From: firstname.lastname@example.org (N. Thornton)
Subject: Re: Strange failure mode in switchmode power supplies: Followup
Date: 23 Sep 2002 17:40:08 -0700
NNTP-Posting-Date: 24 Sep 2002 00:40:09 GMT
> Peter Gutmann wrote:
> The immediately obvious symptoms are that any SMPS plugged into
the power in
> my house starts producing large amounts of high-frequency
> switching noise, both audibly and in its DC output. At 50W it's
> inaudible (but enough to cause the usual unpleasant effects of
> if you sit next to it for awhile), at 100W you can sense it
> not necessarily directly hear it), and at 200W it's so bad you
> being in the room for more than a few minutes without feeling
> What's bad for humans is even worse for any elecronics sitting
> so far this problem has caused about $2K of damage in assorted
> routers, etc etc (the damage is pretty spectacular, one device
> was an online UPS which I'd been using to protect the PC
> high-speed switching component in it had been fried).
Well, I have a few thoughts... but nothing solid.
Rf energy does strike me as one possible. It could be airborne or on
the power line. It could be microwave or lower freq.
2ndly a transformer will not eliminate such hf stuff. TFs have
distributed capacitance all over them, and high frequencies run thru
them like water thru a pipe. Its a well known issue in rf design.
Inductor filters in the power line will have the same problem:
inductors frequently pass hf very well. I would go for an RC power
filter: a non inductive non capacitive resistor in each lead, followed
by a cap across the 2 wires. You can generally drop 10% of mains V and
have things still work. I dont trust your ferroresonant to do it at
all, not for hf stuff.
The other thought is you could have a much greater than normal power
line inductance. There could be a curly section of cable, it could be
wrapped round something (iron), or it might just run next to iron
pipe, and so on. So I wouldnt assume that the usual supply inductance
values will hold true for your supply.
Only R will damp oscillation: LC can do all sorts, epending on the
details of the Ls. Putting C across the line will make a nice resonant
circuit. So put the power thru R before C across.
And finally you could be the subject of an alien experiment, to see
how much rf you can take before your brain fries :)