I never had the opportunity to work with a machine that was IEEE-696
compliant. I'm assuming you did (Danger Will Robinson!)... were compliant
boards reliable? Do you know of any attempts to retrofit compliance onto
existing S-100 designs?
-dq