Path: utzoo!utgpu!news-server.csri.toronto.edu!cs.utexas.edu!usc!randvax!edhall
From: edhall@rand.org (Ed Hall)
Newsgroups: comp.sys.ibm.pc.hardware
Subject: Re: Power Supply: Self Repair?
Message-ID: <1991Jun14.182515.24877@rand.org>
Date: 14 Jun 91 18:25:15 GMT
References: <ONEEL.91Jun10091204@heawk1.gsfc.nasa.gov> <1991Jun11.021150.7238@rand.org> <7079@bwdls58.bnr.ca>
Sender: usenet@rand.org
Organization: The RAND Corporation
Lines: 26
Nntp-Posting-Host: ives

In article <7079@bwdls58.bnr.ca> mlord@bwdls58.bnr.ca (Mark Lord) writes:
[I originally wrote:]
><Once having had a significant fraction of a 25MHz '386 system fried
><by a failed power supply, I strongly second the advice.  Even if you
><manage to repair the supply, it may not be so kind the next time it
><fails (i.e. one strike and it's out)!
>
>Just to keep the discussion more balanced, it might be useful to know that
>at least somebody (me) has successfully repaired switching power supplies
>and *not* had any further problems with them.  Twice, in fact (different
>supplies).  

Hey, so have I.  In fact, the supply which turned my system into a
toaster had previously blown the -12V rectifier.  A couple of diodes
and a new filter cap (who knows what sort of surge the old one
received when the original diodes went...) and everything worked just
fine.  That is, until some months later...  BTW, I checked my
replacement diodes after the second failure (which were on a
relatively uncharred part of the circuit board) and they were still
good.

Just a single data point, mind you, but I like to think that the first
failure was a warning...

		-Ed Hall
		edhall@rand.org