Kiloamps induce curents (was: Economizer circuit)

[Lee Hart <leeahart@earthlink.net>]

Rich Rudman wrote:

A friend of mine works at Fermilab, where they play with REALLY high
power. He has wonderful stories of the little surprises that happen at
millions of amps and millions of volts.

Here's one I ran into: The power transistor on a switchmode power supply
ran too hot. The previous engineer had tried everything; bigger
transistor, better gate driver, snubbers, bigger heatsink, etc. He quit,
and I got the the project (my turn in the barrel :-)

I noticed that the transistor and its heatsink were right up against the
core of switching transformer. The transformer itself didn't get hot;
but was it somehow affecting the heatsink? I disconnected the transistor
electrically, but left it and its heatsink physically in place. A second
transistor and heatsink was mounted on the other side of the PC board,
away from the core.

The unpowered transistor's heatsink got almost as hot as it did when
powered; the transistor and heatsink away from the core was much cooler.
Cause? Leakage flux from the core saw the heatsink as a shorted turn,
which got hot.

Strangely enough, the cure was to replace the thin aluminum heatsink
with a thicker copper one. The copper had less resistance, and so
reduced eddy current losses.

Designers and users of kiloamp controllers; I'd watch out for this.
Anything near the buss bars and high-current wiring is going to have
significant currents and noise injected into it.
-- 
Lee A. Hart                     Ring the bells that you can ring
4209 France Ave. N.             Forget the perfect offering
Robbinsdale, MN 55422 USA       There is a crack in everything
phone (612) 533-3226            That's how the light gets in
leeahart@earthlink.net                  Leonard Cohen