To test the idea that a small fraction of the ring heater current can be flowing through one of the mirror coils I have looked at the trend of these coil monitors before and after the mirror change, and before and after the ring heater was turned on, but without the interferometer locked. So that the monitors of the coil voltage are dominated by the mirror control corrections themselves.
Figure 1 shows that there is a change in the average voltage during the several days when the payload is not connected between Apr 7 and Apr 11. All of the voltage go back to the same level as before the payload is extracted, with the exception of the DR coil which stays at the same level. The time when the ring heater is turned on doesn't change the voltage seen by these coil monitors.
Figure 2 shows the spectra, both with the ring heater turned off, before (purple) and after (blue) the change. The DR coil monitor started to see a large 50Hz and harmonics signal on them after the WE change, while the other coils have remained the same. This suggest that there is cabling issue with the DR coil. It would be good to find a more direct way of electrically check that coil, to see if one of the two cables or the shield is connected with something it shouldn't be.
Note that the other large lines around 350Hz, 1kHz and 2kHz are calibration lines applied to the coils: https://wiki.virgo-gw.eu/Commissioning/Calibration/HistoryOfCalibrationLineFrequencies.