deliberately unmatching output tubes

[porl]

i hope this hasn't been asked before, but i couldn't find it anywhere.

a while ago i was reading about the different sort of sounding distortion you can get when using deliberately unmatched output tubes. whether or not that distortion is a desirable one i don't know, and i imagine it is subjective anyway.

what i would like to know is whether or not anyone here has experimented with using completely different (but similarly speced of course) tube types in a push pull configuration. for example (whether this is a technically possible config or not i don't know, but just for illustration) using a 6v6 and an el84, or 6l6 and el34 together.

i know a lot of people will probably just say i'm being an idiot, but as they say, if you don't ask.... haha

thanks
porl

[Geek]

Hi,

6L6 and EL34 will likely "fight each other" to death, because they require different output impedances and will consume different current.

6V6 and 6AQ5 (not 6BQ5/EL84) will play nicely together because they are electrically identical.

6V6 and EL84, I can't say, because while they like the same load impedances, an EL84 is somewhat "hotter" than a 6V6.

I bet there may be something on this at AX84.com, as it sounds like something they may have tried.



Cheers!

[Steerpike]

Philips made some amps using a PL84 and an EL84. These two are slightly different in things other than heater rating, although Philips' main reason WAS the better heater insulation on the PL84, since in their Series Balanced Output, the 'top' valve has its cathode at a very high potential.

[kenpeter]

If you use separate CCS tails, and bridge across these two high
impedance tails with a suitably large cap. You can balance both
dissimilar devices to the same quiescent current, and this keeps
the output transformer happy..

Even if the cathode bias required to do so is different, this bias
difference can be abused to keep a large electrolytic formed.
I'm assuming you would parallel that with a quality audio cap.
The AC voltage component seen in both tails is then unified...

What happens in the most extreme case: Triode behavior one
side, MOSFET or Pentode behavior the other... Is that the pair
now becomes less differential, and more of a folded cascode.

The higher Gm device holds the AC potential of its cathode or
source at virtual ground (offset by DC bias of course). It will
pass any AC current swing required to hold both tail voltages
steady.

The lower Gm device (usually the triode) exclusively controls
the current swing for both ends of the push pull transformer.
The cathode current of the triode is precisely folded back in a
U, and brought up the other end of the transformer, through
the higher GM device (usually the pentode and/or MOSFET).

If GM are too similar: say 6v6 triode vs 6v6 Pentode, then you
don't quite get pure anti-triode action up the higher GM side
of the fold. You get something blended, neither pure Single
Ended Push Pull nor Complientary Push Pull. And this blended
curve is not necessarily a bad thing, it can be very pleasant.
Just harder to define exactly what it is???

So you got one device in almost exclusive control of cathode
Voltage swing. And the other in almost exclusive control of
cathode Current swing. Each communicating to the other via
its own dominant method of control...

Whats odd: Is that you can drive or ground either grid or gate.
Or drive both ends differentially. And yet the single ended (or
blended) push pull behavior of the folded cascode is the same.
I would suggest driving into whichever end has the least Miller...

The Caveat to this arrangement is PSRR. You don't get the
usual Push Pull cancellation of ripple currents in Single Ended
Push Pull. In fact, it inverts the problem and shoves it right
back up the other end... So you need a well regulated supply.

The opposite of Complimentary behavior is Insultramentary.
Welcome to the strange mirror world of the Anti-Triode! Where
your evil clone on the Anti-side of the listening glass (holding
the other end of the saw) has his back to you, and pushes or
pulls in the same direction...

[Miles Prower]

I've heard of guitar players doing this quite often. Sometimes it involves pulling some tubes from one side of a PPP pair. There was even an article I saw somewhere (may be the Radiotron Designers Handbook (?)) about putting a 45 in one hole and a 2A3 in the other one, this being an analysis on how bad mismatching was concerning distortion. The conclusion it wasn't all that disasterous, and that obsessive pair matching wasn't necessary.

Unbalancing a PP output stage will give even order harmonics as well as odd ordered. Might help guitar amp sonics.

[tubelab.com]

Coincidentally I was having a discussion about this subject just a few hours ago. The discussion was related to guitar amps and deliberate distortion generation. I explained how I used to purposely stuff a 6V6 and a 6L6GC into a P-P guitar amp to get a unique sound. Other possibilities involve a switch or gain control to reduce the drive to one tube without upsetting the DC bias. This way you can achieve some SE sound in a P-P amp without invoking saturation distortion in the OPT.

Understand that just plugging the wrong tube into an existing amp may be an invitation for disaster if it doesn't like the bias level set for the previous tube. In my case it was a guitar amp that I had constructed and I probably would have watched and cheered if it blew up, then fixed it, therefore nothing bad happened. The amp runs at a relatively low 375 volts due to the poor quality of the Chinese tubes available when I built it (about 15 years ago).

[smoking-amp]

"Insultramentary" P-P !!!!????


I think I follow the general idea Ken, but I'm not seeing clearly the way to impliment that one. Seems like it would just give the same SE effect from both sides, but with inverted drives.

Two normal triodes in regular class A P-P would seem to be closer to a "tug o war" P-P.


As Ken mentioned, the DC idle/ bias currents on both sides need to match and the two gm's need to roughly match (to keep the xfmr current balanced at high signal level ).

There is also a scheme around where dissimilar (usually triodes and pentodes) tube types are uses on both sides (4 tubes total) and the bias on one set (usually the pentodes) is set to cut in at a higher signal level.

[porl]

wow... i think i have opened a can of worms with this one... i was expecting 'it won't work' or 'yeah, it sounds pretty cool, but be careful'... insultimentary?? haha wow i think i am a little out of my depth

to clarify, i was thinking of doing this on a guitar amp, where the even order harmonics were what i was after. i was specifically thinking 6v6 and el84, but thought i'd ask more generally, as i want to understand as much about these things as possible.

what i love about valve/tube amps is the fact that something so simple in theory leads to such complex behaviour with such minute changes. the nerd in me relates this to chaos theory

anyway, thanks for all the responses so far!

porl

[Wavebourn]

The can was already re-opened many times. I had my Alligator running, an unfinished prototype with trioded GU-50 and MOSFET modulated source follower. Also, I drew a schematic for one of my friends in Russia, he made a guitar amp for his brother's son with couple of tubes, one pentode and one trioded pentode that shared common cathode incandescent lamp tail. It did not need a phase splitter, a pentode half was running with grounded grid.

[hitsware]

>incandescent lamp tail

mr. noise .....