18 May
2008
18 May
'08
5:48 p.m.
Philip Pemberton wrote:
I've just seen the MPEG -- looks like what
you're seeing is
the counter (or a latch, flipflop or buffer) sticking. It's
holding an address for a few characters instead of a single
character period.
If you run the video frame-by-frame, you see it displaying:
Loading CP/M and HELP...
which turns into
LLLLLnnnnn////nd HHHHHH.......
or something to that effect - the video isn't fantastically clear.
Excellent spot, thanks. Apologies about the poor video - it was taken
handheld using an elderly digital camera (Olympus UZ-2100 - in about 2 years
time, it'll qualify as a classic for this list!).
So you want to look at:
- counters
- buffers / flipflops / latches
- latch / enable logic for the above
That's on top of checking the PSU as I mentioned in my last email.
I've almost ruled out the PSU, the +5v and +12v lines are very stable, even
under load (booting). I've also got the schematics (thanks to a link found
by Roy Tellason), so I can probably find the duff chip -- well, I can try at
least. I've got the 'scope working too, but as it's not a storage scope,
seeing logic waveforms is not easy (impossible, in fact).
So... how do I find a sticky counter? Is it likely to be an on-chip fault,
or a failed track? My guess is on-chip, given that the video is clear &
stable when the machine is crashed.
Also, for clarity: Underneath the video fault, the machine is performing
perfectly. It only crashes when I pull the Ext Video buffer & plug it back
in - which physically disconnects some of the video signals from the onboard
monitor, and clearly does something electrical which causes the machine to
freeze up.
One other thing - all of the chips (bar 4) are soldered to the board. This
isn't a big problem for me, I'm reasonably adept at soldering, even when
I've got lots of legs to free up; but it will be time consuming :\
Cheers,
Ade.
No virus found in this outgoing message.
Checked by AVG.
Version: 7.5.524 / Virus Database: 269.23.20/1453 - Release Date: 18/05/2008
09:31