TP6 is good. The 4558 won;t work until the supply is at least +-5V which you have it this point. I'd still go for pushing the +-40 V supplies
to +-20 V and measuring the +-15 V supplies and ripple.
Well, here are my findings.
Cranking the rails to +/- 20vdc, I was still unable to get the +/- 15vdc supplies up to full value, so I cranked to +/- 25vdc and got the -15 supply to read -12.9, which was what it used to read (the original problem I was tracing). At this point the +15 supply read +14.1.
I was getting a brief glow on the 40W bulb limiter at power-on and I wasn't sure if this was kosher or not, so didn't push the voltage further. I'd say the 40W bulb flares up to about 1/10th full brightness for about 1/2 second at power-on. Maybe a full second before it fades out.
Anyway, I did take ripple readings at that time. I suspect they're probably not good. With the rails at +/-25vdc, here's what I got:
The +40 supply at C16 had a 1.5vpp sawtooth ripple.
The -40 supply at C16 had a 2vpp sawtooth ripple.
The +15 supply had a 520mVpp nearly sinusoidal waveform at it. Clean looking wave.
The -15 supply had about 36mV of noise. The Q11 collector had a 1.8vpp sawtooth ripple, but very noisy looking. Whereas the other side of R25 (which is the -40 supply) had pretty much the same size ripple, but cleaner looking.
Thoughts?