Quote:
Originally Posted by Alistair D
I was thinking about how isolate which section of the amplifier is drifting. The idea I have come up with is to use two meters. Connect one to the 2 gates of TR601, connect the other meter likewise to TR603. Monitor the voltages as the scope warms up. Both meters should drift by the same amount. Repeat this process with TR605/607 etc until you find where the two meters disagree. Hope it helps.
Al
|
Thank you for sharing your thoughts on this. Differential measurements were a good idea. Until now, I just measured the voltages to ground. Unfortunately, I do not have two suitable meters. So I got me an optical PC interface for my DMM, which enables me to log measurements with a 1 second sampling interval. The measurements were conducted with half an hour cool down time for the scope in between the individual measurements to have a kind of reproducible starting point. The results are interesting:
- The issue with channel2 consists of largely off values on startup combined with a nasty drifting behavior which does not stop even after an extended amount of time.
- Since the drifting issue of channel 2 can not be observed with channel 1 and with both channels of the second V4 amplifier from my D755, I consider it to be a repairable issue. In addition to that, it completely prevents the unit from being calibrated with respect to any kinds of DC offsets.
- Swapping TR601 with TR603 did not swap the drifting issue between channel 1 and 2. Thus, both FETs should be OK and the issue comes from another component.
- Swapping TR605 with TR607 did not swap the drifting issue between channel 1 and 2. Thus, both FETs should be OK and the issue comes from another component.
- Measuring between the drains of TR601 and TR603 shows no voltage drift, which means that the drifting zener voltage is passed with a constant offset to both channels over R617 and R618. Since channel 1 works acceptably, the issue is most likely not located here.
- Voltage drifts between the gates of TR601B and TR603B were too small to be measurable with my DMM.
- Measuring between the gates of TR601A and TR601B, as opposed to the gates of TR603A and TR603B clearly showed my drifting issue(see attached image). The graph for TR605 looks like its quickly stabilizing. The graph for TR607 in channel two, on the other hand, only converges very slowly.
- IMHO this narrowed down the issue to the gate of TR607 and the connected source of TR603. Since both FETs are not linked to the issue(ruled out by swapping), the issue should be linked to the connected R615 and R616. So I measured the voltage drop over time across these resistors and their counterparts R612 and R613 in channel 1. The results can be seen in the second attached image. While the resistors in channel 1 seem to maintain a constant offset, this is not the case for the resistors in channel 2. I think, a constant offset should be expected from resistors with equal value and load. When plotting the difference of the voltage drops over the two resistors in channel 1 and 2(see third image), the ones in channel 2 show a curve that resembles the one at the gate of TR607. Considering the trace position is extremely sensitive to changes of both resistors(checked by individual cooling through blowing) the non equal temperature drift of the resistance of R615 and R616 is maybe the source of my issue. The voltage in between the two resistors is exactly the same like at the corresponding measurement position in channel 1.
The shape of the temperature curve of R615 and R616 both do not match the shapes for R612 and R613. Fixing the issue could require to find a pair of resistors with matching temperature curves. I am not sure if it is a good idea to swap all four resistors with some 0.1% tolerance metal film ones like described in my previous post. After all, the temperature drift of these resistors may part of a design to compensate the temperature drift of other components. Maybe it counteracts the voltage drift at the zener diode.
Denis