[Pinörkel]

I did some further debugging and followed the idea of Argus25, cross-wiring stuff. Swapping the 47K DC calibration pots R601 and R602 suggested that they are not part of the error, since channel 1 worked perfectly with the other pot. In the same way, I could also rule out the Volts/Div switch connected to the gate of TR603A via the 100R resistor. Looking at the voltages across the resistors in the gate circuits of TR601 and TR603 over time, I did not find any obvious differences either. Swapping the resistors in the gate circuit of TR603B with new ones also did not cure the shape of the voltage function between the two gates of TR603. A small leakage of C602 should not be critical, since there are resistors to ground anyways. Still need to check C623, but I do not have an appropriate replacement at the moment. Its counterpart in the circuitry of channel 1 is too complicated to reach without desoldering tons of wires.

Now I was kind of desperate, so I tried to fight the issue by using the issue to my advantage. I had noticed that the difference of the voltage development over time between using the original resistors for R615 and R616 and using new carbon ones was approximately the difference of the voltage function I measured with the pair of new carbon resistors to the one I was looking for. So I simply tried putting in the original resistors, but swapping their places to produce the opposite effect. To my surprise, this stupid idea kind of worked. As can be seen in the attached image the voltage difference between the gates of TR603A and TR603B now converges much quicker and stays at the final value. Now, the DC offsets of the V4 amplifier can be calibrated and remain sort of stable. However, this is for sure not the way it is intended to work. I cannot believe that the Telequipment engineers manually selected resistors from the same series with specific temperature functions for each scope. In addition to that, channel 2 still does not converge as quickly as channel 1. All In all, I still do not consider the issue to be fixed but only mitigated by using stupid tricks.

While doing all the component swapping, I also detected a small voltage drift of the +24V rail. Since it does not have a noticeable impact on the operation of channel 1, which I consider to be good, I will leave it that way for now and fix it later.

Since Argus25 also suspected the thermistors I had a look on those. They are probably designed to compensate temperature drifts of R698 and R699, which they manage to do quite satisfactory. Measuring the voltage across the resistor/thermistor pairs did not reveal any significant drift.

Denis