Well duh... Meet the Tube Doubler

[R.G.]

I chewed on myself most of the way home from checking on the cattle this afternoon.

Those of you who have been here for a while will remember the Mu-Doubler and its Achilles heel - you have to have matched FETs.

Why this never struck me before astounds me, but the obvious applies - two triodes in a single envelope are usually better matched than JFETs or MOSFETs. They do vary, but it's not the huge differences you can get with semis.

One ought to be able to make up a Tube Doubler (dibs!) with a single duotriode like the 12AX7 in the MuDoubler configuration and get nice octave doubling. The test circuit I put into the simulator has a 100K shared plate resistor, 1K shared cathode resistor, 47uF cathode cap, 1M grid resistors and a phase inverter feeding it. Produces nice little distorted doubles (octaves) with some higher order junk. Prime candidate for feeding from a back-to-back transformer setup with B+ at 150V.

Looks like all the other tricks like balancing, SRPP loading, etc. work fine too.

Gotta breadboard that one.

[Paul Marossy]

Sounds very interesting! Please keep us posted...  
When you say back-to-back transformer, do you mean a la McTube style?

[puretube]

 :wink:

(another thread I can`t contribute to...)

[Ansil]

haaaaaaaa  rg you beat me too it as usual man i ran across the article i printed a year or more ago as i lost the original one.

i was looing at it and thought "man i got to make a tube version of this." and then i was like. hmmm must have problems with it or rg would have done it already.  good to see i was wrong.

[R.G.]

The breadboard version works. The output waverform is not a true doubling, like with MOSFETs and JFETs. That puzzled me a little until I remembered that triodes are not square-law devices like channel FETs are. They are 3/2 power devices.  The result of running in a signal, then cancelling the signal and keeping only the distortion is not purely a squaring of the signal, so there are other distortion products, but the octave harmonic is definitely the biggest contributor. That may be good - some of the experimenters with the Mu-Doubler have reported that the octave is a little too soft, and needs some distortion after it to thicken up the signal.


I guess I gotta try pentodes next.

[Marcos - Munky]

Cool!!!

[Ben N]

What a neat idea.  I ove seeing stuff tubes are STILL uniquely suited too besides the usual (soft clipping, graceful overload/recovery, etc.)

RG, what kind of splitter are you using, and how are you driving it?

I guess I gotta try pentodes next.

Maybe this is (yet another) dumb question, but why?

Ben

[R.G.]

what kind of splitter are you using, and how are you driving it?

Opamp follower followed by an inverter.

in re pentodes...

Maybe this is (yet another) dumb question, but why?

Just to round out all the bases. Bipolars do a full wave rectified signal in this configuration (Univox Superfuzz); JFETs and MOSFETs do a relatively pure octave; triodes do a non-pure octave; pentodes are likely to do yet another type. I suppose I could look up the equations and gues, but why not just lash one up?