Failure of my classic computer

From: Dwight K. Elvey <dwight.elvey_at_amd.com>
Date: Mon Sep 13 14:24:30 2004

Hi
 Well it happened Sunday. Right in the middle of copying
some disk on my Poly 8813, it died. I tried to reboot
but nothing seemed to work.
 I tried all the obvious things like swapping the RAM addresses
around on the main RAM boards and swapping RAM chips around on
the CPU board before I realized that I was going to have to
trace it out.
 The Poly's have a built in machine video front panel monitor.
With all the boot ROMs loaded, it wouldn't come up but if
I removed one of the ROMs, I was able to get the monitor working.
I did some quick checks of the floppy controller card and
that seemed to work. Placing the ROMs in the extra socket,
I wrote a simple checksum. Again, not problem found. They
all matched my copies that I'd filed away( One should always
have backup of machine ROMs ).
 I guess now, I was forced to look at the symptoms and
use that to track down the issues. Something I should have
done from the start but I kept expecting something simple.
I noted that the processor was clearing the video and
starting the floppy but not doing the steps to track zero.
When the ROM was removed, all seemed to work OK so the
basic processor operation was OK. It was something in the
process of getting the floppy read. I reentered the monitor
and was able to write some simple code to verify that, yes,
the controller could step both in and out. Still, this was
were it seemed to be hanging.
 A while back, I'd made a listing of the code so I figured
I'd look through that. Following the code, I could see
the video clearing and then the setting up of the disk
access. I could see where the drive was selected and then
there was a wait for delay. Bingo, the timer interrupt
was not happening.
 Now I had something to trace down. Following the interrupt
ended me on a daughter card that was tacked onto the CPU
card ( I didn't have the schematic for the daughter card
but it only had 3 chips ). A little more looking and I
found a miss behaving 74LS75. I found a spare on one of my
scrap boards and replaced it.
 After about 4 hours, I was back in operation. Why is it
that the latching and flopping ttl parts seem to have such a
high failure rate?
Dwight
Received on Mon Sep 13 2004 - 14:24:30 BST