6V6 in Class AB1 push-pull mode

[bluefoxicy]

I cannot find anything about appropriate values for a 6V6 pair in Class AB1 push-pull operation. Not in RC-30 and not in the Internet!

Any thoughts? I basically want to match power output with an EL84 in AB1 push-pull mode. A 6L6 (scaled down a bit) and an EL34 would fit together for this well, since they use the same plate load (but one produces 5.1% THD and the other 2%).

[bluefoxicy]

Found it! Page 557 of RC-30 has an additional chart for 6V6GT:

Plate: 285@70mA
Grid bias or cathode resistor: -19V
Screen grid: 285@4ma
Plate load: 8k
Output: 14W

Compare with EL84:

Plate: 300@72mA
Grid bias or cathode resistor: -28.3V
Screen grid: 300@8ma
Plate load: 8k
Output: 17W

Scaling this back to about 250V should land me somewhere in the vicinity of 14W

EDIT: Huh damn, forgot, the two sides of the thing are going to have the same plate voltages...

Well there's one other way to decrease power: Leak part of the plate to the screen grid (partial triode mode). I can scale the EL84's output back that way.

[Steve Conner]

That sounds like it would be difficult to calculate. Me thinks it would be much easier to just hook it up and see what happens. Fixed bias with two individual pots would be a good idea, or cathode bias with a separate resistor and cap for each tube.

[bluefoxicy]

That sounds like it would be difficult to calculate. Me thinks it would be much easier to just hook it up and see what happens. Fixed bias with two individual pots would be a good idea, or cathode bias with a separate resistor and cap for each tube.

Yeah, probably fixed bias, requiring a small tranny for a fixed bias source (can't seem to avoid this...)

The reason for trying to partially triode the EL84 side is that I need 14W from 7W+7W (or 50W from 25W and 25W); 8 tubes with the top signal in a 34W bank and the bottom in a 28W bank doesn't sound appealing to me, and I can't power scale half the OT. I'll probably need to use a small R/C filter or something between B+2(EL84) and B+2(6V6) (yeah, the screens are connected to the same power source, ya know?), but I'll only figure that part out experimentally.

[NorCalTuna]

You can make a bias supply, with a few parts, from the secondary- It's done with a resistor in the BF Princeton Reverb, or with capacitive coupling (.047 uf @ 1kv) in a more modern setting. It's simple- cap off sec., res (27k) to ground, reverse diode (1N4007), filter net (.1 600, 100/100) then to a pot and a threshold res- or just a resistor- to ground. Just check out some budget fixed bias amps- they were too cheap to specify a bias tap- err, a 3w resistor was cheaper.... and the rest of the circuit is the same as a good beefed up bias circuit (unless you like hum injected into your signal) it's compact and simple.