Hi,

This noise arises because the servo is actively working to maintain its position and counteract any displacement.
That is almost exactly what I said from the beginning.

I suspect the PWM output of the servo drive is 10.7kHz, or maybe some multiple of it. The cycle rate of the control loop has nothing to do with the PWM rate. I rather suspect the noise you are hearing
is not 'dither' but rather is one of the audible harmonics of the PWM.

How long are the cables between the drive and the servo?

There is in most manufacturers specs, Yaskawa included, that call for motor side (called load side) reactors if the cable length is long. May I suggest you consider trying load side reactors.
You need to be exceptionally careful here. Most EMI line filters have capacitive input and capacitive input filters are VERBOTEN with PWM output devices. The load side reactor input impedance
MUST be NET INDUCTIVE or you risk your drive.

With the forgoing, then ordinary EMI filters CANNOT be used on the load side of a servo drive. If you are going to try load side reactors then I would follow Yaskawa's instruction carefully.

Line side reactors and EMI filters are very much easier to find and specify.....and any capacitive input of an EMI filter will do no harm on the line side, but not so on the load side.

Craig