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

rjenkinsgb said:

This appears to do exactly what you want?

Dynamic Product Page | Microchip Technology

Click to expand...

maybe . This is the classic Class B FET amplifier with a 100V power supply , great for 10W of ultrasound power., but here I think it needs only 1/4w in a tiny 603 SMD PZT to get over 80'C or ablation energy level. This technology is at least 20 yrs old. https://www.google.com/search?q=SMD...hUKEwjK863RrZ3zAhUyTd8KHaliBYsQ4dUDCA0&uact=5

Given more details, time and motivation, I could find a list of suppliers for parts. http://en.china-yec.com/product/73/

Buk said:

View attachment 133898
And a bunch of "useful equations" on another.
Hopefully I'll get some further feedback once they are back to work tomorrow...

This the standard Crystal model and almost any resonator. Falstad has this model in >Draw>passive>crystal

Click to expand...

most experts will be surprised to read that your garden variety Xtal for a uC that runs on 3.3V or 5V or whatever actually has a Q of 10k and about 30kV inside the crystal lattice which is why they rate the max power to 10 to 100uW. Due to contaminant internal lattice voltage breakdown.

Whereas the PZT is more power but lower Q.

rjenkinsgb said:

This appears to do exactly what you want?
Dynamic Product Page | Microchip Technology

Click to expand...

I think it would, except that it appears to require +&- 10V, and I'm hoping to power this from 3.7V Li-ion battery.

Tony Stewart said:

Given more details, time and motivation, I could find a list of suppliers for parts. http://en.china-yec.com/product/73/

Click to expand...

You're waaaay ahead of me. I'm still trying to digest the post before the post before the post before this post. And my head is already swimming today.

Tony Stewart said:

Ideally, what you would do is combine the PZT in some type of Colpitts series resonant oscillator using an attenuated voltage for feedback in order to get the voltage gain you want, always at resonant frequency with C in series with PZT for low impedance , more motional energy, not parallel which is high impedance.

Click to expand...

I have vague recollections of the Colpitts oscillator from about 30 years ago; but this...

Tony Stewart said:

Here by a clever little change in parts, I boosted the PZT load to 250 mA pk reactive with 88V rms and made the Astable an injection locked loop (ILL) with PZT feedback by a tapped inductor to get some of the feedback and grow to maximum amplitude after reset. http://tinyurl.com/yht2st9w

Click to expand...

goes Whoosh. Right over my head at the moment. I'll try to catch up.

(Thank you for your extraordinary input.)

Tony Stewart said:

Ideally, what you would do is combine the PZT in some type of Colpitts series resonant oscillator using an attenuated voltage for feedback in order to get the voltage gain you want, always at resonant frequency with C in series with PZT for low impedance , more motional energy, not parallel which is high impedance. https://www.maximintegrated.com/en/design/technical-documents/tutorials/5/5265.html

Click to expand...

Okay. I think I get this. Use the capacitance of the PZT in place of one of the two capacitors in the Colpitts oscillator, thus using its capacitance rather than having to work around it?

Tony Stewart said:

Here by a clever little change in parts, I boosted the PZT load to 250 mA pk reactive with 88V rms and made the Astable an injection locked loop (ILL) with PZT feedback by a tapped inductor to get some of the feedback and grow to maximum amplitude after reset. https://tinyurl.com/yht2st9w

Click to expand...

If I understand this correctly, you've done away with the need for the square wave generator; by using the feedback loop and schmitt trigger to perform the switching.

The downside from my perspective is that I will almost certainly need to try several different piezo elements each requiring different frequencies before I hit in the right combination; and whilst I understood the math controlling the earlier RC circuit, I suspect this one is a lot more complicated?

Not quite. This is a trivial injection locked loop. One is tuned by {k/RC}, the other tuned by {LC}^-2

These are two separate oscillator loops that become one frequency as the tapped inductor gets a small part of the low impedance driving reactive load so that when both operate near the same frequency it naturally seeks max gain. The L and PZT C act as a high Q (high voltage gain amplifier.) The real energy depends on coupling factor to tissue and may be limited with a fixed series R. You choose the piezo for the properties you need and the Astable is simple to adjust with coarse and fine tuning. There are refinements that can be made. This was a 10 to 30 minute design. Have you learned how to compute Q? If the tissue starts to conduct, the amplitude of voltage drops quickly. Rather than the fixed 100V Class B amp does not change in voltage unless you add a large series R.

Tony Stewart said:

coupling factor to tissue

Click to expand...

Tony Stewart said:

If the tissue

Click to expand...

Now I'm confused. There is no 'tissue' involved in my project. The inspiration for using peizo to generate heat was a paper about quarterizing tissue; but my project only borrows the idea of using of piezo to generate heat. Nothing more.

Sorry, if I have misled you.

Tony Stewart said:

a high Q (high voltage gain amplifier.)

Click to expand...

I thought Q was quality factor -- the ratio of frequency to bandwidth. Having trouble finding references to Q in the context of amplifier gain?

Yes it is also the impedance gain ratio at fo. Q=X(fo)/Rs for series and Rs/X(fo) for parallel e.g. gain of 30 dB https://www.falstad.com/afilter/circuitjs.html?cct=$+0+0.000005+5+50+5+50 %+0+65868806.91851612 l+208+112+256+112+0+0.000009999999999999999+0.1324343184476945 c+256+112+256+144+0+2.1999999999999998e-11+0 r+320+112+320+144+0+22000 g+256+144+256+160+0 g+320+144+320+160+0 170+192+112+160+112+2+20+4000+5+0.1 O+320+112+368+112+0 r+208+112+192+112+0+1 w+256+112+320+112+0

But if you just want heat, use an SMD resistor

Tony Stewart said:

But if you just want heat, use an SMD resistor

Click to expand...

Its a little more than just that. A quote from another paper:

Cavitation bubble collapse is a remarkable
phenomenon induced throughout the liquid by the power of sound. In aqueous systems at an
ultrasonic frequency of 20kHz each cavitation bubble collapse acts as a localised "hotspot"
generating temperatures of about 4,000 K and pressures in excess of 1000 atmospheres
(Neppiras 1984, Henglein 1987, Suslick 1990).

Click to expand...

Another paper: https://www.sciencedirect.com/scien...ba14d76&pid=1-s2.0-S1350417797000199-main.pdf

As an ex submariner bad news when props cavitate. "Hey world here is where I am ....."

Regards, Dana.

Ano

danadak said:

As an ex submariner bad news when props cavitate. "Hey world here is where I am ....."

Click to expand...

Even for those prop users who lives don't depend upon their invisibility, cavitation is bad news:

And it can be seriously bad news for pumps also:


But in that destructive power, there is also potential.

Your question now seems unrelated. It would take much more than 500 mW.

Tony Stewart said:

It would take much more than 500 mW.

Click to expand...

It depends what you mean by "it"?

Tony Stewart said:

Your question now seems unrelated. It would take much more than 500 mW.

Click to expand...

I really wasn't being cute with my question. I don't know what it is that you think I want to achieve that will take so much power.

Cavitation creates such extreme temperatures and pressures because a vapour bubble collapses from ~50 microns diameter to less than 5 microns in a few 10s or 100s of picoseconds. That 1000 x reduction in volume in such a short timescale involves extreme accelerations which are responsible for the extremes of temperature and pressure (and even the production of light); but the total amount of energy involved is miniscule.


The amount of liquid (say water) in (say) a 1mm length of 1mm OD (0.5mmID) tube is ~0.2ml. To heat it all from ambient to boiling would take 0.0186 WH or 67Ws. 1W for 67s or 67W for 1s. Given total efficiency, at 100v, 0.67A would provide enough power.

Of course nothing 100% efficient; but then what if the goal is not to boil the entire 0.2ml, but rather just a small percentage of it?

And then there's the matter of what happens if the acoustic energy is focused at a point or line in the middle of the tube.

Buk said:

1000 x reduction in volume

Click to expand...

you mean 10:1 reduction from 50:5.

I don't know exactly what outcome you want and then to define the piezo specs and input you need. But your question started from the input without defining the output.

Is it just a piezoelectric mister ?

Cavitation seems to be the phase-change avalanche effect analogy to dielectric breakdown of Partial Discharge that eventually destroys oil-filled transformers by creating H2 and other HC combustible gases, due to contaminants in the dielectric(=insulation) , or voids in dry types.

Tony Stewart said:

you mean 10:1 reduction from 50:5.

Click to expand...

No. Volume. 4/3*25³*pi = 65450 microns³; 4/3*2.5³*pi = 65.45 microns³ 1000:1.

One of the papers I read mentioned 10^6:1. I think 300µm to 3µm. I'll try to re-find the reference.

Tony Stewart said:

I don't know exactly what outcome you want and then to define the piezo specs and input you need. But your question started from the input without defining the output.

Click to expand...

I did identify very early on that I wasn't at all sure about the parameters. You moved so fast with expanding the electronics side of this -- to which I am still trying to catch up with even the basic understanding -- that I have been neglecting the electro-mechanical side.

It doesn't help that as soon as the PZT manufacturers hear that there is no company, no multi-billion/year target market, and no big order in the immediate offing, they stop responding. I'm probably going to end up buying a few different size PZT elements off alibaba and sucking to see. Which means I'll need flexibility in the electronics to vary the frequency and output impedance widely.

Tony Stewart said:

Is it just a piezoelectric mister ?

Click to expand...

No. In fact there is no 'it'. There are a bunch of possible 'its', with one I'm particularly interested in but don't want to discuss.

By way of example, without wishing to turn this thread into a discussion about this either, one of the problems with ICE engines is incomplete combustion producing noxious substances. Some injectors already use piezo elements to inject the ~0.1ml of fuel per cyclinder per cycle.

Imagine if between the closing of the injector valve at the end of one cycle, and its opening 2 cycles later those same piezo elements were driven with a high frequency low voltage signal so the valve remained closed, but the ultrasound induced cavitation within the closed cylinder. Partial vaporisation within the injector raises the temperature and pressure, so that when the valve opens, the fuel instantly vapourises as it escapes into the air flow. Better mixing? Better combustion?

Tony Stewart said:

Cavitation seems to be the phase-change avalanche effect analogy to dielectric breakdown of Partial Discharge that eventually destroys oil-filled transformers by creating H2 and other HC combustible gases, due to contaminants in the dielectric(=insulation) , or voids in dry types.

Click to expand...

Mostly over my head, and far outside my experience, but it sounds about right.

FWIW, I met the guy who used H2O preheated spray with a vortex intake on the carburettor to get 100 MPG on his truck using a PIC to monitor ECG and intake from decades ago.

Partial Discharge effect is when the smallest bubble or contaminant in oil or plastic insulation (air void) for high voltage cables or transformers is effectively the smallest capacitance in series with the bulk and thus charges up the fastest and reaches a breakdown in that tiny gap, causing detonation of the material at the nano level not yet shorting across the grid. But for fuel or any Hydrocarbon, detonation of the lowest level releases Hydrogen, then higher levels of activation with more energetic gas. To prevent breakdown catastrophe in transformer oil , they perform DGA tests on sample oil to look for higher levels of combustible gas, methane, ethanol, acetylene ...

10 ppm is good , 1000 ppm is a concern, 1% is near danger


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.