I also see that you are cleaning 6 records at a time, which means a spacing of at most 33mm. That spacing is OK for 80KHz, about 1.7 wavelengths, but, at less than a wavelength, that spacing is quite inadequate for 37KHz. Also, with 6 records your effective US power per record is down to about 55W. I find that I get better results at 75W. YMMV@terry9
Pardon me for being sceptical about the utility of either of those statistics. I'll tell you my thoughts and then you can explain.
I don't see how the wavelength of a given frequency is relevant to the space between records. If you're saying a given wavelength, say ~40mm for ~37kHz (water, 30-degree C) is too wide to fit between a 33mm space between records, I don't see how that makes any difference. The frequency determines the number and size of the vacuum bubbles generated that will implode against the record in solution - that is the cleaning force. My Elma has 6 transducers on the bottom of the tank and they will generate the same number of vacuum bubbles at a given frequency regardless of the spacing between records. I don't see how a wavelength greater than the distance between records changes ... what ? - the access of bubbles to records, as if that wavelength limits how many vacuum bubbles get to the records? I don't see it.
The claim of higher watts per record is based on having fewer records in a given tank. I don't see why the ratio of records to watts makes a difference. The same number of watts will be output regardless of the number of records. Granted there are more bubbles per record with fewer records but the records are in a fixed position and the total bubbles in the tank at any given time is the same independent of number of records. It's not clear that fewer records 'attract' or receive more bubbles than a greater number of records.
Wrt frequency: There is a correlation between particle size, particle tenacity, the efficiency of particle removal, and frequency. If there was relatively constant particle size on a record we could target the frequency to that. But given the state of used records, there is no such constancy. From a visible glop of something to a few microns, multiple frequencies target a broader range of dirt.
Btw, which Elmasonic model do you have? As you read, mine is the P120H..