Path: utzoo!attcan!uunet!samsung!usc!snorkelwacker!mit-eddie!uw-beaver!uw-entropy!dataio!gtenmc!stumpf
From: stumpf@gtenmc.UUCP (Jon S. Stumpf)
Newsgroups: comp.sys.att
Subject: 3B1 Power Supply fried
Keywords: 3B1 fire power_gremlins
Message-ID: <544@gtenmc.UUCP>
Date: 8 Feb 90 02:44:08 GMT
Organization: GTE Telecom Inc., Bothell, WA
Lines: 45


I have a power supply dead and no other means to verify other components
are not affected as well.  Visual inspection reveals only the power
supply has problems.  Power supply problems were confirmed with simple
tests with a voltmeter.

I am looking for suggestions to repair this machine.  I will send it out
to get fixed, buy new parts and do it myself, but I would prefer not to
buy another PC and cannibalize.  Thanks in advance.  

jss


The gruesome story follows.

My UNIX PC had been powered down for a month since I was going to be
away and no one would be around to monitor it.  Prior to my return,
my father powered it up so I could do some uucp stuff.  He said he came
back after the initialization sequence, logged in, switch the phone to
data and started some other things and walked away.

When he returned five minutes later, he noticed the machine had tried to
reboot and the disk was speeding up (as in accelerating) and winding down
in a 5-10 second cycle.  He powered it down, waited a few minutes, and
powered it up again.  It never got to the point to access the disk.

When he reached back to power it off again, he noticed the top of the
case (over the power supply) was hot.  He said it would burn his hand
if he left it on the plastic.  So, extremely hot.

Taking apart the machine and looking at various components, I noticed
scorch marks on the power supply circuit board.  Testing the components
that had connections in the scorch marks revealed 8 out of 9 blue resistors
(I don't do EE  8-/) and one capacitor out of commission.  There is also
a proverbial black box, a cube (actually), that I have no idea what it is
or does or how to test it.  It is sitting right next to the capacitor
and shares a connection with it.

I have a power strip with a line filter (big deal) and the various fuses
at the power cord entry point on the PC were intact.  If a component,
or combination thereof, was drawing too much power and caused this problem,
how do I prevent this from happening again?

-- 
 jss - Jon S. Stumpf