The Cyber-Spy.Com Usenet Archive Feeds Directly
From The Open And Publicly Available Newsgroup
This Group And Thousands Of Others Are Available
On Most IS NNTP News Servers On Port 119.
Cyber-Spy.Com Is NOT Responsible For Any Topic,
Opinions Or Content Posted To This Or Any Other
Newsgroup. This Web Archive Of The Newsgroup And
Posts Are For Informational Purposes Only.
From: email@example.com (Peter Gutmann)
Subject: Re: Strange failure mode in switchmode power supplies: Followup
Date: Wed, 25 Sep 2002 05:43:55 -0000
Organization: Posted via Supernews, http://www.supernews.com
Sender: firstname.lastname@example.org (Peter Gutmann)
"Harry Dellamano" writes:
> I agree with what you say but if it common mode energy, what will it
>affect in the SMPS? I see the Y caps smoking but after that what else? It
>will not pass a EMI conducted emissions test but it would still be
>operational. Will it cause the SMPS to grunt and groan? I think not.
> The OP statement "one device which failed was an online UPS which I'd been
>using to protect the PC downstream, every high-speed switching component in
>it had been fried" is very suspicious. Maybe he is on a fishing expedition
>or just trolling. Sorry Peter G. but I had to say that. Can we get some
>digital pics of the destruction?
I'm most definitely not trolling... I wouldn't be making something this ugly
up, even if it does come from the Steven King school of electrical problems.
Here's an attempt to reconstruct what was replaced in the UPS, based on the
parts that look freshly-soldered. Note that some of these may have been merely
removed for testing and reinstalled, and there were a few devices for which
access was blocked by other parts where I couldn't read the type:
GT15Q101 x3 (maybe 4)
What looks untouched alongside those is:
To anticipate the question, the repair was done by a friend of mine using
(among other things) his personal stash of spares, so it's highly unlikely
there was any gold-plating involved.
Just for reference, the regulators are a UC3525 and a 3854, and for protection
there's a fairly complex multi-stage custom LC filter on the input (and a
generic Delta inline filter on the output). There's also a DAC0800 and
ADC0838 (finally some parts I understand :-) on the digital control board,
although they don't seem particularly appropriate there, particularly the DAC.
In case this is any kind of clue, but otherwise on a more or less unrelated
note, use of the ferroresonant transformer I mentioned in an earlier post just
destroyed my last remaining hot-spare monitor. I'd been running the computer
off the ferro and the monitor off straight mains power because I was a bit
concerned about running two (possibly conflicting) nonlinear loads off the
transformer (there's enough weirdness involved already without adding possible
interactions between all of those). The transformer was turned on (it's in
another room because of the noise) but the computer was off, and it's old
enough to not run a standby circuit. When I was setting up a new video card,
I got a slight shock, nowhere near the full 240V but quite noticeable. I
assumed it was just a badly-built PC (I've had plenty of these in the past)
and didn't think about it further... until the monitor I was using died. When
I unplugged the monitor cable, I noticed arcing from the metal D-shell to the
outside of the socket on the video card, which explains the dead monitor (the
same thing happened with my final remaining monitor, so it wasn't the
Given the path the current was taking, I dind't stop to take measurements,
sorry :-). Checking the ferro indicates no problem with the earthing, with a
clean connection from input to output, however there was obviously some form
of ground loop through the computer and monitor which wasn't present when
running both off the same mains outlet. This may be a clue (different ground
potential in different rooms? Most odd), when I get back tonight I'll see if
I can spot anything and post results.
In the meantime, many thanks to everyone who's been offering suggestions, I'm
working my way through them as time and means allow.
Go Back To The Cyber-Spy.Com
Usenet Web Archive Index Of
The sci.electronics.design Newsgroup