A question about distortion in SE vs. PP designs:

[BHD]

I have a question:

I understand that SE amps produce more even order harmonic distortion than PP amplifiers, because the PP amplifiers cancel out most of this distortion at the output.

My question is do SE amps produce LESS odd order harmonic distortion, or is it the same amount given the number of output devices? In other words, would two triode output tubes in parallel SE produce the same amount of odd order harmonic distortion as two triode output tubes in PP, or would PP operation produce MORE odd order harmonic distortion?

Any help would be appreciated.

[Robert McLean]

The amount of odd harmonic distortion, in proportion to the fundamental, is the same in both cases. Dont forget the "in proportion to the fundamental" part. In proportion to the total output, the odd harmonics are increased in a PP amp. (since the evens are gone)

Paralleling the SE tubes, assuming you dont change the operating conditions, will have no effect on distortion.

The math works like this :
if each tube has an output given by
O1 = A1*V + A2*V^2 +A3*V^3 + A4*V^4 .....

where An is the amplitude of the nth harmonic, and V is the input voltage

then the equation applies directly to the SE case, and ratio of 3rd harmonic to fundamental is A3/A1

in the PP case you have one tube with the above equation, but the other tube has input = -V,

so O2 = A1*(-V) +A2*(-V)^2 + A3*(-V) + A4*(-V)^4 .....
= -A1*V + A2*V^2 - A3*V^3 + A4*V^4 ....

since it is push pull the total output is O1 - O2
O = 2*A1 + 2*A3 + 2*A5 ...., ie no even harmonics

third harmonic in proportion to fundamental is
2*A3 / 2*A1 = A3/A1, same as SE case.

I have seen text books make statements like " PP amps double odd harmonics for the same reason they cancel out even harmonics", but dont forget they also double the fundamental, which after all is the first odd harmonic.

Also bear in mind that although the odd harmonics are the same in proportion to the fundamental, they are increased in proportion to the total output, and are the sole harmonic content, and so the sound of PP should be different than SE.

[cerrem]

Although I agree with Mr. Robert McLean....
The reality is that the PP amp is typically "driven" from SE voltage amplifiers that do produce even order harmonics...
The PP stage will "pass" these even harmonics that you feed it's grids...it only cancels it's own even harmonics.... i.e, how do you "listen" to just an output stage without other influences...
Also the ratio of odd to even harmonics generation is a function of a number of variables such as plate loading, feedback...ect..

Chris

[dhaen]

Originally posted by BHD
I have a question:

I understand that SE amps produce more even order harmonic distortion than PP amplifiers, because the PP amplifiers cancel out most of this distortion at the output.

Yes.

Originally posted by BHD

My question is do SE amps produce LESS odd order harmonic distortion, or is it the same amount given the number of output devices?

The odd order distortion is caused mainly by the transfer charcteristic of the device. The topology is not relevant in this case. Generally triodes produce less than pentodes.

Originally posted by BHD

... In other words, would two triode output tubes in parallel SE produce the same amount of odd order harmonic distortion as two triode output tubes in PP, or would PP operation produce MORE odd order harmonic distortion?

Any help would be appreciated.

Roughly the same amount of odd order distortion will be present. Even order will be greater.

The caveat is that negative feedback can change things. However, if you produce a circuit that performs well before feedback is applied, you stand a good chance of it sounding "nice" afterwards, if you do decide to use it.

[BHD]

Okeedokee... thanks to those that responded.



And here's why I asked. I recently read an article from Glass Audio about an output stage that's best described as a "parallel single-ended phase inverted" output. How it works is to take two single ended amplifiers and invert their output at the output of the transformers. It combines the outputs of two single ended amps out of phase at the connection to the loudspeakers. The author claims (if I may condense it down) that it gives the advantages of both single ended and push pull topologies, cancelling out the even order harmonic distortion while producing less odd-order harmonic distortion than a standard push pull output circuit. It would, of course, be ridiculously expensive because it would require two single ended transformers per channel. I mainly wanted to know if this approach had any validity.

[Tweeker]

Singled ended transformers dont suffer from hysteresis the way PP ones can, they have to contend with saturation instead.

[BHD]

Hey tweeker,

Well, from what I understand about the term "hysteresis" is that it's sort of the amount of energy needed to actually make a change or movement in SOMETHING. In the context of my experience, a motor with a 4-20mA control signal doesn't REALLY have an infinite amount of resolution, so it might take a .5mA change in control signal to see any real change in motor movement. So you're saying that the hysteresis in a push pull topology amp is greater than a SE version...? I understand that transformers in SE amps are more prone to saturation than PP versions, but can you help me out and expand a bit on the role that hysteresis plays in PP amps?

I'm fully ready to accept that I have a totally screwed up idea of what hysteresis means in the context of tube amps...

[Tweeker]

Each time the magnetic field is reversed, a small amount of energy is lost to hysteresis in the magnetic core. The energy goes to magnetizing the core. The field is never reversed in an SE amp.

[Tweeker]

Wikipedia on Transformers.

Wikipedia on hysteresis.