[plaurids]

Dear all,

this is my first post on Head-Fi. I've just made some upgrades on my Stax SRM-T1S tube electrostatic headphone energizer (recap, added safety resistors to output terminals, replaced plate resistors with constant current sources (CCS's) - see this Head Case post and its parent thread for more details) but I stumbled at a problem when I started rebiasing the output tubes, whose cause I've been unable to pinpoint so far. The energizer did turn on and didn't explode (phew!), but it seems to me that the delay circuit is now taking longer than before to "release". So far, no big deal (I think, but it may be important in what follows, so I'm taking note of it). When rebiasing the tubes proper, the voltage between the channel terminals (Left +/-, Right +/-) and the ground terminal (e.g. in the back), which is adjusted by the Balance (TVR2) trimpots, can be set near zero on the right channel as it should but stays at about +/- 280-320V on the left channel, depending on the left channel's TVR2 adjustment. The left channel's Offset (TVR1) trimpot can be set "around zero" (sort of), but it swings with a much larger voltage amplitude than the TVR1 on the right channel - the voltage fluctuations are certainly far above the 15V bias tolerance interval. The latter problem alone makes rebiasing the left-channel tube near impossible. Like the TVR2, the TVR1 adjustment on the right channel is behaving normally.

What I could observe so far is the following:

1. The problem is not caused by the new power supply caps, since judging by the schematic they should affect both channels equally;
2. The safety resistors are measuring properly after installation - no shorts, no missed contacts;
3. A problem with some of the CCS's is possible but unlikely - I don't know if a faulty CCS circuit could cause the problem I'm observing, but if that were the case, it would mean that precisely the two CCS circuits connected to the left-channel terminals are faulty;
4. I tried to swap the tubes, but the problem stays in the left channel. The same happened after I replaced both stock tubes with a pair of Japanese, short-plated Raytheons (not a matched pair). Therefore, it doesn't seem to be a tube problem;
5. The tube socket leads on the left channel are not cut off;
6. All four red leds on the PCB are lit - the only odd thing (which I don't know if it's related or not) is the additional time the delay circuit is taking to release the energizer.
7. The 2Kohm TVR1 and TVR2 trimpots on both channels are measuring properly (with the amp turned off) at several adjustments on the circuit and don't seem to be mechanically damaged. However, sometimes when turning the left channel's TVR2 trimpot, the voltage drops to reasonable levels for a few seconds and then goes back to 280-320V. This seems to indicate a faulty trimpot, despite the fact that resistance measurements with the energizer turned off seem to indicate normal operation.

Needless to say, the energizer was working normally before I started tinkering with it. Any ideas of what could be causing this? In view of 7.) above, I'll try to desolder the left channel's TVR2 trimpot from the PCB and measure it alone to see if it's really faulty. If that's really the case, I'll probably replace all TVR trimpots with e.g. 2Kohm, 1/2W Bourns 3386F trimpots, which should be fine. I couldn't find 2K trimpots with the same footprint as stock and power rating higher than 1/2W, even though I've got a recommendation of 0.6W of higher based on the rule of thirds. The only other component which could possibly affect the left channel alone in the TVR2 measurement is the left-channel C26 capacitor (ERO, now Vishay Roederstein, model MKP1841 - polypropylene, 0.01uF, 630V), which is connected to ground, but I think this one failing is highly unlikely.

 

[plaurids]

I've found the culprit. A tiny solder blob which seems to have landed out of nowhere was shorting two solder points in the CCS module's PCB. Removed it, resoldered the module to the unit's PCB... and presto! Left-channel voltages back to normal! Phew!