D.G. Nova update... disk drive again
Dwight K. Elvey
dwight.elvey at amd.com
Wed Apr 6 18:19:49 CDT 2005
>From: "Tom Jennings" <tomj at wps.com>
>Well all seemed fine, I was working on Kermit, making good
>progress, then I/O PARITY ERROR AT xxxx -- disk read error. Diags
>etc, another shows up. Then another, but a previous goes away...
>Clearly there is a soft(ish) read error. Probably all those cheap
>ceramic disks on the read/write/amp board. So I will order nice,
>new, correct monolithic caps, and replace all the replacements.
>It's almost certainly in the analog read/write section, and not
>hard logic, said hypothesis boosted by the fact of my fiddling in
>I ran memory diags and all that stuff, but I am assuming it's the
>electronics I fiddled. I'll go do the job Right.
I know that you'd like to blame the capacitors but
if these were all bypass capacitors, it is unlikely that
they are the source of your problem.
I have found that when bringing up older electronics of
this complexity, that has not been operational for some time,
that active parts tend to fail over the next 2 or 3 months
I expect that the culprit is moisture. When the unit is always
powered, the heat of each part keeps moisture from accumulating
within the plastic IC packages. When turned off for some
time, moisture builds up inside the parts. When power
is applied, this causes some electrolysis inside that destroys
I've thought that it might be better to bake the boards in
a dry N2 environment at about 120F for a few weeks before
powering on units that have been sitting for a long time.
Not having any way to confirm if this would work, I've
not been able to verify it.
This would be bad for electrolytic capacitors but these
are often more easily located and replaced.
Good luck, we are all hoping things will be easy to
find for you.
More information about the cctalk