Tube mixing


So what would happen (theoretically anyway) if in a "push pull" tube amp where each tube has its own bias pot, you stuck disparate tubes next to each other. I have 3 sets of power tubes...if I put 6550s and KT120s next to each other in each channel biased properly (or to the amp specs), or a KT88 next to a 6550 in 1-2-1-2 pairs...would this trigger cosmic calamity? Gear meltdown? Audio magazine subscription rejections? Would the tubes sit there and stare at each other in disgust? I'm not necessarily gonna try this, but it seems to be a way to get the advantages of different tubes ALL AT ONCE! Ha...and hmmmmmm....
wolf_garcia
Wolf!! Disparate tubes next to each other? Un-matched quads in a tube amp. Don't do it!. You will upset the cosmic tubeworld balance. I'm going to short circuit as well. (I don't think I will ever be as funny as you.)

Anyway I'm curious to know what membership has to say as well. In the other tube amp I have there are 2 6L6GB tubes and 2 6V6GT tubes used together. The amp is supposed to have a quad of 6L6 tubes. From what other tubeophiles told me the split set-up I have in that amp is ok. I just have less power/channel. I never got around to having a quad of 6L6 tubes put in because I had to pay a tech to re-bias and check out the amp.
I'm talking about one of each in each channel...I think somebody has to ask these questions. No question that 2 matched pairs of different tubes in each side is no biggie and results in what you described, but what I'm proposing is much weirder, and probably illegal.
Here's the problem with different tubes on one side of a push-pull amp.

I assume you're familiar that a push-pull circuit divides the signal into two halves (usually shown as a top & bottom referenced by a center line), amplifies each separately, and rejoins them at the output.

Using a simple sine wave input where the top half of the wave is a mirror of the bottom, a 6550 tube is going to have "X" amount of gain while the KT88 or KT120 will have a different amount. At the least, your output push-pull wave will no longer be a more powerful duplicate of the input push-pull signal. You will have distorted the wave.

Depending on circuit design, this could also make the amp unstable.

Even if the amp is designed to accommodate all the tubes involved, I wouldn't recommend this kind of mismatch unless the situation was desperate (i.e., an audience waiting for a performance and no other possible way than this for the show to take place).
Wolfie, if the Large Hadron Collider is in use at the time you try this, the result will be a black hole that will eventually suck in the entire known universe.

If the LHC is not in use at the time, I agree with Mlsstl that the result would be a significant increase in distortion. The positive-going and negative-going halves of a symmetrical sine wave, for instance, would most likely emerge no longer symmetrical, as a result of differences in transconductance (Gm) and/or other parameters.

The same thing would apply to using a 6L6 and a 6V6 together as a push-pull pair, probably even more so (i.e., even greater distortion, and/or more rapid destruction of the universe), as Gm and several of their other parameters differ widely. On the other hand, while using a pair of 6L6's in one channel and a pair of 6V6's in the other channel would eliminate the possibility of interaction with the LHC, it would most likely result in a significant sonic mismatch between channels.

Best regards,
-- Al
Wolf, I could understand your consideration of this tube complement (?) implementation if you still resided in the land of Maui Wowie, but what is your excuse now...?
...but it seems to be a way to get the advantages of different tubes ALL AT ONCE!
That's one of those ideas that sounds great if you say it fast enough.

Rather like deciding your next car should have an 8 cylinder engine with 4 cylinders from a Chevy and 4 from a Ford to take advantage of the strong points of each.

That's usually not a good engineering approach. ;-)
Hey...I'm not planning to try this as I assumed there would be good reasons not to but somebody had to ask (actually nobody HAD to ask), and now I know and I feel I've learned something. By the way...I'm sure some near sighted tube head has accidentally stuck similar but different tubes in something, so I feel this thread has performed an important public service.
I think the impedance mismatch alone may be a problem, causing either the push or pull side to have a different load. Even with the same bias load, the frequency response can also be different enough, to possibly overheat the output transformers. I've used two different brands in each side (same number), for a blend that I liked better. This was as much of a difference I have tried. But, I wouldn't risk a totally different tube in each push-pull side.