a push/pull tube o.d. for the summer...

[Johan]

I always thought push pull sounded better than single ended tube amps,  so here is something I've worked on for a few Days. a push pull tube od using a 12au7 running at 9 volts for simplicity.
the leds are not for clipping but to prevent the transistor phace inverter and the opamp from clipping, should you use a boster Before it.
 the super-bal opamp stage replaces the need for anoutput transformer


http://www.aronnelson.com/DIYFiles/up/ppdriver.mp3
its played through line6 ampfarm on a blackface fender model set Clean. output is set so only the od distorts. gain on full. all gain/tone Changes are done from the guitar
(if you can bare the playing) enjoy
J

[nocentelli]

playing is great, the sound is great.

[samhay]

Cool - sounds really good.
Is the 10k impedance of the output op-amp stage chosen to load the valves down - it seems a little low?
I haven't seen anyone use the heater tap to generate a reference voltage before. Nice idea.

[Johan]

The low z on the grids is to prevent the grids going negative when starved for voltage. Paia has a nice write up on that phenomenon.

[Johan]

..Oh sorry. Misunderstood. ...i initially had much smaller plate resistors and when i changed them i left the opamp as it was. It worked and i didn't worry about it. .
J

[deadastronaut]

sounds great man...love it, very responsive, cool playing too 

[Seljer]

Ooh, I  like the idea. I just ran a some simulations of the same concept but with JFETs

The blue and green are the voltages on the JFETs' drains, the teal is after the opamp subtracts one from the other.



It gives some mighty fine looking soft-ish clipping, more soft than just running the FETs themselves. Also, if anything is mismatched (the JFETs, the source capacitors, etc...) you start getting crossover distortion but I guess thats just part of the class AB power amp 'mojo'.

Your original setup with the opamp and gain control in the start is probably a better topology. It probably also needs some more work on the EQ of the stages. I'll definitely try at least breadboarding it sometime this summer

[Johan]

Please note that the common cathode resistor(without bypass cap) gives a litle bit of negative feedback in the tube stage

[Luke51411]

Very nice! Definitely adding this one to the list and following this thread intently!

[samhay]

The low z on the grids is to prevent the grids going negative when starved for voltage. Paia has a nice write up on that phenomenon.

Yeah, I figured you knew what you were doing when I saw these values.
Merlin's writen a nice piece on running triodes at low voltages too: http://www.valvewizard.co.uk/Triodes_at_low_voltages_Blencowe.pdf

..Oh sorry. Misunderstood. ...i initially had much smaller plate resistors and when i changed them i left the opamp as it was. It worked and i didn't worry about it. .
J

Fair enough. I guess without a cathode bypass cap, you need fairly large anode resistors.
I assume you tried with a bypass cap?

[Johan]

I did. It got brighter. But not in a good way. ... my original plan also had a tone control in the form of a 100k pot and cap between the inputs on the output opamp but i never got around to figure out the cap value. ..

[samhay]

Did you try running it single-ended? Perhaps a 10k pot across the feedback path of the second op-amp would let you blend in the push-pull effect.

[Johan]

No i didn't since the whole intention was a push pull.  By it would be a Breeze replacing the bias resistor on output opamp with a pot and a large cap between bias and wiper

[samhay]

Indeed - will add this to the 'to breadboard' list.
Thanks Johan.

[Frank_NH]

Wow - this sounds really great (as is your playing)!  I have some preamp tubes sitting around in my parts collection - may have to try this out.  Anyone up for a vero layout?? 

[Bill Mountain]

I have considered doing this with CMOS stages for quite some time.  Maybe even with a small transformer.

Looks pretty cool.  Something else to add to the never ending circuits-to-do list.

[Johan]

One more thing that just hit me. ..The reason for the leds is to prevent the pi from clipping. I've seen transistors in this setup invert the peaks instead of just flattop which sound particularly ugly. But this transistor could be any three legged critter you find around your bench.  npn, pnp, si, ge even fet....Could be something worth exploring for anyone bread boarding this. ...

[GFR]

If you add an add/subtract switch on the output stage, maybe a bias pot, you would have (kind of) a tube version of the Univox Superfuzz 

[samhay]

Have had a play with this concept on the breadboard - jättebra Johan. A very nice alternative to a Valvecaster.

Cathodyne clipping can get pretty ugly, so I figured it might be nice to use a valve to do the phase splitting rather than the BJT. I ditched the long-tailed pair and went for a valve gain stage into a valve cathodyne, which should have more gain anyway. Kept the differential op-amp to do the final summing.
Rather than mess with the final summing/mix, I used a cap/pot to AC-bypass the cathodyne cathode resistor. This lets you go from push-pull to single-ended, which offers some nice differences in tone. The push-pull certainly sounds very nice, so my hat off to Johan.

I don't want to derail the thread too much, but will post an in-progress schematic if there is any interest and Johan doesn't mind.

[Johan]

Thanks Samhay
Yeah. Cathodyne pi is not a favorite of mine either. But as long as it's kept under Control, as i did limiting the level with leds. It's hard to beat as far as simplicity vs funktion goes.
..also. Since the tubes will start clip far before the leds. Cleanup will always be more tubelike than the usual diode kink allows, even if the gain in the first opamp stage is dramatically increased....i will redraw and rescale the schematic too so that film caps can be used throughout and incorporate some of the things mentioned earlier in the thread.  
..so. Now the the concept is proven. Who will be the first building a low voltage Marshall super lead using 12au7 and a lm3886 as output?
....(That's actually how i came up with the concept. Designing a hybrid power amp for myself with a long tailed pair input driving a power ic..)
J
And no. I don't mind People adding their own ideas/versions... That's what i loved about this forum when i first joined. ...discussing and expending on each others work