Sovtek MIG50 odd power tube draw

[trevorus]

I opened up my mig to check out a crackle, which turned out to be a loose solder joint. I decided to also add in my 1 ohm resistors for biasing. I do all this, and go to check out the bias, and I get 32mA at 583 volts on one tube, and nothing on the other. I check the draw with the shunt method, and get roughly the same on the one tube, and about half on the problematic one.

Of course, I switch tubes between the sockets. The issue persists with the socket, not the tube. So I clean the sockets, a spray of contact cleaner, and I tighten them up. Test again, same issue. I break out the oscilloscope, and check signal on both sides. Roughly the same signal showing up on either side. There is a bit of variation on the OT between the sides, but the amp draw mismatch could be the culprit. I still cannot find the cause of that, though.

Please help me figure this little beast out! Thanks, all!

[Pedro Vecino]

I would check:
- If the value of the new cathode resistors is the same
- If they are well connected to ground
- The value of screen grid resistors

Regardless, with this voltage I would install a 100V/5w zener in series with the screen grid resistors to reduce it and adjust the bias again.

[trevorus]

I would check:
- If the value of the new cathode resistors is the same
- If they are well connected to ground
- The value of screen grid resistors

Regardless, with this voltage I would install a 100V/5w zener in series with the screen grid resistors to reduce it and adjust the bias again.

Screen grids are both on spec, and the cathode resistors are fresh 1% ones. I may change them to wire wound, because they're better at handling the heat. Screen grids are 1k wirewound 5 watters.

I may also try returning it to the previous wiring and retest.

[g1]

Try measuring resistance from the cathode pin of the socket to chassis, in case there is a grounding issue.

[trevorus]

Try measuring resistance from the cathode pin of the socket to chassis, in case there is a grounding issue.

I'll test from the socket side, since I've got good connection from the solder joint to the ground through the one ohm resistors.

[trevorus]

Shoot, good connection on the socket. Time to wire it back the way it came, and make sure I didn't introduce the issues...

[trevorus]

Bias voltage on both sides is stable and consistent, but it still has a big mismatch. I'm wondering if a bias balance would be in order, or there is a big mismatch in the OT. The resistors between the bias supply and the tubes are all consistent and on spec... This is getting really confusing. The closest I can get is 35 mA and 25mA. I'm wondering if a variable resistor on a split bias supply would be in order...

[g1]

check out the bias, and I get 32mA at 583 volts on one tube, and nothing on the other.

I check the draw with the shunt method, and get roughly the same on the one tube, and about half on the problematic one.

The closest I can get is 35 mA and 25mA.

This is a little weird, as it seems things are getting progressively better? First no idle current on the "low" tube, then 1/2 current, now within 10mA.

Do you now get the same readings when swapping tubes? Maybe try one tube only at a time, tube A then B in socket #1, tube A then B in socket #2.
Previous readings suggested the tubes were perfectly matched, this should help verify if that is the case or not.

[trevorus]

Yeah, it was a bit inconsistent. The off readings are consistent with the socket, regardless of the tube that's put in it. That's one thing that is really odd to me. I'm biasing on the OT shunt, and eliminated the 1 ohm resistors entirely.

One thing that's changed a bit is accidentally zapping my pilot led by shorting B+ a little too near it with the ammeter... I'm keeping one hand in my pocket, for good reason.

[trevorus]

Ok, now I need to dig deeper. The hot tube is the only one amplifying. This thing is driving me nuts!

Edit: Preliminary diagnosis, a lifted/broken trace on that tubes screen resistor!
Edit two: Problem solved, bias in spec, and it sounds wonderful again. Thanks for letting me bounce ideas off of you g-one and Pedro.