I made the following changes:
1. Re-tested all valves and changed the ones measuring low. I ordered new ones as have some 6N3P unmatched sections (V18). This could the reason why I can't calibrate the uA voltmeter given the mismatch between triodes.
2. The Ug1 pots changed to 10T wirewound
3. The Ua and Ug2 pots changed to new 1Meg ones.
I tested all valves per manual and the voltages are ok. I re-calibrated the transconductance meter and struggled to get the 450mV reference. The voltage drops quickly from 500mV down to 20-50mV when adjusting R157. The frequency was set to 1400Hz and voltage left to 500mV. It seems to work fine.
The filament supplies are ok. I can set grid voltages accurately, however there is a load impact when switching from "Ug1" to "Ia". For example, I was testing a CX301a and set the voltage to -9V in "Ug1" that changed (measured at the grid with external voltmeter) to -8.8V.
So what is not working:
1. I can't measure the Ua on the meter. It sounds like a bad contact on the rotary switch. Any tips for dismantling it and cleaning it? It seems like a complex stuff. Im scared with the amount of connections and whether I could bring all the wafers together again!
2. The anode shunt resistors are out of tolerance (>0.2%). I found 2 measuring out of 0.2% - R50 (4R889) and R46 (97R78). I contacted the seller to see if he can send me them. When measuring a CX301a traced with the Utracer V3, transconductance was very close, however the Ia measurement was way off. This led to investigate and found the resistor drift issue.
3. The screen current resistors are within tolerance (<0.2%). However, when following the screen current calibration/test from the manual (e.g. I used a variable CCS as an external load with a series precision current meter) the meter goes into full scale regardless on the range. Something is wrong here
4. The meter seems to be working fine. Although is not the originally specified meter it has 745R impedance but the R100 measures differently from what it should be. However, the meter seems accurate when measuring Mains, 250V and Us. Also the transconductance values are ok when tested some sample valves. This is leading me to suspect that R100 was chosen to match the actual impedance of the meter in place.
Frustrated, seems to be a long process to hopefully restore this beauty of tester.
Hope any folks out there can give me some advice