Thomas's "check connections" and PowderMonkey's comments are both good, the latter if you have an AVR with 120V, single leg sensing, which is a bad idea.
You didn't mention an AVR. Is is using the stock harmonic excitation?
Is this a new problem or have you never tested large 120V , asymmetric loads? I ask because many (smaller than your ST-7) ST heads moan and vibrate badly with such a load, and regulation not good.
My first thought is to go over the basics via ohm meter:
I'd check the resistance of each set of the stator windings looking for a difference in values and documenting the resistance for future reference..
I don's see brushes or the rotor as likely culprets due to the 120V leg voltage asymmetry. But it never hurts to check connections, and document resistance of the rotor coils for future reference.
Best Wishes,
Bruce