Driving piezo actuator

[behoolder]

Hello!

I'm planning to create some simple tip/tilt mirror using three piezo actuators from Thorlabs.
This should in principle be a realisation of design Thorlabs gives as application idea: https://www.thorlabs.com/Images/GuideImages/7563_Piezo_Stacks_9.jpg
As an educated guess I choose this one:
https://www.thorlabs.com/thorproduct.cfm?partnumber=PA3CKW

With parameters:

Displacement (Free Strokeb) @ 100 V   3.0 µm ± 15%
Load for Maximum Displacementc   65 N (15 lbs)
Resonant Frequencyd   415 kHz (No Load)
Impedance at Resonant Frequencyd   600 mOhm
Anti-Resonant Frequencyd   510 kHz
Capacitance   65 nF ± 15%


So the idea is to glue them with, say, epoxy, to a mirror in this triangle fashion, create hw & sw to move it and put it inside imaging setup, say: a telescope. Few pixels of movements in image plane should be sufficient so I do not care for large movement stroke, but I want to update mirror position with at least 100Hz frequency, like having it maximally pointing left and 10ms later pointing max right. Precision is not very important, I guess positioning with +/-10% accuracy will be fine for a start.


Now as we already got here I am aware that maybe XY problem... But I think the question itself remains valid:
Is there any easy solution to drive such a piezo actuator with such requirements?
By easy solution I mean either:
- a ready to buy module (like for steppers you have i.e. DRV8825)
or
- an easy to solder by myself set of elements (like lm298 + some resistors)

So far I've found this:
https://www.tindie.com/products/Fyberlabs/piezo-haptic-flex-module/
But I cannot see DRV2667 to be available to buy anymore, also I am not exactly sure if haptic driver like this would be good for actuator like I want

I can also see something like this:
https://pl.farnell.com/maxim-integrated-analog-devices/max77501ewv/boost-driver-2-8-to-5-5v-wlp-30/dp/3941815?st=haptic%20driver

But sorry, 0.4mm balls is much too much for my skill: if it has no leads I cannot solder it
Also it doesn't have analog way from input to output which I'd prefer over playing waveforms. (but maybe it's just prejudice).

But maybe all I need is something entirely different?

What I've found looking for answers here is this thread:
https://www.eevblog.com/forum/projects/need-schematic-to-build-a-piezo-driver-0-150v/msg1388451/#msg1388451
But high current (15A???) requirement seems inadequate for my needs. Anyway - maybe all I need is full bridge mosfet driver?

I've also read this:
https://www.eevblog.com/forum/projects/_high-voltage_-linear-piezo-driver/msg916378/#msg916378
But I am not sure also - i do not need big bulky thing, I thought it can be done by something smaller?



Sorry for being so unsure about anything: it seems like I have most of pieces somewhere in my brain, but I think I need some external push in the right direction, so please help me!



======================
And by the way:

This is my first post here so please allow me to shortly introduce:
I am a C++ developer, having just a little experience with embedded stuff.
What I can do is to write a I2c/SPI driver for Arduino for some simple IC, provided with datasheet.
Provided with an oscilloscope I can even check if it works.
I can solder THT elements if necessary.

As a hobby I do astrophotography, but instead of getting actual pictures I mostly do some diy stuff about it: steppers control, optical encoders, control & analyzing software.

[ledtester]

but I want to update mirror position with at least 100Hz frequency, like having it maximally pointing left and 10ms later pointing max right. ...

I watched this crash course in piezo actuators:

Piezoelectric Actuator Testing: Displacement vs Voltage -- Photonymous Labs
https://youtu.be/vKC99y3fZjA

Would a high-voltage triangle waveform generator work for you? Like it would go from 0V to 100V and back to 0V and then repeat?

Concerning TI chips... I found the following chips in stock at TI:

- DRV8662
- DRV2700
- DRV2665

(To see TI's inventory levels just create a free ti.com account.)

All of these are QFN20 chips. Options for using QFN chips are:

- use a QFN socket
- apply solder paste with a stencil and reflow
- using a hot air soldering gun
- buy an evaluation module

Sockets which breakout a QFN package to individual pins are available but also kinda expensive -- like $50-$100 a piece.

Here's a good video on hand soldering a QFN package -- also discusses the solder paste/reflow  procedure:

https://store.curiousinventor.com/guides/Surface_Mount_Soldering/QFN

The eval module for the DRV8662 is $99 and apparently is in stock at TI.

This video uses the DRV8662 eval module:

Haptic Feedback with Piezo Buzzers! -- Carl Bugeja
https://youtu.be/egLA0GgF3xw

In experimenting with it the guy managed to blow up the DRV8662 and replaced it using a hot air gun.

[WannesS]

The TI DRV2667 is ideal for your purpose. 65nF is within spec, especially at only 100Hz, it can probably do a bit more.
There's a excel file "DRV8662, DRV2665, and DRV2667 Design Equations (Rev. A)" to validate your parameters.

There are 4 parts in the same family:

• DRV8662 & DRV2700 have only an analog input
• DRV2665 & DRV2667 have i²c and analog input

All of them fit your requirements.

The Fiberlabs Piezo Module you found on Tindie would be ideal as a breakout board, but it is still out of stock at the moment.
The TI piezo drivers were out of stock for a long time, but are available again since mid 2022, and are now reasonably priced and in stock on digikey and mouser.
I suggest you message the seller and ask if they could build another batch.

The next easiest solution is a TI evaluation module. The cheapest one in stock is DRV8662EVM. TI will only sell you 1 for $99, mouser has 11 in stock for €130

Next best option would be to design and build your own board. The QFN20 is somewhat hand solderable with a fine tip iron, but I would recommend to use solder paste with a 0.1mm stencil.



As for the piezo actuators, using epoxy is the best way to attach them. They have a maximum curie temperature, so if you can get actuators with pre soldered wires, and don't use a heat cure epoxy.

They don't expand/contract linearly, you'll have to compensate for that in software.