Left channel gets warbly and weak after hours of running

[ohshitgorillas]

After a months-long hiatus of use, before which everything worked perfectly, I fired up my Eros 2 the other day to find that the longer the amp ran, the left channel became weaker, warbly, and odd sounding. I have isolated the problem to the Eros 2 left channel by swapping tubes tubes, cables, etc.

I tried to identify a few less-than-ideal solder joints and hit those again and thought I fixed the problem because the amp works just fine for the first ~45 minutes... eventually I sat back down with a test record and oscope and noticed that the left channel positive peaks were clipped.

Resistance checks passed, all LEDs light up, however I am getting 206V at OB vs 185V at OA  (run time ~5 minutes). I'm also seeing 1.7V at OkD vs 1.3V at OkC... the latter two are within spec but that's a pretty big difference considering everything else is within a few percent between L and R channels. All other voltage checks are within spec.

Any help would be awesome, thanks in advance!

[Paul Birkeland]

206V at OB and clipped peaks are the same problem fore sure.  I would definitely try replacing the 6922 before doing much else.  What are OKA, OKB, OC, and OD?

[ohshitgorillas]

I don't have many spare 6922 but I did pull one from my headphone amp at work and got the same results.

OKA: 99V
OKB: 98.5V
OC: 96V
OD: 95.6V

[Paul Birkeland]

You have a substantial amount of bias voltage appearing on both sides for the 6922 that's not really permitting it to draw sufficient operating current.  You have 3V of bias on both channels and I'd expect about half that.  Do the IA and IB voltages at the front C4S board still appear to be around 220V? 

I would next be looking at the little jumpers that go to pins 4 and 5 on the 6922 socket, as well as the voltage between the + and - pads on the 6.3V supply of the power supply board.  Both your channels are experiencing the same issue, so we would ideally start with what they share first.

[ohshitgorillas]

Actually the right channel works just fine... it's only the left that becomes audibly affected. I'll check those pins tomorrow regardless.

IA: 217V
IB: 217V

[Paul Birkeland]

Both your channels have the same issue, but I would imagine that the right channel 6922 plate voltage is just low enough to work properly. 

[ohshitgorillas]

Well it's taken me forever to get around to this, but I did eventually check on these.

I measured 6.25V all the way through on the 6.3V power supply, at the PSU board and on the EF86 tubes... but when I measured the individual voltages, I am getting 50.8V on the -6.3VDC pad and 57.4V on the +6.3VDC pad. I've poked around here but don't see anything obviously wrong... anything I should check in particular?

Another voltage check for sanity:
IA: 215V
IB: 214V
OA: 186.3V
OB: 208V
OC: 92.6
OD: 90.5
OkA: 94.8V
OkB: 93.3V
OkC: 1.295V
OkD: 1.81V

[Paul Birkeland]

The 51V and 57V are normal readings.  I would try a different 6922.

[ohshitgorillas]

Thanks for your help. I only have one other 6922, I tried before and got the same left channel clipping, but didn't check the voltages. Those are below:

IA: 216 V
IB: 215 V
OA: 181 V
OB: 201 V
OC: 97.9 V
OD: 95.8 V
OkA: 96.3 V
OkB: 96.7 V
OkC: 1.372 V
OkD: 1.84 V

[Paul Birkeland]

I would suspect that there's an issue with the connections on pins 4 and 5 of the 6922 tube socket that isn't allowing that tube to heat properly, as it doesn't seem to be properly conducting. 

[ohshitgorillas]

Pins 4 and 5 of the 6922 socket look good to me... but while checking I noticed that the left output resistor on pin 6 is fractured. I can also see some corrosion around and through the heat shrink jacket around the resistor on pin 1. I've attached some photos... please forgive my decision to use a non-standard tube socket as I wasn't aware the rotation of the pins was different from manufacturer to manufacturer.

I'm also noticing, thanks to these photos, how much uncleaned flux is on these solder pads. I should use my phone camera for this more often...

Edit: While some isopropyl and a toothbrush cleaned up the crap around the right channel/pin 1 resistor jacket, it caused the left channel resistor to break clean off. Going to begin the process of hunting through my stuff for a replacement and I hope I have that value. Edit2: I do not.