Bias drift or runaway Help needed!

[Neil]

I just bought a Fender twin reverb ri. Its quite old so I replaced all the tubes (pre and power) with a matched set of JJs from Eurotubes.

I set the bias to around 39 - 40 ma. All the tubes were close to each other. Using the amp shortly after I noticed it was very hot to the touch. I rechecked the bias and found that it was around 44 ma but when I left it for a while (15+ mins) it would start to creep up. After a while it reached the mid 50's and was still climbing. I kept backing the bias down but it would keep climbing. I turned off the amp and allowed it to cool before rechecking the bias. I got a reading of 6ma and it soon started to climb back up towards 30 and up.

I switched off and reinstalled the old fender groove tubes and set them to a nominal 30ma. After a while they too started to climb (although maybe a little slower).

To check bias I use a biasrite tool and a fluke DVM set to the 300ma dc range. All tone and volume controls to zero, reverb off, nothing plugged into the input. Speaker plugged in as normal. I leave the amp to warm up for a few minutes before turning on the ht.

First off, is this normal? (I cant see how it can be). If not then what do I need to check? From my rudimentary knowledge it seems as though either the bias supply is unstable or there is a signal coming from somewhere (which presumably is DC or the dvm wouldnt read it - leaky cap?).

Can anyone offer some advice?

Many thanks
Neil

 

[Wakarusa]

When the quiescent current starts to drift does the amp still sound normal (i.e. quiet), or does it start to hum/squeal?

Since you're measuring current in the tubes and not actual bias voltage the problem could be a loss of bias or just as easily the amp oscillating.

First test: monitor bias voltage while things are drifting. Does bias drift too, or hold pretty steady?

-- standard disclaimer --
you have to open it up to measure C-. There are lots of ways to die when you do this. If you aren't comfy hanging out with high voltage, get thee to a tech.

 

[Neil]

Hi,
Thanks for the interest. It sounds the same when the bias starts to drift. No strange noises. I tried it again with the PI tube removed (to decouple the preamp) amd the same thing happended. I havent measured the bias voltage yet. I guess that is the next thing.

THanks for the warning about high voltages. I may regret saying this but I am ok with them. I used to work for Mullards making electron optic devices that worked at around 6kv. I then worked for English Electric valve working around devices that ran at similar voltages but where they made klystrons that we would spot knock at over 1 megavolt. It scared the c@#p out of me then and it does now!

I'll get back when Ive tracked the bias voltage
Cheers
Neil

 

[Neil]

Well I checked the bias voltage on pin 5 of the tubes and it is steady. It drifted less than 0.5 volts in over an hour (and it got more negative which would reduce idle current).

The only difference between this test and the previous tests was the reverb was unplugged (so the chassis could be slid out) and there was no biasrite installed.

Unless I have missed something either there is
- a dc signal getting to the tube resulting in an apparent increase in bias current,
- some contact problem between the tube socket and the tube that is allowing the tube to drift away from the bias voltage. (on all 4 tubes?)
- there is a problem with the biasrite (I contacted the manufacturer and they claimed a problem of this type is not possible)
- there is a problem with the meter used.

Any other ideas?

Thanks
Neil

 

[Wakarusa]

One item unclear from your last post --
With the biasrite not installed was there any evidence that dissipation in the tubes was rising? This is a tough one cuz if the plates aren't glowing there isn't much to see. Two thoughts:

1) There's a definite possibility that the PI coupling caps are leaking DC, but if this were a steady leakage it would tend to offset the bias voltage and be visible on your meter as a drift in C-. When the leakage is intermittent or varies a lot it is more difficult to see on a DMM (an o-scope is really handy here), but usually makes noise that you can hear -- something you say isn't happening.

2) Keeping in mind that this is remote-controlled troubleshooting, I'd look for oscillation in the power section. It's something that can cause power output to rise (as seen in your biasrite) and may or may not be something you can hear. Also, it is possible that the biasrite is causing/contibuting to the oscillation (if that's what's going on) because its installation lengthens the leads from the grid stopper resistors to the tubes -- possibly just enough to reduce the quenching effect of the grid stopper and allow oscillation.

Next things to try -- solder a 1 ohm resistor between the power tube cathodes and ground. Measure voltage on the cathode pin with the amp operating and the biasrite NOT installed. If you have exactly 1 ohm resistors, then mV = mA. This is exactly what your biasrite does but eliminates any foolishness (longer leads, etc.) in the grid circuit.

If the cathode voltage rises as before then its time to dig out the o-scope and go poking (or time to see the tech). If the amp doesn't run away as before, then I'd suspect the biasrite's effect on the circuit.

Oh yeah -- one more dumb thing to try before we dig out the soldering iron: does the amp exhibit the same runaway with a different set of power tubes installed?

 

[Blue Strat]

Do all 4 power tubes exhibit the same behavior? If so, the amp or bias probe are definitely at fault. If not, you have one or more defective tubes.

The bias supply cap (as well as the PI coupling caps) is another suspect. If this cap is leaky (not visibly leaky necessarily) or installed backwards, your bias would drift.

This seems unlikely if the voltage on pin 5 is relatively steady. Note that a change of even one volt on pin 5 can result in a several MA change of the power tube idle current.

Mike K