Hi Marco,
    Try a COMMON MODE CHOKE between the power supply and the spark
gap.  Run the ground and HV lines (RG-58 stripped of braid) in
parallel.  Wind them around a big transformer core in parallel.  Wind
them a second time around a second transformer in antiparallel.  This
should tame the common mode and differential mode spikes.  (think
giant line Corcom filter).  This must be the only ground path between
your power supply and the spark gap.  Don't forget the capacitive
coupling between the power supply line (isolation) transformer primary
and secondaries.  If you float your power supply with an isolation
transformer, connect it to earth ground with at least a 10 MegOhm
resistor.  This will prevent DC voltage accumulation from high voltage
corona rectification effects.  May also save your isolation


----- Original Message -----
From: "Tesla List" <tesla-at-pupman-dot-com>
To: <tesla-at-pupman-dot-com>
Sent: Wednesday, December 22, 1999 1:32 PM

> Original Poster: "Marco Denicolai" <Marco.Denicolai-at-tellabs.fi>
> Hello everybody!
> >From the dungeons of the Thor project, I came to the surface to
tell you
> briefly
> about what I'm doing.
> I have put together in a 19" rack a down-scaled version of my
switched power
> supply (that is, one half of the final version). That is capable of
> up to 10 kV DC (from 0 to 10kV) at a current of 1A max. All
electronics is
> mounted on PCBs, the HV transformers too (!). The unit has been
> fine for
> a month now: the voltage stability is about +/- 150V.
> The problem is that yesterday I connected it to my old TC (removed
to old MOT
> supply): after a few bangs the IGBT driver chip died. Changed the
chip, a few
> bangs and again: rest in peace.
> The transients from the spark gap gets induced back into the
switcher primary
> side, with a devastating effect: I know I should have tought about
that, but I
> was so much concentrated about getting the switcher to work that I
haven't had
> time to think about protection.
> What I am going to do now is to try converting my schematics from
> into PDF and make them publicly available on my web side, with part
list and
> explanations.
> This way I hope to involve also some of you guys in helping me
designing a good
> protection strategy for the power supply.
> I'll be back soon, but for now Merry Christmas to you all!