Generate a 100v peak to peak sinewave (at a few milliamps) from a 3.7 v DC source

[Buk]

When PD occurs repetitively and is visible on the outside in air we call it corona , that's the same thing as PD in oil and cables used for high voltage.

Ah yes. I seem to recall seeing a video of a grid transformer undergoing coronal disharge at night, but I can't find it.
The best I saw was this:

 

[Tony Stewart]

Just as ionization from an excessive voltage across a conductor gap takes some time ~ microsecond, I believe this is the bandwidth of the medium where risetime of the envelope is inverse to the BW at some resonant frequency where the ratio of fo/BW-3dB and the gain out/in are both = Q, so likely your cavitation response time has a similar response behavior. i.e. step function delay of energy excitation to cavitation instant.

 

[Buk]

the ratio of fo/BW-3dB and the gain out/in are both = Q,

I've read a bunch of stuff on this, but just reading doesn't give you any feel or intuition for the subject. I'm back to playing with falstad and some simple amp circuits to try bolster my understanding.

One thought that crossed my mind -- please be gentle if it is completely cockeyed -- could I use a transformer to isolate the oscillator from the capacitance of the PZT and would that allow me to have a variable frequency oscillitor with a wider range -- I think I need to range from 20kHz upto 2 or 4MHz -- without needing to completely re-tune the circuit for difference PZT elements?

 

[Tony Stewart]

Like any Crystal or MEMs or tuning fork or PZT oscillator, there will be a series in phase resonance, parallel 180deg phase and possible harmonics.

You can sweep any ceramic capacitor with rated voltage and hear piezo effects from a low impedance audio amplifier or half bridge buffer. . You can do this cheap and dirty with 6x 74HC hex inverters in parallel. or use the fancy SMD bridge drive from #19 or similar as total ESR and wire ESL , R determines the maximum current at resonance.

 

[Buk]

Like any Crystal or MEMs or tuning fork or PZT oscillator, there will be a series in phase resonance, parallel 180deg phase and possible harmonics.

You can sweep any ceramic capacitor with rated voltage and hear piezo effects from a low impedance audio amplifier or half bridge buffer. . You can do this cheap and dirty with 6x 74HC hex inverters in parallel. or use the fancy SMD bridge drive from #19 or similar as total ESR and wire ESL , R determines the maximum current at resonance.

Was that a yes or no

 

[Tony Stewart]

Maybe. You could use a low impedance switcher drive an autotransformer that steps up the voltage and power level for a Piezo load, but resonant voltage is fo=1/(2pi*sqrt(LC)) so the transformer inductance will alter the characteristics unless you choose the impedance needed.

But for constant voltage one uses a bridge with low RdsOn switches (like CMOS logic but lower R) or a feedback amplifier with very high GBW)