I was recently given a pair of used vintage Tungsol 6550s. They look nice with good getter and even test well on my (admittedly limited) tube tester (also made by Tungsol).
With my recently built amplifier, which is a single-ended pentode design with fixed bias, the tubes will stay steady at 60mA/400V. However, if I bump up to 70mA, one of the tubes will eventually start drawing more and more current until it pegs the meter (above 100ma). By that time I'm frantically shutting the amp off, hoping to get there in time before the 10 ohm cathode resistor or fuse pops.
Of course I don't know the fully history of these tubes, nor the type of abuses they encountered in the past, but I am curious behind the physics of a runaway tube.
I've had this happen in the past with a different amplifier using a NOS 1625 with self-bias, but I assumed that was from a gassy tube.
With my recently built amplifier, which is a single-ended pentode design with fixed bias, the tubes will stay steady at 60mA/400V. However, if I bump up to 70mA, one of the tubes will eventually start drawing more and more current until it pegs the meter (above 100ma). By that time I'm frantically shutting the amp off, hoping to get there in time before the 10 ohm cathode resistor or fuse pops.
Of course I don't know the fully history of these tubes, nor the type of abuses they encountered in the past, but I am curious behind the physics of a runaway tube.
I've had this happen in the past with a different amplifier using a NOS 1625 with self-bias, but I assumed that was from a gassy tube.
Grid leakage in some form - gas, grid emission, contamination. But there is always SOME grid leakage. Max grid return resistor for 6550 is 50K with fixed bias - you can get away with higher if dissipation is kept down.
I agree it sounds like grid leakage. A lower value grid resistor may be called for, as noted above.
Yes. Could be due to gas (released when the electrodes heat up) forming positive ions or the glass becoming conductive (due to electrolysis?).
Ah yes, grid contamination. That is another cause of thermal runaway. If the grid gets too hot then it can start emitting electrons - especially if during manufacture or use it managed to pick up some cathode material. Gold plating the grid can reduce this - one of the few places in audio where a little gold plating is actually useful!
100K
Since I'm working from home today, I plugged the 6550s back into the amp. Set the bias at 50mA - having meters on board is fun, btw - and monitored. Eventually one of the tubes went past 100mA, while the other stayed at 50mA.
I re-adjusted the bias for hot one back down to 50mA. A few moments later, it was drawing 20mA.
Tung Sol specs 50k max for the grid leak. That's sort of a tacit acknowledgement that grid current is an issue... You might try dropping the value, just to see.
Ok then.
What makes a tube run away is finally meeting a "really hot" tube.
Or just plain fear.
</humor>
What makes a tube run away is finally meeting a "really hot" tube.
Or just plain fear.
</humor>
- Status
- Not open for further replies.