Tim Williams wrote:

"Jon Elson" wrote in message
servers.com...


That's insane! Try some power FETs!



How 'bout IGBT's?...

IGBT's are great ar 20-25 KHz, but their appeal rapidly drops above that.
There are "ultra fast" version that can go a little higher, but I think
100 KHz
is totally out of range. Also, FETs are massively tolerant of overloads and
short-duration short circuits, while IGBTs are very INtolerant of those
conditions. (Don't ask me how I know that, but it involves parts of IGBTs
passing my head simultaneously on either side, and a noise that sounded like
a pistol shot indoors.)




They are ten times better than
tubes for this frequency range (50 - 100 KHz). You can get 600 V 30 A
transistors under $10. Want to compare the price with a 4CX250?



No, because I already have said tubes, curtesy of Mark Winlund.
Forgot to mention, I also have a dead printing arclight supply, 10kW.
Damage seemed to be on a contactor so I'm sure the 200/220/240 800V
transformer is still good, as well as the several chunky capacitors. If I
FWB that winding, I get 1200V at oh, 5A? Obviously I can't use too much of
that on a pair of 4CX250R's, so it'd be nice to have some beefier things.
If I go with a doubler, I've got 2.4kV. Anyone got a 2kW pentode on hand?



Running conservatively, you could get 4 KW out of such a transistor
(400 V, 10 A).



Unforunately, 600V is not only under the minimum output voltage (say I did
choke input filtering, and wow....what a nice choke that would have to
be....) but well under the peak voltage seen in such a circuit. Not sure
about 30A either, but in class C things come in short bursts, I've got to
imagine such a transistor would only be able to run +250VDC 5A... 1.25kW
ain't bad, but if I'm going to buy some parts I might as well get some that
can handle the whole thing.


You can get 1200 V and higher FETs, only a little more expensive. You
could definitely
go higher on the current and voltage. If you are running the FETs on
short pulses,
the usable current would go up, too. Also, the drive requirements could
be much
simpler, and you could probably have 12 V CMOS control circuits that
could shut it
off within microseconds if overloaded. I guess you could do all that
with tubes, too,
with only a little translation circuitry.

Jon