Vasile Surducan wrote: > On Wed, 13 Aug 2003, Dave Tweed wrote: > > Also, your <1nF load draws over 300mA rms of reactive current at 25 kHz, > > which is going to need relatively large wire to keep losses low. > > I think you have mistaken with 10^3. Is just about 250 microamperes... No, 1/(2*pi*f*C), given f=25000 and C=1e-9, definitely gives 6366 ohms impedance. With 2KVrms applied, you're definitely going to have 314 mA rms flowing. Note that I'm talking about reactive current, not just the in-phase current that represents power delivered to the load. The coil has to carry both types of current with low loss. > > Yeah, you're going to be better off getting a broadband audio power > > amplifier and using a beefy audio transformer to step the voltage up. > > This is not working. One of my coleague has project and built such an > amplifier 5 years ago. Just in 0...10KHz but up to 10KV. The system is > totaly instable with capacitive loads. It works only with resistive loads. Have you tried adding feedback with an appropriate compensating network? > > An alternative approach would be to create a DC supply of 3KV (900 watts) > > and then build some sort of H-bridge driver out of high-voltage tubes. > > But controlling distortion with this setup might be a problem. > > This is not feasible at all. And you know well why :) Do I? Hey, I was just throwing out ideas, trying to be helpful, with the sketchy requirements you gave us. If you already know how to do this, why were you asking here? Seriously, if you already have coils that work over the entire frequency range, the obvious answer is to electromechanically switch among them, using relays or a motor-driven switch. How frequency-agile does the solution need to be? -- Dave Tweed -- http://www.piclist.com hint: To leave the PICList mailto:piclist-unsubscribe-request@mitvma.mit.edu