Circuit to find out the resonance frequency of a piezo element

[ewinek]

Hi all,

At the moment I am working in a project making use of a piezoelectric pump. The resonance frequency of the piezo element is about 22 kHz, I have a basic driver to do the job but the resonance frequency changes depending on the load of the pump, the temperature, the air pressure and, obviously, from device to device. The device operates fairly well even though I use a passive circuit driver, which is not hitting the resonance frequency. However the aim is to have the best possible performance and reliability, which might be compromised over time if the device is not used at its resonance frequency.

At the moment I have an alternative active solution using a microcontroller, tracking the maximum current consumption (impedance is the lowest at resonant frequency) and therefore sweeping the frequency continuously to find the sweet spot. I would like to get rid off the microcontroller and use a passive solution to match the right frequency.

Any ideas? In the attachment you can find the recommended driver from the manufacturer of the piezo pump (Murata).

[nfmax]

That circuit looks like it’s doing what I have done in the past, and was about to suggest: create an amplifier with a negative output resistance. It’s sensing the current taken by the piezo, amplifying it, and feeding it back with positive polarity. The whole circuit will oscillate at the resonant frequency of the transducer, tracking any changes e.g. with temperature. Much the simplest way!

[the Chris]

Hi.

I once worked on a drive system based on a piezo in resonance. In most applications, the resonance frequency was detected just like you discribed by scanning the current of a frequency sweep for lowest impedance.

However, there was also a pure analogue frequency detection circuit available. This is about 15 years ago, so my memory is very vague. But the basic principle of function was that the voltage drop over the current shunt (or more precisely, it's inverse) was used as an input into a VCO that drove the piezo over a MOSFET. I can't remember if the input got any additional treatment, but it worked fine for compensating the slight changes of heat up dependent frequency drift. One just had to limit the driving time as the circuit would otherwise adjust frequency during permanent operation until heat damaged the polarization of the piezo.

Maybe this helps as a starting point.

[Conrad Hoffman]

You might also consider putting an inductor in parallel with the PZT to knock out the bulk capacitance. At resonance the system will look nearly resistive and be an easier load. Probably easier to sense as well. I think I've used Micrometals cores for this in the past, which can be better than ferrite.

[Marco]

https://www.google.com/search?q=pll+piezo+driver
https://github.com/leacog/PLL-ultrasonic-driver

A simple positive feedback driver won't necessarily lock the right resonance.

[moffy]

When working on laser gyros which have a piezo driven oscillating wheel at the center, a simple phase shift network with a couple of opamps was used to detect and provide the resonant frequency. The phase changes abruptly around resonance so choosing the right phase shift will ensure you track the resonant frequency.Afraid I can't remember the configuration exactly.

P.S. We might have used a piezo element on the oscillating wheel as a motion sensor.

[coppercone2]

is this a phase lock loop problem?

[ewinek]

Thank you all for the contributions. I will explore the proposed alternatives and come back with the results

[nfmax]

https://www.google.com/search?q=pll+piezo+driver
https://github.com/leacog/PLL-ultrasonic-driver

A simple positive feedback driver won't necessarily lock the right resonance.

Indeed. Note that U2 in the Murata drive circuit is configured as a band pass filter, presumably to persuade the transducer to oscillate at the intended resonance.