UK Vintage Radio Repair and Restoration Discussion Forum - View Single Post

Pinörkel · 11th Aug 2019, 6:55 pm

I just tried to switch R615 and R616 with some 1% tolerance metal film ones I had lying around. The result was only sort of a success. When previously properly zeroing out and then switching the scope on after a 30 minute cool-down, the 1x trace of channel 2 was now nearly at zero(off by 0.5 tick), whereas it was off before by about -0.5 major divisions (5 ticks = 1 division). The 5x trace is now off by 0.5 major divisions, whereas it was off before by about -1.5 divisions. So a clear improvement here. What bugged me was that the general shape of the voltage drift function over time did not not change. IMHO this means that the resistors with the non-matching temperature coefficients only amplified the actual issue but are not responsible for the qualitatively differing shape of the drifting voltage function. To verify this, I harvested a pair of matching carbon film resistors from an old TV board and tried those. The result was very similar to the one I got with the metal film ones, just a slightly higher temperature drift. The next thing I suspected was the TR603 FET which acts as a source follower to provide sufficient current for TR607. So I swapped in the original TR603 instead of the one I got from MrBungle. After re-zeroing the channel and a cool-down, the result was close to the previous ones which means TR603 should not be the source of the issue.

The attached image shows a comparison of the measurements taken so far. A comparison measurement of the TR607 gates in the V4 of my D755 showed that they qualitatively behave exactly like channel 1 in my D75. So, I am pretty sure that channel 1 should be used as the reference target here.

Summarizing, this does not leave many options:

The issue shows at the gate of TR607, so it should be located before this FET.
According to my observations, it is unlikely that R618 at the drain of TR603 or R706 are the issue. Blowing on them did not result in visible trace drifts and their temperature drift converges much quicker than the phenomenon I am looking for.
The issue could be located before the gate of TR603. If so, a voltage drift should be measurable between the gates of TR603A and TR603B. At least with my DMM I can not detect such a change, only fluctuations with a magnitude of 0.01 mV to 0.02 mV with the total difference being at 2.00 mV. Maybe my measurement equipment is not up to the task. So, I looked further and measured the temperature drifts of the voltages across R604 and the center pin of the R602 DC balance pot to ground. Both exhibit exactly the type of drift I am looking for. However, so do R603 and R601 in channel 1 and channel 1 is working near perfectly.
A quick measurement indicates that the temperature drift function of the new carbon film resistors resembles the one of the original ones. So it is somewhat unlikely that the newer resistors I swapped in have qualitatively non-matching temperature behavior.

At the moment, I have run out of ideas but I am not going to give up this easily. Any analysis ideas and thoughts are welcome.

Denis

11th Aug 2019, 6:55 pm	#55
Pinörkel Pentode Join Date: Jun 2019 Location: Dortmund, Germany Posts: 161	Re: Telequipment D75 scope. I just tried to switch R615 and R616 with some 1% tolerance metal film ones I had lying around. The result was only sort of a success. When previously properly zeroing out and then switching the scope on after a 30 minute cool-down, the 1x trace of channel 2 was now nearly at zero(off by 0.5 tick), whereas it was off before by about -0.5 major divisions (5 ticks = 1 division). The 5x trace is now off by 0.5 major divisions, whereas it was off before by about -1.5 divisions. So a clear improvement here. What bugged me was that the general shape of the voltage drift function over time did not not change. IMHO this means that the resistors with the non-matching temperature coefficients only amplified the actual issue but are not responsible for the qualitatively differing shape of the drifting voltage function. To verify this, I harvested a pair of matching carbon film resistors from an old TV board and tried those. The result was very similar to the one I got with the metal film ones, just a slightly higher temperature drift. The next thing I suspected was the TR603 FET which acts as a source follower to provide sufficient current for TR607. So I swapped in the original TR603 instead of the one I got from MrBungle. After re-zeroing the channel and a cool-down, the result was close to the previous ones which means TR603 should not be the source of the issue. The attached image shows a comparison of the measurements taken so far. A comparison measurement of the TR607 gates in the V4 of my D755 showed that they qualitatively behave exactly like channel 1 in my D75. So, I am pretty sure that channel 1 should be used as the reference target here. Summarizing, this does not leave many options: The issue shows at the gate of TR607, so it should be located before this FET. According to my observations, it is unlikely that R618 at the drain of TR603 or R706 are the issue. Blowing on them did not result in visible trace drifts and their temperature drift converges much quicker than the phenomenon I am looking for. The issue could be located before the gate of TR603. If so, a voltage drift should be measurable between the gates of TR603A and TR603B. At least with my DMM I can not detect such a change, only fluctuations with a magnitude of 0.01 mV to 0.02 mV with the total difference being at 2.00 mV. Maybe my measurement equipment is not up to the task. So, I looked further and measured the temperature drifts of the voltages across R604 and the center pin of the R602 DC balance pot to ground. Both exhibit exactly the type of drift I am looking for. However, so do R603 and R601 in channel 1 and channel 1 is working near perfectly. A quick measurement indicates that the temperature drift function of the new carbon film resistors resembles the one of the original ones. So it is somewhat unlikely that the newer resistors I swapped in have qualitatively non-matching temperature behavior. At the moment, I have run out of ideas but I am not going to give up this easily. Any analysis ideas and thoughts are welcome. Denis Attached Thumbnails