I see. I understood. B+ to be steady at 107.5V, and TB+L and TB+R to be steady at ~81.5V but TPL and TPR is having a larger drop than expected. My first thoughts lead to the current mirror as TB+L and TB+R are after the multiplier.
Some other thoughts. If TB+L and TB+R were purposefully set to 80V and not the max voltage, that's OK. If it's not then there may be some drop due to current flow through the multiplier (more current = more drop). That may also explain the large drop through (TB+L)-TPR and (TB+R)-TPR. Easy way to tell is to measure TB+L and then measure the output of the TL783 to get the drop across the cap multiplier. It should be a few volts at most. If it's not, then it's Q1P and Q2P.
After verifying that, measure the tail current to make sure it's steady, if so, it's definitely the transistors in the mirror or the tube(s). It's unlikely it's all the tubes that were tried, so I would tend to think the transistors.
If it's not steady, that may explain the variances through the HV chain due to current draw.