My OP contained an error in that the voltage on the +20 regulated supply
line wasn't zero but in fact was 27V. Or maybe it was zero the first time I
powered up the unit measured it and then something "snapped" and it went to
27V. [Maybe I measured the wrong thing the first time -- we'll never know.]
In any event, the last 3 times I've powered the unit the voltage is a
clearly unregulated ~27V which results from the output of U420 (the error
amp) being stuck as high as a 741 will go in spite of the imbalance in the
inputs with the voltage at the inverting input almost 2 volts higher than
the voltage at the non-inverting input. The voltages at the pass transistor
Q420 are completely consistent with it being perfectly sound and just being
driven into (near) saturation. There are no shorts in this transistor.
Neither is there any problem with the current-limiting transistor either.
All voltages in the +20V regulation circuit are consistent with a failed Op
Amp and with no other hypothesis I can propose.

As far as I can determine, all of the other voltage regulating circuits are
functioning "properly" -- that is their outputs are all perfectly
proportional to the 27V on the 20V line as they're supposed to be.

Based on the serial number this is an early unit. The Op Amp at U420 has
completely different markings from the other 741's in the unit which are all
of a single type. The soldering on U420 looks exactly like the others and
there are absolutely no signs of any hand-work in replacing it; I have to
believe that it's the original as it came from the factory.

The voltage on the +5V line is ~6.7 volts. There is a single TTL IC (7402)
and an MC1711CL which IIRC is some kind of ECL gate. Neither of these would
be very happy with almost 7V on Vcc. Everything else on the board is either
discrete transistors or 741's (or similar) so I don't think that any of the
other ICs would have been damaged. While the sine shaper might not put out
a good waveform with the higher voltage, none of the other outputs are
working either so I suspect there's a fundamental problem with the main
oscillator. That's why I'd like to simply substitute a a good +20V supply
for the output of this circuit and see if the rest of the unit is intact.
It may in fact not be and it may simply not be worth the time to debug it.
I've been thinking that I could easily rig up an LM317 (as you suggest) or
an LM7818 to substitue for this circuit.

My point in checking for oscillations was that sometimes an oscillating Op
Amp can give some very strange readings on a simply DMM.

Norm