# Is there a replacement for high voltage capacitors?

Does anybody know if there is a replacement for tuning capacitors in an LC circuit that can support several kV (let's say 4kV)?

The LC tuning circuit is used as a load for a power amplifier. C varies between 30pF and 1nF, L is fixed, and the frequency range is 1-5MHz. The load is driven with 10us pulses with a period of 10ms, with a max. power of 200W.

The current tuning circuit uses vacuum variable capacitors and step motors to tune the capacitors and I would like to remove any moving parts.

I've been looking at varicap diodes, but these are for small voltages.

• I think you are asking how to turn an automobile into a hovercraft. – Andy aka Apr 11 '17 at 7:46
• I think I agree with @Andyaka. Maybe you'd want to describe your whole application so that anyone has a chance to say "hey, you see, you're using your vacuum cap as {XYZ}, but you can do {XYZ} instead by doing {ABC}". – Marcus Müller Apr 11 '17 at 7:51
• @Andyaka I guess it is better to ask than to keep this to myself. – Cristian M Apr 11 '17 at 7:51
• It sounds like you're wanting to have a tunable final amp for a high-powered transmitter with 0 moving parts. that's a bold idea, but I think the solution is likely to get a bit more complex (and probably less elegant) than you're wanting. Let me know if that's right, and I'll put together something for an answer, but I won't guarantee you'll like it. – Robherc KV5ROB Apr 11 '17 at 7:52
• @RobhercKV5ROB the curse of competent engineering: "You can have two of {good performance}, {doesn't explode}, cheap, but never all three" :) – Marcus Müller Apr 11 '17 at 7:53

It would be possible to implement a "somewhat clunky" version of what you're wanting using digital control, solid-state HV switchgear, and multiple fixed-value HV caps...something similar to the illustration below (incomplete, but I think it conveys the intent).

simulate this circuit – Schematic created using CircuitLab
Now, the caps themselves aren't a problem, you can find all of them for under $10USD at you favorite mail/internet-order component supplier. HOWEVER solid-state switching of 4KV+ is an WNTIRELY different story! The transistor I labelled in the diagram will cost about$30/each for ordering quantities of 25+ (which you'll need), not to mention the diodes & the control circuitry (did I mention that all the MOSFETS I found had a Vgs MAX of +-20V?...you'll have to match the control signals to your 4KV signal frequency)

With that all said, there will most definitely be some unwanted noise/parasitics in the circuit from all those diodes & MOSFETS, any you may end up having to gate both ends of each cap (for better isolation of "switched off caps")...but it would be 4kV capable, solid-state, and (somewhat) adjustable between 5pF & 1nF+.

(P.S. the 1uH inductor in the illustration is just a placeholder for your fixed inductor in the amplifier tank)

• at freqs <10 MHz, and depending on what OP really needs for his research setup, one might get away with a single set of two FETs and a load that is switched using a L/capacitor-based phase shifter to emulate a variable cap; not sure about the feasibility of that. – Marcus Müller Apr 11 '17 at 9:39
• @MarcusMüller wouldn't that phase shifter still need a variable capacitance/inductance? ... thus still necessitating a similarly complex switching mechanism to above if I'm not missing something painfully obvious (which I sometimes do). – Robherc KV5ROB Apr 11 '17 at 9:43
• Use a resistive voltage divider to drive a lower-voltage variable cap? – Marcus Müller Apr 11 '17 at 9:47
• ...and then how to you get the low-voltage phase-shifted wave to effect the HV wave though? From my understanding of the question (rewrite), it's sounding more and more like he's either working an antenna tuner, or a dummy load; both of which require handling the entire 4kV/200W output of the amp (either altering it, or absorbing it, depending on tuner or dummy load)... Or are you saying to use another transistor to feed a HV copy of the phase-shifted signal back into the main circuit? – Robherc KV5ROB Apr 11 '17 at 9:58
• What losses in the C_drain_bulk of those high voltage transistors? – analogsystemsrf Apr 11 '17 at 16:27