At long last I had some time...
Thank you Fraser, thank you RadioAct, and all the contributors, for all the information, pictures, schematics, testing, data and instructions.
I went through almost all of it... Directly controlling the varactors proved I was on the right path, and the TG could be saved.
I do not have a good way to measure the 3840MHz, so I set the tracking generator to 500MHz -40dbm feed that in to a Rigol DSA815 with a 10MHz GPS reference.
I have what was called the "Type 1 early" Advantest tracking generator. But the ugly kind with no adjustment arm, just a 1206 0 Ohm resistor to move around... back and forth all around, it did not help, the VCO stayed at -11.21Vdc...
I built up solder, then I wicked away solder.
Finally I decided to stack a capacitor on top of C312 to simulate moving the dielectric trimmer.
See Fraser's pdf:
https://www.eevblog.com/forum/testgear/advantest-r3131a-spectrum-analyzer-9khz-3ghz/msg864931/#msg864931Measurement at TP401 "TUNE", with cover held on by gravity
-11.21V with 0.5pf -5.21V with 2.0pF -5.70V with 4.7pf -9.83V with 15pF, -10.18V with 22pf 0805 Package
I am staying with the 2.0pF, I don't know the original value of C312 but this gives it some swing the op-amp is running of +/-12V
For any one having low +12V, check the power supply fan, mine was pulling down the +12V supply.
While typing this the VCO has moved to -5.38V and holding the 500MHz it is set to.
After a Self Calibration the voltage is -5.51V~-5.52V.
Next is to put it all back together and see if it stays working.
Again thank you all for the information.
:-+Is the test normal after TG repair?