An old Mullard EL84 still running at 70 mA?

[overtone]

This is a question about an older Mullard EL84 which exhibited, what for me is, a strange behavior. I would like to understand what is happening, so I hope it is OK too post here in the theory section.

I was checking my stash of old 60s / 70s EL84s with an adaptor thingy and the 1 Ohm resistor method to get an idea of what has accumulated here. I gave each tube five minutes to settle down and watched them stabilize at between 36 to 52 mA, measured at the cathode.

The Mullard tube in question looked like it would stabilize at about 48 mA, but it just did not settle down and the reading kept on creeping up. At 60 mA I was wondering what tube I could match it with. At 65 mA the paralleled tube to the right of it started showing slight signs of red- plating. At 70 mA I switched it off thinking that it could be an evil tube sent to unleash hell's fire in my living room.

My impression was that the increase in mA was exponential. The photo, from a second try, is taken at about 65 mA. For some reason the area on the second tube, which I perceived as the beginings of red-plating, shows up as lilac-blue on the photo. The evil tube looked very content all the time, not even a blush.

So my question is not, are some tubes bad, but what is actually happening? Is this what happens when a tube “runs away”? Or is this normal behavior for a tired tube?

(The amp is a home cooked Rocket / AC30 circuit running with 11.8W plate dissipation. The codes on the Mullard are "?x1 B6D BRITISH )

[pdf64]

This is with a cathode resistor shared between 4 tubes, plus 1 ohm in series with each cathode?

[overtone]

It is a shared 50R 25W cathdoe resistor for those four tubes.
The 1 Ohm resistor is only on the adaptor, nearest you in the picture.

[MWJB]

Just a thought, some bias probe adaptors cannot take high idle currents for a sustained period of time, I found this out whilst trying to set idle current on a pair of KT90's set to 60mA per tube...reading settled for a while then crept up...& kept rising! I assumed a fault with the amp...and I would like to say that I got to the root of it before I changed the PI coupling caps in a fit of panic...well, that's what I'd like to say anyway! ;-) The bias probes, when cool, would give the correct reading, but it drifted as the probe's 1ohm resistor was put under load & heated up. AB'ing with the shunt method showed no rise in idle current.

There's a big difference when you square up 40-50mA compared to 30-35mA.

[Steve Conner]

It's probably an old, tired tube, going into thermal runaway.

[ThermionicScott]

Is this what happens when a tube “runs away”?

Yep. Gas in the tube is being ionized by the electron stream. Positive ions are attracted to the negatively-charged (relative to the cathode) grid, causing it to become less negative. This reduces the bias on the tube, allowing more electrons to flow and ionize more gas molecules. I think you can see where this leads.

A properly-manufactured tube will have very little residual gas inside, but over time, gas leaks in through the pin seals or is liberated from hot tube elements. (That's one of many reasons to heed the maximum plate dissipation and bulb temperature limits on the spec sheets.)

In theory, the getter should be absorbing the excess gas molecules, but getters wear out and don't seem to be as effective during tube operation as when they are flashed at the factory.

The good news about cathode bias is that the bias ratchets up if the tube starts drawing more current, so it can hold an unstable tube in check to some degree. A good idea for your build would be to give each tube its own cathode resistor, so that "cool running" tubes won't encourage a leaky one to run away.

- Scott

[overtone]

Had I not been measuring the current in the first tube, I would probably have assumed that the second tube is faulty because it was showing red, when in fact it is just caught up in the ride, like a run on the bank. „Mit gefangen, mit gehangen“ as they say around here.

Thank you for all the informative replies gentlemen. It lead me to some interesting reading today about current hogging and positive feedback in sheep herds...
Then I was going to ask “what about” separate resistors too, so thanks double.

[pdf64]

'A good idea for your build would be to give each tube its own cathode resistor'
I think that's far better than a shared cathode resistor. It allows quite unmatched tubes to reach their own balance. And is more robust, because if one tube dies, the others aren't forced to run hotter.
The minor downside is that more bypass caps are needed, and much higher values are required, because there's no cathode signal cancellation that occurs when the cathodes are linked. Pete.