I have rewired the test stand (service cap wiring only, not power supply wiring) to anode bias mode as shown in the diagram. This was not how it was wired before, there are several things I have removed that appear to be newer additions. I also removed the burned cable between TBHT(-) and COIL(-), and between the multipin feedthrough and ground (probably related to the other burned cable).

I tried to measure the resistances across the feedthroughs (as opposed to across the resistors themselves) but they all seem wrong. I get 27Ohms across the TGHT and TBHT circuits instead of 100Ohms. Between the TGHT and TBHT I get 100Ohms as expected. Across the coil circuit I get 6Ohms instead of 100Ohms. Probably only the 6Ohms really needs to be fixed, but if the resistors are damaged, then the varistors might be as well.