Bias radically changed by itself

[daz]

This happened the other day and i can't figure it out. The bias of my 6V6 build went from 18 mV to 9 mV. The amp sounded lousy and i have bias points and pot on the back of the amp. BOTH tubes went down so it's not a tube issue. The pot is a screwdriver type shaft so i couldn't have turned it accidentally and haven't even been back there for a while when this happened. I rebiased then pulled the chassis and tried turning the amp on with the meter hooked up showing the bias and chop sticked everything related to the bias so i don't think it's a bad joint. Only think i can imagine is a cap, but can a cap do that intermittently and work correctly most of the time? This hasn't happened but once. Checked the mains too and yes, the power was low but thats a daily event and 1. it's never affected bias radically like this before and 2. i have been checking it since and the bias hasn't changed more than a fraction of a mV since even tho the mains have changed back to full 120+.

[Leo_Gnardo]

Only think i can imagine is a cap, but can a cap do that intermittently and work correctly most of the time?

A cap that's subject to overheating can do all sorts of mischief. I hope that's not the case with your bias filter cap.

Let's hope the gremlins have departed your amp for good now.

[daz]

I wish it were, but only if i could know it was for sure. Then i could replace it and know i'm good to go. The intermittent stuff is problematic.

[nsubulysses]

The last time I found myself in this scenario it was because I was using used tubes to get someones amp going because they needed it for a show and I had no 6L6 in stock. I thought the amp was the problem but it was the used tubes.

If the bias remains unstable you might try different power tubes (I know you said it's not the tubes) to see if there is any difference. Or you have to monitor all DC voltages at the tube socket to see if any are changing. Smash the amp a lot with your hand too (or rubber mallet) in case there are intermittent connections

Also verify your bias measurements with an additional measurement just in case. Cathode resistor, bias probe, transformer shunt method, etc.

I have also installed bias pots (little trimmers like this) in amps that I've noticed I can tap with a chopstick and make the bias change. Multi-turn pots are better for stability since it takes 12 (or however many) full rotations to exercise the full range of the pot.


Just throwing some things out there.

[tubeswell]

What does the bias voltage look like when the the tube current drops off?

[Enzo]

And does the screen node B+ drop off at that time?

Just because you can poke the thing and not see a change, doesn't mean ther is not a bad solder joint. Some such only appear through thermal expansion and attraction. Poking moves things but does not change their size.

[Enzo]

And not to be overly picky, but the CURRENT changed, we don't yet know if the bias did or not.

[olddawg]

I wish it were, but only if i could know it was for sure. Then i could replace it and know i'm good to go. The intermittent stuff is problematic.

Is the cap that expensive or inaccessible? I would just replace it. At the same time.. get out the magnifying glass and bright light.

[Leo_Gnardo]

Is the cap that expensive or inaccessible? I would just replace it. At the same time.. get out the magnifying glass and bright light.

Another thought, I'd expect a failing cap to lose voltage, either by becoming conductive or losing its uF value, and drive bias current up, not the other way around.

[g1]

Agree with nsubulysses, tubes are usually the first and easiest thing to check.
Many times one bad power tube will pull another down with it, so you can't rule it out just because they both show symptoms.

[daz]

Lots of questions and suggestions but so far it hasn't happened again so maybe it was a fluke and i'll just keep monitoring it. The bias check i do is DC mV across 1R cathode resistors. Maybe I'll try chopsticking it again and this time tap the bias pot.Forgot to do that last time. I'll also chopstick the B+.

[Dave H]

The bias check i do is DC mV across 1R cathode resistors.

That's measuring the bias current. I'd also measure and record the bias voltage i.e. the voltage from the bias supply at the junction of the 6V6 grid leak resistors. It will be around -30V or so. Now if the bias current shifts again you'll know if it's the bias supply (pot etc.) or something to do with the tubes or screen voltage.