Push Pull Question

[paolojm]

If I have a Class A biased push pull pair is it viable to ground one of the PI outputs to create a pseudo single ended amp at half the power rating? AFAIK the OT will still be DC balanced just that only one side will pass signal.
The crux of my question is what load will the valve passing signal "see"?
I understand that in Class A with both sides passing signal into a 8k P-P OT primary both valves will "see" 4k. I'm just not sure, with only one side passing signal, will it see 4k or 2k and hense whether I'd need to change impedance setting as well when doing it.
Any guidance from you guru's would be greatly appreciated.

[hamfist]

I hope someone answers this, as you ask some great questions here !

[WesKuhnley]

I'm not entirely sure, but to get your suggested PI/power stage to work like that, you probably would have to worry more about plate voltage and bias than loading. You would probably have to change the bias for both output tubes to maintain DC current balance in the primary, and lower the B+ if it's too high for the tube to operate in class A (it probably is). An easy way to do this might be to use a VVR type circuit with two switchable resistor values rather than a pot, smaller and probably cheaper than the power resistors you would have to use otherwise. Someone call me out on this if I'm wrong...

[paolojm]

I'm sorted for B+, I've sourced a power transformer with a 190/260 Vac secondary, I'll using the 260 (350ish Vdc) for Class AB and 190 for Class A (250ish Vdc). A DPDT will be used to switch the secondary and bias resistors accordingly.

[SatelliteAmps]

It is possible. I don't know that the power drop will be dramatic enough to be worth it, but there are other uses that could make it amusing. You could use an on/off/on and run two different types of power tubes individually and be able to switch between them.

As far as what the transformer "sees," it will see one side of the transformer. What value it sees depends on how balanced the tranny is. Some are spot on, some have some sloppy winding and one side will be farther off than the other. Check with a meter and see.

[ntasher]

In push-pull operation, the dynamic load (the transformer primary) is virtually tied only between the two plates of the tubes. So an 8K load is virtually connected between two current sources with opposite signs driving opposite signals. When one side becomes higher, the other becomes lower. This keeps the voltage to current curve of the tube within it's recommended operation zone. When one of the tubes stops generating signal, the other tube "sees" the entire 8K load on it's plate and is knocked out of the recommended operating zone. You can think of it this way - the other tube used to push or pull current in the opposite direction so the voltage/current curve looked OK for the first tube. Now there is no other half to help pull or push the current so the working tubes operates at a different current/voltage curve - a non optimal one.
The plate load should be halved - to do so you need to switch the output tap to a higher rated impedance. From 8 to 16 or from 4 to 8 (i.e. connect 8 ohm speaker to the 16 ohm tap).
Nir

[donnyjaguar]

Think deeper, Ntasher, but you're on the right track. It'll work okay. Instead of grounding the output on one half of the PI I'd just disconnect the coupling cap to see if I like the tone enough to be bothered with it. The current flow on each power tube is primarily from the center tap through the tube to ground although the static tube won't unload like it normally should so in essence the other tube will see a lower working impedance. The static tube is basically acting like a resistor so think of the primary as now having a secondary winding with a load on it. I haven't measured the difference in power between these two modes, but I would hazard to guess it'll be closer to 1/4 rated output than 1/2. This probably doesn't bode well for longevity of the working tube either. It may sound way kewl, but probably not for long.

[andrekp]

Another option would be to leave the tubes running as is (so you don't have to worry about the OT "seeing" a different configuration) and changing the bypass cap circuitry on the output tubes.

First, you need to have each tube have a separate cathode circuit (the resistor and cap - they are shared in some amps). If you've got that (or change to that), then you can do a nice reduction in power by lifting the ground on the cathode caps. For a further reduction, tie the lifted ends of the caps to each other (i.e. from the cathode of one tube, though that cap, to the (previously grounded) end of the other cathode's cap, to the other cathode - as if they are running in series between the two tubes.) You can put this whole thing on a switch for all three settings.

You might like this as well, and you don't confuse your transformer that way.

[ntasher]

andrekp - your suggestion simply lowers the tubes gain - a master volume will have the same effect and I guess that is not the intention in the original post.
donnyjaguar - The fact that DC power comes from the OT CT has nothing to do with dynamic (AC) analysis since we are talking about a transformer here - there is magnetic coupling within the primary windings - so when you AC ground one end (by having zero signal generated by that output tube) the other tube "sees" the entire primary winding (due to that magnetic coupling). You are correct that power drops to about 1/4 of the original power since we know P=V*V/R and V dropped by half, while R remained the same. If we would reduce R by half, we get half the original power, and better loading of the tube.

[paolojm]

Some interesting suggestions there guys, thanks.

Just to clarify, the purpose of this is not actually the volume reduction, as halfing/quartering the power wont yield that much, but it's more for the tonal variety that it might yield.
The idea would be that one amp could yield that AB growl and edge (I know it wont quite be a 50 watt, fixed bias growl but no-ones perfect), the warmth of the Class A push pull and bluesy tone of a single ended power stage.
I had suspected that I'd need to alter the load as a few of you have suggested. I guess the sure fire way to find out is to build it, take all the measurements and see.

[RedMan]

You may want to check and see how your phase inverter is clipping. Generally it's a bit assymetrically. That's OK in push/pull since it will balance out but you may or may not like the sound if the waveform is too lopsided.

[IanSchets]

Maybe a bit late but I've just recently registered and came across this topic while browsing the forum.

What you want looks like what Peavey is doing with their Valveking series. These amps have a Class AB /A pot that allows class AB and A and anything in between. I have the schematic for the VK112 and what this pot actually does is controling the amount of the inverted signal to the second powertube in what looks a normal push pull power amp. The pot sits between two caps in the signal path to the second powertub. The signal to the first powertube is coupled by a cap that has halve the value of the caps in the inverted signal path.
It would be an easy mod I think.

Ian.

[donnyjaguar]

It sounds like Peavey is making a tradeoff between the two extremes with a view to not impacting the durability of the output stage itself. Ntasher, I think we're both saying the same thing. I mentioned the normal DC path merely as a starting point but in so doing may have obfuscated the issue. Go for it paolojm, and be sure to take some baseline measurements before you start tweaking.

If you can split the two connection to the primary centre tap this opens up some possibilities. :)