hamilton@intersil.uucp (Fred Hamilton) (12/10/89)
<Do we really even NEED lineater lines anym
Well, after all the QUESTIONS I was posting when my FRANCES was sick and I
was desperate, I feel obligated to post some of my findings for others who
may be in similar states.
I was having two problems with my LUCAS/FRANCES combo:
1) The machine was locking up (not GURUing, just freezing up) randomly.
After a few days of this I asked myself, what have I changed since the last
time it was working correctly? The answer was: the 74LS245 chips.
When I first built the thing I had a mis-matched quartet of 74LS245s.
When the system developed a serious problem, I replaced them with fresh,
identical LS245s. After a long time I fixed the big problem and started
getting mysterious crashes. I put the old "gangly" (this week's
vocabulary word) chips back in and it worked fine. I was just about to
try the *CAS hack. So before anyone tries the *CAS hack they may want
to try replacing the LS245s with another brand.
2) When I ran programs that used the 68881 extensively (TurboSilver), I
would usually get a 0000000B floating point guru sometime during the
rendering. I had the luxury of being able to try different 68881/68020
combinations, and the one I have now works fine. I don't think the old
881/020 set was bad, I just think LUCAS may work better with some "good"
ones than others with slightly different timing specs. The guy at
Impulse said he had heard of other LUCAS boards having floating point
problems. This may be a fix.
Well, that's it. Good luck out there.
--
Fred Hamilton Any views, comments, or ideas expressed here
Harris Semiconductor are entirely my own. Even good ones.
Santa Clara, CA