"This kinda sounds like the arcing problem as described. The arcing is from the winding to the carbon brush wiper: "

No, the wiper was nowhere near it. It is either between the windings or from the windings to the core.

From what I see there are probably 100 turns, if not there must be at least 50. That means only a couple volts between each turn. This is not hopeful.

Since it all seems to be at the boost end, maybe I can eliminate that part and just have something that will cut the line voltage. It beats nothing.

There is another possibility, I could put the rectifiers back in the "isolation transformer" cabinet and use the DC supplies which are + - 90 volts to power a class D amp. I know they can get hairy to design like for high fidelity audio but for these lower frequencies it should not be all that difficult. Or I could tap off an existing class D amp before the output filter (except for certain Crowns et. al.)and use that to drive bigger transistors, set up a feedback attenuator to match the level and make a bigger output filter. It would be easier than designing from scratch. It would also automatically be isolated.

But honestly that is a pipe dream. Maybe I should just think about eliminating that part of the winding and hope the rest is alright. Seems like that would be the hottest part of the winding and would fail first, no ? For the amount of time I use it...

Soon, I'll take the cover off and have a closer look at this arcing. But I am sure it is not to the wiper.