V6B (the A side is the master volume for CH3) will only effect ch1 and 2 if the Effects loop is in use, it is completely out of the circuit (grid network and plate network) when set to hard bypass on the back panel, so if it goes sour it will not spoil the mix if turned off. V5B (the A side is used for gain stage of CH3) is the reverb return but may influence signal if it begins to fail as there is no relay contact or switch contact to remove the plate network from the signal path. I would probably start with removing one preamp tube at a time to find the issue (if the one pulled does not change the noise, put it back in and pull out another one). Also note, what may seem apparent on one channel but does not influence the others may be a tube that is used I all three channels. Why? Ch1 and CH2 use a different voltage bias on the preamp tubes compared to CH3. V1 may be the actual issue since it is operating at lower voltage for CH1 and CH2 but is bumped up for CH3. Change in voltage may be causing the issue.
Reason why I suggested the Rec tube, it runs in parallel to the Silicon diode bridge. the heater circuit which is also the cathode feeds the unfiltered DC rail as does the Silicon diodes. However, the Silicon diodes will not remove leakage current if it is passing through the tube rectifier. You can operate the Mark V with the Rectifier out if you are set for 90W mode. Not recommended for long term or the other power modes. The Choke on the DC supply will filter some noise and transients but it will not filter out audible noise. (Choke is the small transformer next to the large Filament transformer).
I suspect it is a preamp tube more than the rectifier tube.
Also consider this, some electrical noise will come in through the power line. Reason I cannot operate the dishwasher at the same time I play the guitar, motor noise can be heard through the speaker.
