Asymmetrical waveforms !

[Nick]

Yep, I agree with what izzy says. I wonder if the problem Steve is the phrase "Distortion Cancellation". It may be that you are thinking it takes the signal works out which part of the signal is distortion and changes that. If it could do that it would be a wonder of modern science. What it actually means is that the distortion present in the individual output device is canceled by adding the other half.

Where it gets a bit confusing is James's woodwind signal. By the nature of the instrument it generates a fundamental signal and a number of distinct harmonics. They look like distortion, but of course they are not, they are the sound of the instrument. Adding the distortion generated by the amp to that signal change the same harmonics and so look like its affecting the original signal in a intelligent way.

[steve s]

Thanks gents, I realise my issue is my lack of obviously not understanding the process

I was thinking it was similar to negative feedback, where feedback is combined with input but out of phase
And the cancellation of distortion and gain takes place
And it probably is related but that not what you both are saying,

But I spent a few hours searching last and it appears there is a lack of detail describing the cancellation or distortion reduction? process.

Without that knowledge I'm struggling to completely understand why its only theoretical, and also explain why Morgan Jones told me that push pull causes a low level of signal cancellation and change.

[Nick]

I was thinking it was similar to negative feedback, where feedback is combined with input but out of phase
And the cancellation of distortion and gain takes place

It is similar.

explain why Morgan Jones told me that push pull causes a low level of signal cancellation and change

Well, you would first have to get him to explain just what that means. Its certainly true that push pull causes a lower level of signal change than single ended. Is that what he meant, or something different?

But I spent a few hours searching last and it appears there is a lack of detail describing the cancellation or distortion reduction? process.

It may be because what you are searching for doesn't exist?

I had a quick look and this doc from 1960 perfectly describes what I am badly describing, and the key is the generation of composite curves, which as you can see from that document are far more linear than a single valve.

http://www.samhallas.co.uk/repository/e ... cs_2_7.pdf

The only thing I can think of that may be related to what Morgan was talking about was the flux in a PP output transformer goes through 0, so there is the hysteresis loss at that point, and that may be better or worst depending on the core material. One of the reasons I use nickel cores in the output transformer in my DAC.

It may be the problem is your use of "cancellation" I am not sure that means what you think it does in this context.

[Nick]

It may be if you used "reduction" instead of "cancellation" it would have a different meaning to you?

BTW, my personal view is I don't like most push pull valve amplifiers I have heard, there are exceptions but they are few, and in most cases the problem is the phase splitter not doing a very good job, and that WILL cause the output stage to cancel some of the signal some of the time. But that is due to the circuit, not push pull itself. Transformer phase splitters don't have that problem (and the amps that I do like seem to use them) but can bring their own.

[Nick]

Also, the transformer means that the general level of feedback available is low, so I think that it can only work to push the distortion down a little, but by doing so increase the level of higher harmonics, so sounding worst but measuring a bit better (unless you start weighting the distortion products as the BBC did when comparing amps). Solid state amps can use much higher levels of feedback so don't have that problem.

[steve s]

Nick, that link was worth posting, explaining what I was missing extremely well..
Thanks

I think it's been very much a case of me looking for explanations and looking a little to far outside the box..

But for me it was worth it, the explanations now all make sense
And after quite a few days of scouring the net for explanations, it appears many just trust what said rather understand the process fully.

The one interesting thing I read was that to work properly, both sides of the circuit need to be completely balanced, with one contribution saying in reality thats nearly an impossible to achieve.. but I suspect again good enough is better than the alternatives for most.
I'll keep reading that link as there's alot in it,

Anybody got any nice push pull txs to sell, I may well revisit

[steve s]

It may be if you used "reduction" instead of "cancellation" it would have a different meaning to you?

BTW, my personal view is I don't like most push pull valve amplifiers I have heard, there are exceptions but they are few, and in most cases the problem is the phase splitter not doing a very good job, and that WILL cause the output stage to cancel some of the signal some of the time. But that is due to the circuit, not push pull itself. Transformer phase splitters don't have that problem (and the amps that I do like seem to use them) but can bring their own.

I'm rather slow at typing so catching up..

the seemingly competent descriptions of push pull ive looked at these last few days show a transformer phase splitter.
I've always wondered how the opposite ends of a valve give exactly 180 split, and how differing gain is
counteracted, I'm sure there are simple explanations.

[pre65]

Talking of phase splitters.

My first build valve amp (832a PP) uses a different type of phase splitting than usual in that the input valve only feeds one half of the 832a. I forget the technical name for it. Also I tried feedback (switchable on/off) and preferred it without.

And, the output transformers were Audionote ones so of reasonable quality.

At an early Owston Nick said of my amp "I can see why you like it", but I never got to understand what he really meant.

[steve s]

Been reading rod Elliotts site, as he seems to have a honest approach to the effectiveness of valve phase splitters (compared to most )
According to rod, most seem to require some form of adjustment to provide equal gain etc especially when accuracy is required
And he references the fact that many types should include some form of adjustment to retain symmetry.

Which did have me thinking about how they react to incoming asymmetry for a few minutes


I also could not help but to copy this section, his reference to push pull valves balancing

"While most valve power amps provide a means to adjust the negative grid bias voltage, few provide a pot that allows the valves to be properly balanced.  Not providing this adjustment is a real nuisance, because matched valves must be used.  Many offerings of 'matched' valves must (or so it seems) consider them to be 'matched' if they look the same - they often certainly don't behave the same.  Yes, valves should be matched, but the problem is where?  At full power?  Half power perhaps?  At what voltages (plate, screen, control grid)?  There are too many things to match, and valves just aren't that simple."

https://sound-au.com/valves/analysis.html#s4

[Nick]

I've always wondered how the opposite ends of a valve give exactly 180 split, and how differing gain is
counteracted, I'm sure there are simple explanations.

If you mean a Cathodyne or Concertina then its simpiler than you think. If you have equal resistors at the top and bottom, consider the current flow. All the current that flows through the anode must flow through the cathode as there is no where else for it to go. If a signal on the grid (a instant signal value not a wave over time) means the current changes, then the current change through the anode resistor will be exactly equal to the current change through the cathode resistor. Because the current is equal and the resistance is equal the from ohm's law the change in voltage across the resistors will be equal. Because one resistor is referenced to 0v, and the other b+ the change in voltage across the two resistors will have to move is the opposite direction WRT 0v, so you get the two voltages moving in opposite directions by exactly the same amount, So 180 out of phase.

few provide a pot that allows the valves to be properly balanced

Well, again I would question the use of the word properly there. At best it can only balance them at one point in their curves, and so while it may be better than nothing, its still far from perfect.

There is no doubt there are problems in PP amps, the advantages PP offers does come with a cost. Many of those problems are also present in PSE amps, though not those of the phase splitter.

Which did have me thinking about how they react to incoming asymmetry for a few minutes

This is starting to become an obsession . Maybe seek professional help.

[steve s]

Thanks I'd thought I'd do a little more research today, but yours is a nice simple explanation, right at my level..

Interesting comment you've made on pse amps, again in my limited experence,I found it clear to hear the differences and the similarities in the effect on sound of the amps.

What's highlighted so far in this thread ( excluding my lack of understanding !!) is that push pull amps should be better/ lower distortion etc , but are more difficult to get right.

And most aren't quite right, ...whatever right is of course

This is starting to become an obsession . Maybe seek professional help.

Too right.. but this forum is group therapy, sort of of half way house for the audiolic

[steve s]

The only thing I can think of that may be related to what Morgan was talking about was the flux in a PP output transformer goes through 0, so there is the hysteresis loss at that point, and that may be better or worst depending on the core material. One of the reasons I use nickel cores in the output transformer in my DAC.

Just coming back to this point, you are more than likely correct but I also wonder....
Along with all morgan's tech knowledge he was a very skilled listener, reconising what good sound was in the same way we do.

It appeared to me he recognised the differences in the SET's at our meetings, compared to the average pp amp. And like me, he didn't put the blame totally on added 2nd harmonics.

I don't think he was one to spend vast amounts on components too,
So maybe he was comparing sets to the type of push pull you and me have referred to, the ones that dont quite hit the mark to our ears, for whatever (unknown ) reasons..

The 'great' thing about audio today is that there is always someone with an opposing view, having read far too much about audio these last few days
It difficult to come to any real concrete conclusions about anything.

[Nick]

So maybe he was comparing sets to the type of push pull you and me have referred to, the ones that dont quite hit the mark to our ears, for whatever (unknown ) reasons..

Yes, maybe, but I find it unlikely that given the unknown reason he would then not try to find what it was. Or maybe he had ideas but just kept them to himself.

[steve s]

Yes, maybe, but I find it unlikely that given the unknown reason he would then not try to find what it was. Or maybe he had ideas but just kept them to himself.

If I remember correctly he said something simple, on the lines that the process of of cancelling distortion had effects the music signal..


Unfortunately I was that surprised at that, I didn't ask for details
And I have regreted not asking ever since.

That was the 2nd time he came up.

I expected to see him in May, but he didn't come to Retrotech.

[Nick]

If I remember correctly he said something simple

In that case it should be simply measurable, and probably fixable. Maybe worth looking at Some of Lynn Olson's designs. He has addressed a lot. You can do a lot with a CCS that wasn't available back in the day.

https://www.meta-gizmo.org/Tri/LYNN%20OLSEN.HTML
https://audio-talk.co.uk/phpBB3/viewtopic.php?t=7934

Both the amp Mark made and that balanced pre I did years ago (that was the inspiration for the output stage in my DAC) did something that at the time made me realize that there was something possible with push pull beyond the normal.