OK, so I rewired the heaters and that did help marginally in reducing the hum, but it is very much still there. One question I have is around 'ground' and 'PE' on the schematic. Obviously PE is the physical chassis earth, and ground is the ground return path for the circuit. I currently also have 'ground' attached to the chassis. I've tried at the same point, and separate points to the physical earth, but I'm wondering whether perhaps I should lift that from the chassis? Also, in the schematic, it shows 'J5' connected to 'ground'. What is J5?
Another interesting thing is I can't seem to see the noise at the output on my scope. There is a very small amount of noise visible, but it is not effected by either volume control (which the hum is), and it seems to be quite high frequency, where the hum is sounding to me like 50Hz (but possibly 100Hz). This makes me think it is some sort of ground loop issue, but I'm not really sure. Another odd symptom is that putting the lid back on the chassis actually increases the hum, rather than reducing it, which is another reason I suspect grounding. Any thoughts?
I tried unbolting the power transformer and moving it further from the XLR's but there was no discernible difference in noise. I have also identified a very microphonic 6BA6 in one channel which seemed to be causing some of the motor boating symptoms, but I tried swapping it from one channel to the other and there was no real discernible effect on the hum either, although the motor boating did seem to follow that tube. Also, just how closely do I need to be able to match those 6BA6 tubes? I seem only to be able to get within about 2V, which is less than 1% discrepancy, so I figure it is OK, but I'm curious.
Last couple of things... I just noticed trawling through the BOM document. It mentions using thermal paste and an isolator for the LM317. I used thermal paste, but missed the isolator. Could that be leading to hum somehow? Also, R27 is getting massively hot.