I finally found a problem, the magic 47K resistor. From all appearances completely normal. However, it is open. I've never seen this in a metal film resistor, even the carbon resistors crack first before they open. Now, I may replace this defective part and something else will show up, but no doubt this was contributing to the problem. Who knew? My explanation while looking at the schematic goes like this: R4 is fed +85V from the diode train of D1-D3. R4 drops almost all the voltage and sets up about +1.8V to the 1K Ohm DC offset null potentiometer P1. The null DC voltage is wiped off the pot and sent to R3, another 47K Ohm resistor to combine with the audio input to turn on Q3 or Q5 depending on the polarity of the input waveform. The other function of R4 is to allow current flow to the base of Q1. When Q1 conducts it sends +1V to the emitter bias resistors R15 and R16. These 47 Ohm resistors reduce the emitter bias voltage on the differential pair of Q5 and Q6 to +.6V each allowing them to conduct the negative going input signal present at the base of Q5.
Thoughts?