Although I agree with most of what @theaudioamp has suggested, I'm yet to be convinced this is actually the fault of the phono section, even though the phono section is clearly able to amplify the noise.
First off the 12AX7s have nothing to do with this. They are not somehow inherently sensitive to RFI. How their grid circuits are set up is an entirely different matter...
The length and exposure of the 20 foot XLR cables is a clue.
I think that what is happening is there is noise being picked up and injected into the ground of the ARC (which plays the same sound at a much lower level...). Since the ARC does not support the balanced standard (even though otherwise is balanced) a noise is manifest.
To sort out what is going on first short the inputs of the phono section to be sure its OK. Removing the top cover and no additional noise points to the idea that the phono section really isn't the culprit. But if shorting the phono inputs shuts it up then the phono section needs mods to protect it from demodulating RF (demodulation is what causes nearly all RFI problems). I've seen this sort of thing caused by poor tonearm cables too, but if the phono section is adequately protected from RFI its usually no concern.
If shorting the inputs has no effect I would be trying to use different interconnect cables. A lot of 'high end audio' manufacturers have a poor grasp of how balanced line is supposed to work so I would be trying a cheap microphone cable like you can get at Guitar Center since such a cable will be wired correctly and will also have an effective shield. Note that this cable is between the preamp and amp; the idea here is that noise is being injected elsewhere from the phono section but is able to be amplified by the phono section; the noise is in the grounds of the preamp and is being passed to the phono section.
