PI balance setting adventures (what test freq to use)

[MKB]

Hi all, did a bit of testing recently on an ODS build phase inverter (setting the balance), would be interested in your views on the issue.

According to the PI balance sticky, HAD recommended using the Ampeg SVT balance set method, which is basically to set the amp with a 40Hz tone at 1/2 rated power into a dummy load, and adjust the PI trimmer to minimum distortion using a distortion analyzer. One post in the sticky recommended the same procedure, but use 100Hz. Other posts recommend a test frequency of 1kHz. What adjustment frequency is best?

I first set the PI balance on this amp using the ear method, and the pot adjusted to approximately 11 o'clock. This was marked on the pot.

I have access to a HP 330A distortion analyzer and a 3311A function generator, so I tried the 1/2 power 100Hz setting (thinking the 40Hz might be a bit low for the Bassman OT my ODS has in it). The balance pot was then set to minimum distortion (there was a sharp null noted on the analyzer). Using this method, the minimum distortion setting on the PI trim pot was not at 11 o'clock as by ear, but closer to 9:30.

Finding this interesting, I repeated the same analyzer test but with frequencies at 250Hz, 500Hz and 1kHz. As the test frequencies increased, the adjustment on the PI balance inched closer to the earlier setting I achieved by ear, with 1kHz lining up with the ear setting. I left the setting at this point.

There might be other variables I am not taking into account (such as distortion from the function generator), but this was interesting results nonetheless. Am curious if any of you have seen such variability in your adjustments, and if the test frequencies made that much of a difference in your testing.

[JazzGuitarGimp]

You should be able to rule out function generator distortion simply by taking the ODS out of the equation: just output the generator into the analyzer and check at all frequencies you're interesting in using to test the amp.