Wow! As always, Kevin is terrific.
FWIW, my speculation is that a little bit of 60 Hz and/or 120 Hz ripple is present on the DC that is supplied to the filaments of the tubes. Perhaps its magnitude is a bit greater than when the amps were new, due to aging of the filter capacitors. The excessive heater-to-cathode leakage of the tubes, in turn, couples some of that ripple to the cathodes of the tubes, thereby causing it to enter the signal path.
Thanks for the update!
Best regards,
-- Al