Ultrasonic transducer requirements

Hi all,

I'm looking to build a large ultrasonic parts cleaner, and the cost of existing units big enough to take engine blocks etc is simply far too high. Total power consumption of units this sort of size is around 3kW - so i'll be looking to replicate that sort of power.

The type of transducer I'm using is this sort of thing - http://www.ebay.co.uk/itm/40KHz-Ultrasonic-Transducer-for-Ultrasonic-Cleaner-50W-/221191921193?pt=LH_DefaultDomain_0&hash=item338010de29

What do these require, I guess a 40kHz (or other frequency listed) sine wave, saw tooth or similar? How would one get a sine wave through transistors etc?

Many thanks!

I guess a 40kHz

Good guess considering it is in the title. :slight_smile:

How would one get a sine wave through transistors etc?

No need, just hit it with a square wave and let the natural resonance of the transducer turn it into a sin wave.

Grumpy_Mike:

How would one get a sine wave through transistors etc?

No need, just hit it with a square wave and let the natural resonance of the transducer turn it into a sin wave.

Driving the transducer directly with a square wave might be OK for low power transducers, but is will result in high peak currents, and corresponding RFI and losses in the driver, which will be unacceptable at high power. I suggest you use either a series resistor (which will itself introduce a loss), or a series inductor and flyback diodes. If you use a mosfet half-bridge or full-bridge to drive the transducer, then the mosfet body diodes will serve as the flyback diodes.

Saw a construction article long ago in Radio Electronics magazine, I think, for an ultrasonic parts cleaner. I seem to recall that they made the piezo transducers part of a tuned LC circuit, with added capacitors.

polymorph:
Saw a construction article long ago in Radio Electronics magazine, I think, for an ultrasonic parts cleaner. I seem to recall that they made the piezo transducers part of a tuned LC circuit, with added capacitors.

Sounds good - driving the transducer from a square wave fed through a series LC circuit tuned to the transducer frequency should work well.

I think it is useful to point out:

A square wave drive through a series tuned circuit is fine because a series tuned LC circuit presents a high impedance to the harmonics contained in the square wave.

Whereas a parallel tuned circuit presents a -low- impedance to the harmonics, causing losses in the drive transistors.

A bare piezoelectric ultrasonic transducer's equivalent circuit is a capacitor in parallel with a series-tuned LCR circuit. With just half-an-H-bridge square wave drive, the transistors must charge and discharge that parallel equivalent capacitor on each half cycle.

That equivalent circuit as described above is going to have two resonances: series tuned and parallel tuned resonance. Ideally, I think you'll want the low impedance series tuned resonance.

I've never bought from this company, but $8.50 each seems a lot less expensive than $25 each:

I'm doing a lot of Google searches for ultrasonic cleaner design. My head is spinning. The lower power designs mostly just get away with overrating the components and driving with a raw square wave.

I must add that an ultrasonic cleaner is potentially hazardous. Probably not those crappy plastic "jewelry cleaners", but the higher powered ones. Note that this video suggests a test that results in a layer of aluminum foil being destroyed. Now imagine cavitation going on inside your hand if you reach into a running ultrasonic cleaner.

Maybe we're overthinking this. In the ultrasonic cleaning world, 50W -is- low power.

Polymorph - you're right, I once put a hand in one of those low power 'steam generator' haloween type ultrasonic baths - which didn't feel nice to say the least.

...and yes, true - 50W is very low power, I'm looking to build a much, much more powerful bath - which leads me to a problem; driving multiple transducers.

Would they be in series, or parallel - and surely, wouldn't all the transducers have to be in perfect sync else waves will cancel eachother out?

To have them in sync, merely provide the same drive signal to each. They don't have to all be on the same driver transistors. In fact, it might be better if each has its own H-bridge or push-pull driver, and then just send one clock signal to them all.

Thanks Polymorph - that makes sense - whilst they may not all be 100% in phase with eachother, the amount of phase shift between them should at least remain constant, surely?

The reason behind my thinking is the non matched impedance of PCB traces etc to/from multiple driver chips may cause one or more transducer to become out of phase, but it should stay out of phase by a constant value?

At 40kHz? How long are your traces?

The wavelength of 40kHz ultrasound in water is about 8.5mm. If you have all the transducers driven at exactly the same frequency and with zero (or constant) phase difference, then you will get standing waves caused by interference between them. This will result in areas on the object to be cleaned that have little or no ultrasonic waves, and therefore do not get cleaned. To get thorough cleaning, I think you will need to either selectively turn some transducers off, or selectively adjust the phase - possibly just by having the option to drive each transducer in-phase or 180 degrees out of phase with respect to its neighbours.

polymorph:
At 40kHz? How long are your traces?

I guess that was a stupid question then! Not long, 4-50mm. Upon reflection, the problems I've listed shouldn't be an issue at those frequencies.

dc42:
The wavelength of 40kHz ultrasound in water is about 8.5mm. If you have all the transducers driven at exactly the same frequency and with zero (or constant) phase difference, then you will get standing waves caused by interference between them. This will result in areas on the object to be cleaned that have little or no ultrasonic waves, and therefore do not get cleaned. To get thorough cleaning, I think you will need to either selectively turn some transducers off, or selectively adjust the phase - possibly just by having the option to drive each transducer in-phase or 180 degrees out of phase with respect to its neighbours.

Would this explain why when you watch a cheap cleaner, you get waves 'standing' on the surface constantly around that size peak to peak, (or more or less depending upon working frequency), yet with the more expensive units the water is perfectly flat - as they must be as you say, varying the phase of the transducers to ensure the whole part gets cleaned?

jtw11:
Would this explain why when you watch a cheap cleaner, you get waves 'standing' on the surface constantly around that size peak to peak, (or more or less depending upon working frequency), yet with the more expensive units the water is perfectly flat - as they must be as you say, varying the phase of the transducers to ensure the whole part gets cleaned?

That sounds plausible, but I am not an expert in this area. Reflections from the sides of the bath will also be a factor in creating standing waves, but I think these could be moved around (maybe fast enough so they are not easily seen) by using several transducers and varying the relative phases between them.

dc42, very good point.

Hi guys!
any progress? i'm here to learn about ultrasonic transducers and how to drive them properly. Got some fresh thoughts about this topic?