sstc changing frequency issue

[Jeg]

I recall this in case someone knows the way to improve the power of the above circuit. In the meanwhile I reduced the power supply to 48v and it works in a safety voltage level. One more question is why frequency changes according to the power supply level? Does it have to do with the current that loads the mosfet's gate capacitance? More voltage more current faster switching capability, higher resonant frequency?

Tnks
Jeg

[TinselKoala]

I've found this kind of circuit to be very sensitive to a number of things. Coupling between primary and secondary is very important, on my coil the primary has to be at just the right height and spaced just right, too tight or too loose will really affect the power throughput. The positioning of the pickup-feedback coil is critical too, mine works best when this coil is actually underneath the primary-secondary arrangement, and I use just a single turn for this pickup. Also the mosfet is critical, probably because of the capacitance. I finally settled on the same IRFP460 and found that the IRFP460N variant would not work as well. The circuit winds up driving the mosfet gate with a swinging AC signal and the gate protection zeners don't always work, as Jeg has found out. I've put an analog ammeter in my circuit and I know never to allow the current in that part of the circuit to go over six amps or the mosfet will do what Jeg has described.
Looking at the secondary's HV field output by using a scope probe connected to a bit of aluminum foil some distance away from the coil, and also the mosfet drain signal, can be very diagnostic. When the circuit is running properly the mosfet drain will be a series of sharp fast spikes and the field pickup from the secondary will show a beautiful and clean sine wave.
There is something about the "4 Mhz" frequency that this circuit likes. Push it in either direction and you get different performance. Higher frequency and the output is more like a radio transmitter (duh) and you get less corona and sparks and stuff but you can get good effects lighting external CFLs and such. Lower frequency and you need larger coils, more control over coupling, more terminal capacitance, and again, the character of the corona display changes.  There are better designs for lower-frequency SSTCs than this one, I think.

However this design may work really well for a "plasma speaker". I got amazing fidelity at very low volume from an audio-modulated "Jacob's Ladder" ; excellent speaker effect even when the HV output was so low that no corona was visible.

[Jeg]

...When the circuit is running properly the mosfet drain will be a series of sharp fast spikes...

If you set correctly the two coils at primary's side in means of resonance ala Don Smith (wavelength related), there are no Spikes anymore. Just a clean sinewave.

[TinselKoala]

If you set correctly the two coils at primary's side in means of resonance ala Don Smith (wavelength related), there are no Spikes anymore. Just a clean sinewave.

Maybe that's why you are blowing mosfets and not getting much output, whereas I don't blow the mosfet (any more, knock on wood) , it switches cleanly at a low voltage. To get good output from a secondary, any secondary, the rate of change of the current in the primary has to be as fast as possible. Driving with a sine wave is very _inefficient_ in terms of making HV output and it also stresses the mosfet more than the right kind of spikes.

In the image below, the top trace is the Drain trace at 10V/div  and the bottom trace is the pickup from a piece of foil a meter away from the secondary, displayed at 50 V/div. Input was 22 volts, current not recorded.

[Jeg]

This is mine. Check the time between peaks...
The time diff between the dotted lines is 25.5ns, but if you notice they are not centered. The actual diff between peaks is 18.5ns