Mysterious power tube behavior

On my Willsenton R8, I was using JJ 6L6GCs. One of them got a bit noisy, and then soon after startup the bias would go to zero; bias in the other positions remained steady. I switched it to different sockets, same behavior. I threw in a set of known good KT77s, all 4 positions bias fine. Put the 6L6s back in, that one tube still caused bias to go to zero.

The mystery is that sound quality, other than the noise, was never affected. And the misbehaving 6L6 tested fine on my recently calibrated Hickok 600; the reading were exactly the same as the other 3 tubes.

Anybody got any idea how a tube with zero bias could act perfectly normal? Visual inspection of the tube didn't show any differences either. I've been using tubes a long time and I've never seen this.


I am no tube expert but did have what I thought was a tube that hummed. It’s on a secondary system in my office and not loud enough to annoy me or be heard from my desk, but the audio was fine.

After reading your post I just walked over the amp and moved my ear around the whole thing and realized that the noise was originating from the transformer and not from any of the 8 tubes!

That tube probably have bias circuit open. Therefore no current you can read there. The reason that tube works is simply because for push-pull amps most tubes are "cold biased" to like 50mV anyways. 


So the tube has an internal bias circuit that could be malfunctioning and the tube would still test good? The bad tube shows no bias in any of the 4 sockets, and a known good tube biases fine in all 4 sockets.

Only if it's in PushPull setup or "cold bias". It won't work single-ended. 

I would test an input resistor usually 1k and it's between ground and pin5 on bad tube. In case with KT77 it should show value, but if in case there's no value in 6l6, then you know that control grid is compromised there within the vacuum body. I'm just hesitant on why tube tester scores a good tube in that case. 

If the bias reading follows the tube then it's not the bias circuit -- the tube is bad. The fixed bias bias circuit uses a 10 ohm resistor at each cathode and the voltage is measured between cathode and ground (the voltage is converted to milliamps by multiplying by 100). So if it's a zero reading then there is no plate current. 

As to why the tube "tests fine" on the Hickok... I can't say unless I know what you are measuring. A Hickok 600 is a mutual conductance type tester and not an emissions tester. Emission testers sometimes use a low voltage low current power supply to do a static test and give a "Good-Bad" result only. The test current can make the tube work but when it's installed in the higher power supply current and voltage of the amp it doesn't work.

Also, I don't think you are hearing the bad tube. It's partner may be doing all the pushin' and pullin'. A 6L6GC tube will swing the full voltage signal at about 2 watts and under, which is plenty of juice for moderate listening levels on 88dB and greater speakers.

If that tube is bad, do not use it. A dead tube will drastically change the load impedance on the output transformer, possibly swinging the voltage over maximum dissipation which can damage the other tubes.

While we're at it... in fixed bias designs, both the 6L6GC and the KT88 can tolerate a grid to cathode resistance of 100K. If for some reason that resistance is a lot higher, it places stress on the power tubes and may be a factor as to why the tube went bad. To check, simply remove the power tubes and measure the resistance between pins 5 and 8. If it's above 125-130K then you should contact the manufacturer (damaged tube pin, bad solder joint, bad potentiometer). If all power tube sockets measure like that or higher, then shame on the designer.

Replacing the tube is a must.