With about half mains voltage , for a bit of personal safety. About 20V over
C-E of one driver Q and 120V over C-E of the other, whenever powered up.
Plan of campaign will be to try biasing on the off one and see what happens.
Could a leaky driver transistor be the most likely cause of non-oscillation?
one sliver-mica cap could have metal migration , others are polyester and
ceramic. Otherwise loads of diodes and coils.