Tube problems 6922

Otis: I would post this question at Audioasylum.com. There are a lot of DIY's that start from scratch at the site and that should be able to figure out what is going on in your situation. Many (a high percentage) of the 6922 type tubes that I have tested by ear recently have problems in the 60hz to 250hz region, though I hear them as being peaks in the frequency response (which you may be hearing as a drop out at the edges of the peaks). Static (or rustling) is also quite common with these tubes when they have some miles on them as well as microphonics on the softer sounding types. You should be able to sort out the weak tubes by dropping the heater current from 6v to 5v and seeing how much the mu drops per side. A drop of over 8-10% is a weak tube from what I recall, though I have not tested a tube myself for over 25 years and may be rembering this incorrectly. Anyway, give the AA site a try under the proper category.

Thanks for your reply, one thing I forgot to mention is plate voltage. The tester has two ways to test, with plate of @ 90v and @ 180v, The preamp runs at from 115 to 137v on the plate. The drift shows most often at 180v, and to a lesser degree in the preamp and seldom at the 90 volt level. Further thoughts?

Otis: Perhaps you picked up a batch that have already been pre-sorted (rejects in other words). I have always wondered where the less than stellar tubes from the tube dealer's inventories end up though I have seen some of them some sold on ebay at reduced prices.

Additional thought is "gassy" tubes, although I can't guarantee that would cause regular variations. While monitoring the filaments in reduced ambient light, also note if there is any visible ionization in the bad tubes compared to the good, just for kicks.

I might guess a thermal effect - perhaps either:

Control grid position or geometry changes, thus altering plate current, or

Connection problem with the filament, causing it to separate when hot but reconnect quickly as it cools, thus varying emission.

Perhaps you could watch the filament carefully to see if there is any visible brightness variation.

You didn't say if variation was fast or slow, but I am presuming slow. Interesting that so many should have that characteristic, though.