Construction of an ultrasonic cleaner tank

Thanks for the references. Now to find the ideal container.

The sides and liquid surface act a near-perfect reflectors, so there will be a large number of virtual reflected transducers "interfering" with each other even if you only have one transducer. Because the wavelength is small compared to the size of the tank and transducers, these reflected waves -- as well as multiple waves from multiple transducers --- will reinforce and cancel in a classic interference pattern. You can reduce this effect with tank design: don't let width/height/depth be integer multiples of each other, avoid parallel walls, a convex wall is better than a flat wall and a flat wall is better than a concave wall. Placement is non-critical; just avoid having the distance between a transducer an a wall or another transducer be an integer multiple of another such distance.

With a fixed power level, you can have the problem of small items getting too much energy and being damaged while large items that absorb a lot of energy don't get enough and don't get clean. That's why you often see a power sensor and a closed-loop power control.

If possible, sweep the frequency of the transducers. Separate sweeps for each one that are not synchronized are even better. This isn't as big a problem at 40kHz as it was at 10kHz or 20kHz, but it is still something to consider.

Use regular slow-cure (not 5-minute) epoxy to attach the transducers. Make sure that the surface are really, really clean. Clamp or weight the transducers during curing to keep the adhesive layer thin and don't disturb them until the epoxy is 100% cured - give it overnight.

Here are some web pages on the topic;