How to control the driving frequency when working with high power ultrasonic transducers

Question

I would like to know the simplest solution to dynamically tune the driving frequency of an ultrasonic transducer to its resonance frequency.

My amplifier has a power meter that I can read an optimize via LabView but I would like to know if I can do the same using a multimeter and a controller that changes the frequency of the function generator. Can this idea be applied using a multimeter and a controller?

I have the following setup and in this question there is more information about the problem:

Answer 1

It all depends on how you want to drive the transducer i.e. the application: -

It appears that you can drive at resonance or anti-resonance which indeed does make it pretty similar to how you would use a crystal in an oscillator. The graph above is taken from this interesting website.

I can't determine from your question what application you have but, from the link (in the other question) to the type of transducers you use it seems you will be series resonating the transducer and this means it has low impedance at resonance due to L and C being in series. This means that the type of control circuit will look like this: -

Taken from here and this site also has some very useful information and an ebay link to a cheap one: -

But, if you are still intent on building your own you can use the series resistance method to generate a feedback signal to the front-end of a power amplifier. Clearly the series resistance need only be about 1 ohm to prevent excessive power losses. The signal will be maximum at series resonance and importantly in-phase with the drive voltage to the transducer. This means a simple power amp will do the job but, with a method of controlling amplitude.

Amplitude needs to be controlled or the PA will go into saturation and it may damage the transducer. It's a bit like a Wein-bridge oscillator needing amplitude control to ensure sinewave purity. The fed back signal could be adjusted with a pot but, given the Q of the transducer, this is probably best achieved using a JFET: -

Regards the PA itself, make sure that the phase angle between output and input is small at resonance or the transducer will not run quite at perfect resonance. This is usually done by ensuring the PA has at least 10x the bandwidth of the running frequency.

Answer 2

An idea once I had, but never tried, is to treat an ultrasonic transducer like a high-power crystal oscillator.
Your typical crystal oscillator circuit looks like this:

^{simulate this circuit – Schematic created using CircuitLab}

Simplistically, the crystal (along with the 2 capacitors) provides a 180 degree phase shift at its resonant frequency and this determines the output frequency of the oscillator.

So why not try something similar with your transducer?
You would of course need to use something significantly more powerful than a little logic inverter, and you would probably need an additional band-pass filter to make sure you don't end up with one of the transducer's harmonics, but I imagine it would look something like this:

^{simulate this circuit}

You may need to introduce some method of kick-starting it if there isn't enough 'natural' noise in the system to get it going, and the filter may need to have some gain to compensate for the low voltage across the sense resistor.

Answer 3

It seems you aren't the first person to have this problem.

These guys have thought about it and patented a solution.

The solution is basically a microprocessor that controls the frequency of the drive signal, and a current detector in the drive circuit.

Generate approximately correct signal, then hunt up and down while watching for a maximum of current flow.

You could do this by hand using a multimeter with a small adapter. You put a current shunt in series with the drive signal ground, then use a small adapter circuit sort of like this:

^{simulate this circuit – Schematic created using CircuitLab}

This converts the drive current to a voltage that you can measure with your multimeter.

If you had a multimeter that could measure AC current at the drive frequency of the ultrasonic transducer, you wouldn't need the adapter. But, I don't think there are any multimeters that measure AC current for much above typical powerline frequencies.

The diagram is much simplified and is only intended to show the concept. You may need to use two stages to get enough gain, and the output filter could be made much better.

Given that you mention a 50Ohm drive, you may be up in the MHz range with your ultrasound, so maybe an opamp won't cut it and you'll have to use something better suited to high frequencies instead.

Answer 4

What you can do is a very simple broadband high-energy excitation (i.e. a single pulse).

This broadband energy excitation will make the transducer oscillate on its resonant frequency.

If your multi-meter is capable of measuring the AC frequency content then you have your answer.

Alternatively you mention in your linked question that you can measured reflected energy, you could also consider making the measure of the echo energy. Once you have a rough estimation of your resonance you could put an obstacle at a distance equivalent to the time needed for the drive induced oscillation to have died-out and then measure the energy from the reflected wave. When you reach the maximum you will have tuned your driver circuit.

Note: this is a very generic answer as your your diagram is also very generic. Low side shunt measurement like in other answers is usually easier to handle but given that you are using a bench top PA it might not be compatible with your setup. I do believe that the above is generic and flexible enough for you to adopt to your setup constraints. Other answers also mention that multi-meter cannot measure frequencies above AC line, while true of there is one that i know of (Fluke 170) going up to 100kHz. Bottom line without knowing your setup giving clear directions is not easy. Last but not least a DMM might not be the best tuning tool, a simple scope would help you converge faster.