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

[Buk]

What are the chances of being able to generate a 100v peak to peak ?MHz sinewave @ ~5 milliamps from a single Li-ion cell?

The application is to drive a tiny PZT device.

 

[Beau Schwabe]

Yes, at What frequency?

 

[Tony Stewart]

Let's see 500 mW load , ok
100V/5 mA = 20 k ohm ok
100V/3V = 33:1 voltage boost ok
Out/ in impedance ratio = 33^2 = 1k thus Zin = 20 Ohms ok
Self resonant frequency of a step up transformer at x MHz with a Q >3 ok
- flyback would likely be too fast



Yes I think you can do it easily with a small gapped core. . 3.6V logic 74ALVCxx is about 25 ohms. You can gang a hex inverter in parallel to get 4 Ohms or use a proper half-bridge for a transformer 10 uH primary interleaved and segmented secondary windings.

wait a minute 20 kohm at 1MHz is only 5 pf
how can it only be 5 mA?
It must be really tiny.

 

[Buk]

Yes, at What frequency?

Unclear. It depends what size PZT elements I can find; but I think in the 1 to 10 MHz range.

 

[Buk]

wait a minute 20 kom at 1MHz is only 5 pf
how can it only be 5 mA?
It must be really tiny.

I'm way out of my comfort zone here. By "It" above, you mean the PZT? If so, my only answer is 'maybe'.

I'm working from a paper where they describe using a PZT element to cauterize tissue (my application is quite different, but uses PZT to generate heat); but it is somewhat confusing. In one part of the paper they show this diagram:

Here, very clearly the PZT elements are 3.2mm round and 190µm thick.

But in another part of the paper they have this diagram where the PZT element(s) are embedded into the side wall of a hyperdermic needle:

If that needle is large enough to embed a 3.2mm diameter PZT in its sidewall; I'd drop-kick any doctor that approached me with it in his hands. Geez!.

Update: To get the proportions of the diagram above the needle would need to be 10mm in diameter:

Into your thyroid?

In another similar paper, they show this image:

And this diagram:

The recess in which the 4 PZT elements are embedded is described as being 2000 x 300 x 135 µm in the sidewall of a 20 gauge needle. In this paper they describe the drive circuitry:

And frequencies that range from 700kHz to 11.5MHz are mentioned at various points.

My antisipated constuction I hope will look something like this:

The 4 yellow disks (rings) are shown as 5mmOD 1mmID and 0.25mm thick. But that's entirely guesswork. I'm in contact with a Chinese PZT manufacturer and they are looking through their inventory to see what they have available 'off the shelf'. I cannot afford tooling for a custom part; expecially as its a possibly half-baked idea I am pursuing.

My (extremely limited) understanding of PZT is that the resonant frequency is inversely proportional to the size; but that needs to be qualified by the orientation of the excitation.

As I said. Waaaay out of my sphere of knowledge on this.

 

[Nigel Goodwin]

But in another part of the paper they have this diagram where the PZT element(s) are embedded into the side wall of a hyperdermic needle:
View attachment 133834
If that needle is large enough to embed a 3.2mm diameter PZT in its sidewall; I'd drop-kick any doctor that approached me with it in his hands. Geez!.

Cardiac ablation is normally done via catheters inserted in both wrists, and positioned in the heart - one is for monitoring, the other does the ablation. So obviously it has to be small enough to be inserted in a vein, and pushed through to the heart. This is essentially the same technique as an angiogram or the fitting of stents.

However, I suspect it's not as small as you might hope for

Even assuming it's 3.2mm for the PZT above, I don't see as that's an issue as it's there to try and save your life! - and they would inject local anaesthetic first, and perhaps even cut a 'hole' for it. I imagine there's probably a lot of procedures that use that size, and larger, needles.

I knew RF heating was used, but can't say I knew it was via a PZT - or had even thought about it.

 

[Buk]

...Y

...

My response to you questions was so stupid there's nothing left to say?

 

[Beau Schwabe]

Before you mentioned any details in my mind , here is what I thought.... a single Li-ion cell is generally going to have plenty of current, so step it up with an adjustable voltage regulator 7V should be about where you want it but design it so you can go to 12V. Pick your frequency and setup an LC tank resonant to that frequency. In my test case I used a 22pF cap and a 10uH inductor for a resonant frequency of about 10.7MHz. Next your going to need a micro that can generate a 10.7MHz SQUARE wave and drive the LC tank with the output voltage from the regulator. If the LC tank has a 20k load on it then you should be able to 'dial in' between the regulator and the 10.7MHz square wave a Sine wave output that is 100V at 5mA. Attached is a pseudo schematic ....

 

[Buk]

Attached is a pseudo schematic

Thank you for that.

One (of possibly many) questions. Any particular reason for using a microcontroller to generate the square wave rather than VCO (say SN74LS624)?

If the project goes anywhere, it will need a microcontroller, but doing anything else on a µCPU whilst maintaining a high frequency loop gets messy.

 

[Nigel Goodwin]

Thank you for that.

One (of possibly many) questions. Any particular reason for using a microcontroller to generate the square wave rather than VCO (say SN74LS624)?

If the project goes anywhere, it will need a microcontroller, but doing anything else on a µCPU whilst maintaining a high frequency loop gets messy.

There are plenty of hardware peripherals in modern PICs that will avoid any 'messy' code, and essentially leave the processor completely free to do other things.

Try looking up the 'NCO' (Numerically Controlled Oscillator).

 

[Buk]

3.6V logic 74ALCxx

'scuse my ignorance of these things, but typing 74ALC into mouser finds nothing, and I have no idea what I am looking for. Hints?

 

[danadak]

http://www.analog.com/media/en/technical-documentation/product-selector-card/2pb_6015f.pdf

http://www.ti.com/product/OPA454

http://www.eevblog.com/forum/projec...-source-(emitter)-amplifier-on-opamp-outputs/

Some possibilities....

But you need a slew rate of ~ 600+ V/us for a 100V sine. Maybe this is a better starting point -

http://www.analog.com/media/en/technical-documentation/data-sheets/AD815.pdf

Slew Rate Calculator

This calculator computes the minimum slew rate necessary to output a certain voltage at a certain frequency, calculates the maximum output frequency an op amp can deal with at a given voltage, and calculates the maximum voltage that an op amp can output at a certain frequency.

www.learningaboutelectronics.com

Regards, Dana.

 

[Buk]

Some possibilities....

Thanks Dana. Your links are most helpful.

 

[Tony Stewart]

'scuse my ignorance of these things, but typing 74ALC into mouser finds nothing, and I have no idea what I am looking for. Hints?

/I had a dyslexic moment

ALVC

this is 3.6V logic used by ARM logic at 3.3V -10% it is 22 Ohms max for RdsOn on complementary MOSFET switches. Typically it is lower.
Often specified at common supply voltages with -10% tolerance

Whereas older 5.5V logic is a bit more than twice this resistance 50 Ohms +/25%

 

[Tony Stewart]

There is no need for any uC to do this. But you will need to understand series resonance, fo and Q with low impedance driver and RC piezo equivalent impedance cct that will affect resonant f and , Q and Vout.

Here I simulated using 5V logic (50 Ohms) with compl. buffer . ~ 1 Ohm ( as Beau has also done) and same LC values, except C must include the PZT C value. Tuning both the Relaxation RC Schmitt Trigger Oscillator and LC values, I used a large C in series with a 4pF to fine tune the high Q resonance C effective sum load.

(mouseover wheel tunes values) ^Z undo ^Y redo

Similarily I arbitrarily chose RC values, so I could fine tune f with the smaller R.

The output voltage in theory with ideal inductors might be 150V rms with no load. (if you pull off the 22k PZT load) in theory my emitter followers pushed over 300mA rms to create > 600Vpp) ! The model is only as good as the assumptions, and the PZT impedance is not simply 22k. It is a combination of RLCsCp

If you don't understand Q resonance, I can link to an RLC nomograph. or PZT impedance, look up crystal impedance.

Computing / building the piezo PZT RLC values is essential before doing anything.


Good specs make a better design.

 

[Tony Stewart]

even the contact with skin will add pF and so the design will keep changing .... that is until you establish the design specs. !!
Consider HBM for ESD fingertip area is about ~ 100 pf. so 5% area of fingertip might be ~ 5pF

 

[Buk]

If you don't understand Q resonance, I can link to an RLC nomograph. or PZT impedance, look up crystal impedance.

I also did a sim of of your circuit above and wrote a short script to calculate a range of RC values from an input frequency 100kHz, 1MHz, 10MHz, (trying to cover all bases) they produce voltages close enough to 100V for those frequencies. Update: And it just dawned on me I got 200V p-p!)

I found a reference to an equivalent circuit on the reference pages of one of the two manufacturers I've been in contact with:

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

Computing / building the piezo PZT RLC values is essential before doing anything.

even the contact with skin will add pF and so the design will keep changing .... that is until you establish the design specs. !!

Indeed. I've been trying to establish the right combination of thickness, no. of elements, polarisation, but data on this use of piezo is thin on the ground and what I have found in the literature is mostly behind expensive firewalls. I've emailed a couple of authors, often they'll send a copy.

I'm going nowhere fast until I get some help selecting the right piezo elements for the job.

 

[Tony Stewart]

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

 

[rjenkinsgb]

This appears to do exactly what you want?

Dynamic Product Page | Microchip Technology

 

[Tony Stewart]

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 The symmetry of duty cycle may not matter in this case but is measured by the 2nd harmonic being present and attenuated.