Fake PI

[LeftyStrat]

treat a PI as differential stage, but grab only one side of the output.
might try putting a 1m pot across the outputs of one you have up and running and see what you get as a summed out, swing it back and forth between the sides.

Look at a Ampeg see saw with the fixed resistors in the plate circuit

Just had a chance to look this up. This definitely looks worth trying.

[teemuk]

Might I suggest though that you use a 12AU7 or other small triode as the "power tube" and use a reverb driver transformer for your OT. Then you could include the NFB network just like a full amp. Drive a dummy load with the OT--a 10 watt wirewound resistor should do fine.

You don't need a transformer, a differential opamp stage is enough.

There are plenty of practical examples of OP's idea...

At least Crate Blue Voodoo amps and Legend Rock-n-Roll amps use a basic differential with output taken only in single phase. You can wrap feedback around such circuit even without using transformers. Basically, the thing is the same as Laney's "KLIPP" circuit. It overdrives smoothly and quite symmetrically but a convincing tube power amp emulation it isn't.

http://www.prowessamplifiers.com/schema ... end_50.pdf


Vox Valvetronix scheme is even more advanced, it takes a push-pull output from a dual triode stage and combines it with a transformer or - like in later revisions - a differential amplifier. It's an even more realistic simulation of push-pull amp clipping. Hughes&Kettner "Quantum" amps used a similar scheme but also introduced some diode clipping to simulate grid conduction and resulting DC bias shifts + crossover distortion.

Peavey T-Dynamics circuit does about similar PA emulation thing with plain clipping diodes, simulates grid conduction again with clipping diodes and also introduces DC bias shifts for introduced crossover distortion. To tell the truth, you really don't need tubes in the circuit to create a very realistic PA simulation but you can handily use those to substitute gain stages that otherwise would be solid-state (e.g. Valvetronix / H&K Quantum amps). The overall circuit architecture actually makes WAY more difference than what devices are used in the circuit.

Oh, which reminds me, for convincing simulation you also need a somewhat complex filtering stage that mimics the non-linear response to speaker loads.

My personal scheme implements most of that and also introduces voltage sag emulation:
http://www.ssguitar.com/index.php?topic ... 2#msg11412

I know it's entirely heretical but it actually performs way more realistically than a simple LTP stage with single output.