If you had this supply and switch, what would you build?

["Tesla list" <tesla@xxxxxxxxxx>]

Original poster: "Mike" <induction@xxxxxxxxxxx>

Hello,
         Given the refinements coils have gone through, I would like to 
know what state of the art design would different people use if they had 
this power supply to use.
At work is a 20,000 volt at 30 amp DC continious, 480 volt 3 phase 
induction heater. Assume the induction oscillator section is disconnect, 
making the DC supply available.
Also available is a water cooled switching tube type Y676A/9013 a tetrode, 
designed for fast and high power switching.
It is rated at 30 Kv standoff and switching 150 amps. If the tube is placed 
in an atmosphere to precent arcing to terminals, it can switch 75 Kv. But 
we will call the switched voltage 20 Kv as that is what is available. The 
tube socket is also in hand. The Filament is 15.5 volts at 215 amps and 
that transformer is available.
If a cap bank across the 20 kv output of 40 Uf is in place, this should 
provide a decent supply of current peaks, so it should be pretty stiff.
Given the voltage, we can rule out solid state switching unless you have 
some real nice goodies in a series string with optical gating. This tube 
switch is used in electron accelerators.
The tube data, grid / plate switching plots are all available where I 
scanned them to at:
 www.hot-streamer.com/mike2004/
 Because this would be  3 Mw system wire size becomes an issue. CW would 
melt down but pulsed, this would be interesting.
So, 20 kV , 20 amps CCS, pulsed at 150 amps, what would you put it to?
Mike