How to build a constant current source using vacuum tubes

[parsec]

I need to discharge a lot of current into an inductive load from a large capacitive bank. Ideally the current waveform would be as close to a flat-top as possible.

Typically this seems to be done either by switching many caps in sequence using SCRs into a filter network to remove ripple or by simply using many caps to increase the power supply's time constant well beyond the required discharge time.

I have toyed with the idea of using some sort of solid state regulator (e.g. zener based) but typical junction devices can't handle the currents I need (kiloamps).

I know that large vacuum tubes are still manufactured for the radio industry etc. Would it be difficult to make a regulator or constant current source using a valve and some clever feedback? I have no idea where to start designing such a thing.

 

[vk6kro]

Transmitting valves are quite high impedance devices.
Typically, they require very high voltages on their anodes to work properly and then they draw relatively modest currents.

See this one:
http://www.uksmg.org/content/russia.htm

They use a voltage of 2000 volts or so and only get 0.4 to 0.6 amps flowing.
Larger devices use as much or more voltage.

I have seen advertisements for huge SCRs that will control hundreds of amps.

 

[parsec]

Ah right, I figured these valves could handle large currents being so large.

I'm not sure I could use SCRs to throttle kiloamps in any sort of practical (i.e. cheap) way...

Perhaps I could use a discharge of some sort to regulate the current?