Well, I am puzzled. All transistors from Q90 through the finals now have correct bias and indicated voltages (given a slightly reduced power supply, so typically 0.2v lower everywhere) and I can hear a CW or modulated signal in a nearby receiver. As I adjust RV1 the signal in the nearby receiver goes up and down. It was and is again set for maximum, so I think everything through Q90 is probably OK. I've checked R83 as suggested, put another cap in parallel with C80, and nothing makes any difference. I confirmed the power meter is still working with another QRP rig. Other than the area I have been working around Q90, there are no "distressed" looking traces or components.
So what now? If the finals were "dead" would they still bias correctly?
=Vic=
Vic WA4THR