'PI' type distortion in a single ended amp - possible?

[makepeace]

Excuse the subjectivity in the following description, no other way to do it unfortunately. Naturally the sounds of an amp is based on the sum of its parts and its nigh on impossible to single things out when defining the sound of an amp. However, lately I've been reading a lot about what people claim is one of the 'secrets' to their favorite overdriven guitar tone. Basically - push pull, and the distortion of the phase splitter stage.

In the 'major' manufacturers of pop guitar amps list, there seem to be two generally preferred methods: long-tail and cathodyne - with long-tail generally dominating. My main question is this: what are the main factors in the cause of PI distortion and would the way the PI distorts be able to be recreated in a single ended amp in some way?

[Chuck H]

I think it's important to note that all the amps that use a PI also operate in push pull. Duh, right? But it's important. With a single ended amp the output tube distortion is, ideally, half saturation and half cutoff. With some non linearity at the center to peak voltages. In a push pull amp the top and bottom of the wave form are both saturation. Some might argue that the tubes limit is the tubes limit whether it's saturating or cutting off. But the two behave very differently. For one thing, when a tube is saturated it's still conducting. Not so when it's in cutoff. The only time a tubes cutoff cycle is heard in a push pull circuit is when crossover distortion is present. In a push pull circuit each tube should handle exactly half of the wave form. In a perfect push pull world one tube cuts off as the other conducts. But the point where the tubes swap duties is affected by many things and there is often, if not usually, some audible crossover where cutoff is present. This can be a small or large artifact affecting the distortion sound for better or worse. And all this comes along with any amp that has a PI. I think that's my point. If the desireable distortions were all in the PI then the simple post PI master volume would be the universal answer. The post PI master allows for PI distortion while trimming the power tube distortions. Most players find the tone to be a compromise at best. So other systems like power scaling and attenuator circuits are emloyed to get the more desireable power tube distortions back into the overall tone.

I wouldn't trouble trying to get a push pull effect from a single ended amp. The resulting circuit would either fall short of the tonal goals or be complicated enough that you might just as well build a push pull amp. A nascar and a pickup truck are similar in many ways. Four tires, steering wheel, headlamps, bumpers, etc... But the pickup isn't likely to win any races and the nascar is useless if you need to move furnature. Lowering the pickup won't make it win more races and would detriment it's use as a truck. Putting an open bed on the nascar wouldn't make it a useful truck and would only make it less useful on the track. They're two similar things that are still different enough that no one should bother trying to make one into the other.

JM very opinionated 2C

[printer2]

Much easier to make a P-P amp sound like a SE amp. Run the amp in Class A and shut off the signal to one of the output tubes.

[loudthud]

Whatever these "Secrets" are, you might have to re-create them and just use the SE power amp to amplify them. Please provide links to some of these discussions. The three things I can think of are odd harmonic generation (the low level kind), crossover distortion and duty cycle distortion (where a non-50/50 square wave is created from a sinewave input).

I've been thinking alot lately about the 5F6A. It seems like the overdriven duty cycle modulation takes place at the input to the PI due to grid current. Some is generated in the 2nd stage, but not a giant amount and at that point the output stage is already heavily clipping. A single ended circuit could do this depending on how it is biased. I'd try a combination of fixed and cathode bias.

Crossover distortion might be created by anti-parallel diodes in series with the signal and some kind of resistive element in parallel that shorts the diodes out until overdrive generates a signal that causes the resistance to increase. You could do this in series with the speaker but the resistive element would have to be very low impedance. Better to do it at the input to the power tube. An opto might be too slow and a JFET might not withstand the voltage.

One last thing is ripple infiltration. When a push pull amp clipps, supply ripple is coupled to the speaker on the flat tops. A single ended power amp only does this on one polarity when the tube is conducting.

[teemuk]

If you are never going to overdrive the power amp super heavily you can always just throw in a circuit like LTP to the preamp and use only one of its outputs. It introduces most of the characteristics commonly associated with push-pull amps.

e.g. Laney Klipp, Legend Rock-n-Roll series, Crate Blue Voodoo, etc.

You can also extend the idea and fit entire real push-pull stages to the preamp.

Personally I feel that solid-state amps have dedicated vastly greater effort in developing such schemes and making them sound and feel like the real deal (whether its done in all SS or hybrid formats) while the same idea has mostly been left unexplored in tube amplifiers.

[printer2]

If you are never going to overdrive the power amp super heavily you can always just throw in a circuit like LTP to the preamp and use only one of its outputs. It introduces most of the characteristics commonly associated with push-pull amps.

e.g. Laney Klipp, Legend Rock-n-Roll series, Crate Blue Voodoo, etc.

You can also extend the idea and fit entire real push-pull stages to the preamp.

Personally I feel that solid-state amps have dedicated vastly greater effort in developing such schemes and making them sound and feel like the real deal (whether its done in all SS or hybrid formats) while the same idea has mostly been left unexplored in tube amplifiers.

I do not think taking only one side of the output would work, you get no cancellation of the second harmonic generated.

[J M Fahey]

I do not think taking only one side of the output would work, you get no cancellation of the second harmonic generated.

I don't get it: if you find push pull sound of distortion so important, why then you try to create it "single ended" ?

[printer2]

I don't get it: if you find push pull sound of distortion so important, why then you try to create it "single ended" ?

Just using up an output transformer I guess? I would just go P-P.

[Chuck H]

Right. The best reason to do do something (that is a luxury anyway) is because you can. If it's possible to create the "big power amp" sound in a preamp or single ended (so it can be padded as a preamp) then it WILL be discovered. It hasn't yet. And I'm not looking for it. That doesn't mean it won't happen.

[eschertron]

Excuse the subjectivity in the following description, no other way to do it unfortunately. Naturally the sounds of an amp is based on the sum of its parts and its nigh on impossible to single things out when defining the sound of an amp. However, lately I've been reading a lot about what people claim is one of the 'secrets' to their favorite overdriven guitar tone. Basically - push pull, and the distortion of the phase splitter stage.

In the 'major' manufacturers of pop guitar amps list, there seem to be two generally preferred methods: long-tail and cathodyne - with long-tail generally dominating. My main question is this: what are the main factors in the cause of PI distortion and would the way the PI distorts be able to be recreated in a single ended amp in some way?

As a newbie to amp theory and design, suffer me to ask a clarifying question: is "PI distortion" equivalent to clipping, or is some of that distortion related to the asymmetry of the signal introduced by the PI? the discussions that I've followed recently talk about how different PI designs come closer to - or intentionally avoid - a perfectly matched, out of phase signal. Does any of that matter, and if so, to what extent?

[NateS]

Consider the AX84 firefly. 12AU7 push pull power amp. They're still triodes, not pentodes, but you do get even harmonic cancellation of harmonics resulting from overdriving the final stage. I think that may be in the spirit of the original goal.

[makepeace]

As to the reasons/rationales; said most eloquently by Printer2 and Chuck.

Because I have SE OTs, and because I can. Or want to can.

I think I have a couple of ideas. Will try them and let people know how it goes.

Thanks for the responses!

[NateS]

I do not think taking only one side of the output would work, you get no cancellation of the second harmonic generated.

True, but you can reinvert one of the outputs and sum them.