HP 3326A Two Channel Synthesizer repair

[TerraHertz]

I picked this up for US$80 and it arrived yesterday. It was sold as "powers on and has ERROR 192."  Which from the manual is "ACAL 192 Amplitude CALibration. An amplitude calibration is unsuccessful and the HP 3326A is unable to calibrate internal circuits."

So far I'm just doing the usual clean up and righting minor mechanical issues. I'll update this thread as repair progresses, but for now a few photos and a request. The unit's front power button is missing, switch shaft snapped off. (Deja vu, buttons again...)

For the broken switch the service manual gives just a HP part number 3101-2748 and manufacturer code 28480. Which unhelpfully looks up as "HP head office Palo Alto CA." Great.

Anyway, does anyone recognise the switch from the last picture below? Gooargle isn't helping with the PN.

The burned and open track on the power supply PCB  in pic 4 was for +15 Raw, so no +15V rail. No wonder it couldn't cal! Whether that's due to some fault load, or just a track designed a little too thin for its normal loading, remains to be seen. If tracks get warm, over time they can delaminate from the PCB, so suddenly they have very little heat dissipation and get red hot.

Overall, after a cleanup it's pretty nice. Hopeful I can get it working easily.

[TerraHertz]

Purely mechanical and probably boring. Fixing the two bent BNCs from the front panel.
The photos are self-explanatory I hope.

[MadTux]

Couple of my HP3326A boxes have developed trouble with PSU lately, Opamp directly inhibiting 3kOhm pullup on Darlington transistor is shit design IMO.
Darlington transistor gain has likely degraded over the years in some of my boxes, makes PSU unreliable and 3326A shutdown and restart at random. One got so bad, that it even didn't reach +15VDC on bench.
All regulator transistors tested fine with curve tracer, with some variation in gain, though.

Solution was to power the board externally with 5x HP361x PSU and put PNP inhibition drive transistor between LM1558H Opamp and Darlington transistor and decrease pullup on +15V down to 660Ohm. Transistor is some kind of 2N2907 in weird EBC config from front, bulk purchase from china 

+15VDC is rock solid now, -15VDC and +5VDC gets same fix applied soon.

Diodes the other way round, in my shoddy drawing