Granny, wow, that's mystifying. Sounds like you've tried pretty much all of the logical things.
Cheater plugs would have broken a ground loop, if that was the cause of the problem. And the Lightspeed wouldn't be susceptible to ground loop problems, because its power supply is optically isolated from the signal path, as I understand it. And also because it has a two-prong AC plug, with no safety ground connection.
You've probably already tried this, but if not try connecting the preamp and amp to the same dedicated line, and try doing that on each of the two dedicated lines.
I assume, btw, that nothing is connected to the outputs of the amplifier other than the passive speakers. I see that the Atlas is fully balanced, so if for instance a powered sub were connected to its outputs in the wrong manner, hum problems (or worse) could result.
The only other thing I can think of is that I see there is no switch on the back of the amp to select between the RCA and XLR inputs. Perhaps therefore the RCA input is wired directly to the same circuit point as pin 2 of the XLR connector (the non-inverting XLR input signal), in which case a jumper might be needed between pins 1 and 3 of the XLR connector when the RCA input is being used. Perhaps the manual addresses that. Perhaps the EMI environment is different now that the powered electrostatic speakers are no longer present, resulting in a difference in the amount of hum that is picked up by the floating input on pin 3.
Just some wild guesses, but I can't think of any other possibilities.
Regards,
-- Al