So what's the deal with single ended hifi amps?

[Radio Wrangler]

It is interesting that people latched onto single-ended versus push-pull difference for their desirability factor, and seem quite unaware of class A/AB/B differences.

I suppose it's a matter of what is visible to the non-technical.

Douglas Adams postulated that the highly erratic and non-linear maths used in Italian bistros in the calculation of bills could be used to create a faster-than-light spaceship drive. He could have used technical explanations as practiced in hifi boutiques.

I once had things explained to me thus:

"This "feedback" idea looks at the output of an amplifier and compares it with the input signal to determine the error. Then it uses that error as the input to the amplifier proper which drives the speakers. So all you get to hear is amplified errors. No wonder it sounds terrible. You need to spend tthe extra for a more advanced amplifier which doesn't use feedback.They don't have errors."

There's enough alternative logic there to visit a neighbouring galaxy.

David

[Ted Kendall]

Originally Posted by barrymagrec

...and if that`s the sound you like, go for it.

It`s not High Fidelity though......

Precisely. And if you really want single-ended sound, why not use a Linsley Hood Class A and save yourself a packet?

[Argus25]

Originally Posted by Radio Wrangler

There's enough alternative logic there to visit a neighbouring galaxy.

David

As long as its sounds plausible.

Everybody knows that if you push a Cat off a table it always lands on its paws and invariably it has been observed if you drop buttered toast it always lands on the buttered side down.

So if you strap the buttered side of the toast to the Cat's paws and push that assembly off a table, it just sits there floating above the floor, a perfect anti-gravity machine.

[barrymagrec]

Originally Posted by Ted Kendall

Precisely. And if you really want single-ended sound, why not use a Linsley Hood Class A and save yourself a packet?

Now you`re talking quality amplifiers - I still don`t think I`ve heard anything better than the Linsley Hood class A driving a Quad ESL57.

[mark_in_manc]

Originally Posted by Argus25

Everybody knows that if you push a Cat off a table it always lands on its paws and invariably it has been observed if you drop buttered toast it always lands on the buttered side down.

So if you strap the buttered side of the toast to the Cat's paws and push that assembly off a table, it just sits there floating above the floor, a perfect anti-gravity machine.

That's very good. Although years ago I pushed a friend's cat off his kitchen worktop - unconsciously copying what my Mum did when ours went somewhere it shouldn't, I think - and it landed splat on its back. Apparently it had been in a road accident and was no longer as lithe as it once had been. Oops.

[Herald1360]

L-H class A?

It's push-pull.....

Now, what about two SETs driven in antiphase with the speaker between the outputs? Worst of all worlds?

[barrymagrec]

Originally Posted by Herald1360

L-H class A?

It's push-pull.....

N

No, it`s single ended with an active load - In the original Wireless World article JLH explains the topology.

[kalee20]

Originally Posted by Radio Wrangler

I once had things explained to me thus:

"This "feedback" idea looks at the output of an amplifier and compares it with the input signal to determine the error. Then it uses that error as the input to the amplifier proper which drives the speakers. So all you get to hear is amplified errors. No wonder it sounds terrible. You need to spend tthe extra for a more advanced amplifier which doesn't use feedback.They don't have errors."

True to an extent - consider an amplifier which generates ONLY second harmonic distortion, and you try to fix it with NFB. The new circuit will have (less) 2nd harmonic, but will now have 3rd, 4th, 5th, ... which weren't there before.

However this is OT as the thread is about SE vs PP, not about NFB or no NFB.

(Incidentally, if I decide my music is a bit to quiet or too loud and I get up and adjust the volume control until it is my 'target' volume, am I guilty of closing an overall NFB loop and will quality then immediately suffer?)

Originally Posted by Argus25

This was the way push pull was initially designed by RCA ...They had not cottoned onto the energy savings of increasing the negative bias for each valve... They stated that for the
push pull connection the power output was the same as running the tubes in parallel but the drive voltage requirement was double.

Irrespective of being able to chang the bias and the load, interesting that they had not cottoned on to the cancellation of even harmonic distortion (and cancelling of DC) with pure Class A! That alone would make PP worth the trouble of getting the extra drive voltage, which back then was I daresay with a transformer with turns ratio 1 : 3+3 for PP (rather than just 1:3 for SE) and later by an extra valve as phase inverter.

[Synchrodyne]

Originally Posted by Radio Wrangler

I once had things explained to me thus:

"This "feedback" idea looks at the output of an amplifier and compares it with the input signal to determine the error. Then it uses that error as the input to the amplifier proper which drives the speakers. So all you get to hear is amplified errors. No wonder it sounds terrible. You need to spend tthe extra for a more advanced amplifier which doesn't use feedback.They don't have errors."

There's enough alternative logic there to visit a neighbouring galaxy.

Here’s another one about NFB:

“Almost by inspection you can see that the feedback amplifier has the capacity to go on compounding its error residual. When an amplifier is processing a complex, harmonically rich input signal---music---and not a steady-state single sinewave tone in a lab test, something could well go wrong. That cascade of residual errors will intermodulate at low levels, but it will intermodulate in a fantastically complex manner.”


So those initial IM products are quite selective – they intermodulate only with their own kind after going around the feedback loop. As I recall, an associated comment was to the effect that whilst NFB in general was bad, degeneration was OK.

I was of the impression that at any given instant, a “complex, harmonically rich signal” could be resolved into a superposition of several – or even a multitude of – steady state sinusoids. Maybe the laws of physics are different in that neighbouring galaxy.


Cheers,

[bikerhifinut]

Originally Posted by Radio Wrangler

I once had things explained to me thus:

"This "feedback" idea looks at the output of an amplifier and compares it with the input signal to determine the error. Then it uses that error as the input to the amplifier proper which drives the speakers. So all you get to hear is amplified errors. No wonder it sounds terrible. You need to spend tthe extra for a more advanced amplifier which doesn't use feedback.They don't have errors."

There's enough alternative logic there to visit a neighbouring galaxy.

David

Wouldn't that be it somehow amplifying the difference in some way rather than cancelling the bad stuff? Bad analogy I know.
Anyway, having fiddled with lowish power SE amps, I can understand how one would appeal to a listener. My gripe is the way they run out of steam way too early and when they get to that state they really can sound unpleasant.
In a way the distortion argument is a little bit spurious, as the biggest distortion generator in any Audio System is usually the loudspeaker, compounded by its interaction with the room its in? Add into that mix a bit more distortion from an LP source and the couple of percent from an SE amp running in its comfort zone isn't really that much. I may have made a couple too many assumtions there and am happy to be corrected on this.
SE amps are often used with very efficient horn loaded speakers with single full range drivers, again this will significantly alter the presentation of the sound.
It all boils down to whether you like them, and if it gives you a nice warm feelgood factor who am I to say that's wrong?
One thing I would be interested in is what a pair of good modern ESL's would make of one, or even a well fettled pair of ESL57, assuming you had an SE amp capable of driving the load. I think they'd sound pretty dire as the ESL speakers would lay bare all its characteristics?
Horses for courses as always, and for the most part in my experience, SE amps tend to be built by DIY enthusiasts I look upon them as part of the overall experience.
Theres more than one member here who likes their mullard 3-3 builds, I'm not knocking that.

Andy.

[G8HQP Dave]

Originally Posted by joebog1

Sound quality ? Yes very nice, about the same as my push/pull 6BW6 valves. It did have better bass!!, BUT when I added negative feedback, the bass went away, about the same as my push/pull 6BW6 amp.

Yes, a lack of feedback causes high output impedance which means more bass around the speaker resonance. Some folk like this, some do not, but it isn't hi-fi.

[G8HQP Dave]

Originally Posted by Synchrodyne

I was of the impression that at any given instant, a “complex, harmonically rich signal” could be resolved into a superposition of several – or even a multitude of – steady state sinusoids. Maybe the laws of physics are different in that neighbouring galaxy.

If you spend any time on a typical audio website you will find that Fourier denial turns up with disappointing regularity. Then you get Nyquist denial when they start talking about digital audio. Apparently music signals are so much more complex than anything seen in medicine, defence, telecommunications, instrumentation or basic science so it is hardly surprising that these 'conventional theories' cannot cope.

[The General]

Well this thread has generated a fair few posts already & thank you for your responses. Most of them support what I already suspected to be the answer to my question, so I'm going to call audiophoolery/fashion on this one. After all, virtually all the 'big names', Leak, Quad, Rogers etc. go the push-pull route.
Mrs General had a take on it, she opined that it was a fashion for going back to the most basic roots - a bit like holidaying in a yurt rather than staying in a Radisson.
Oh & I liked David's explanation of NFB in post #21 - gave me a chuckle this morning.
Mark

[The General]

Mind you, I did have a Grampian PA receiver once which had a pair of EL41s in the output with a reasonably chunky output T/X. The EL41s turned out to be in parallel SE rather than P-P. No idea why. I moved it on as it was a bit too rusty for me to tackle at the time.

[ms660]

Talking of single ended....

Dansette...EL84…TC8H...& Barbara Randolph 'll do for me.

Lawrence.

[ajgriff]

Originally Posted by bikerhifinut

In a way the distortion argument is a little bit spurious, as the biggest distortion generator in any Audio System is usually the loudspeaker, compounded by its interaction with the room its in?

Andy.

Yes and one of the most significant factors in the creation of speaker distortion is the difficulty associated with passive crossover design. Unfortunately, for practical reasons, the mid to high crossover point is usually in the middle of the frequency range most important to human auditory perception. The solution is to use active crossovers. I would therefore propose a three way active stereo system using six single ended valve amplifiers. That's four big output valves plus two even bigger ones for the bass drivers. This set up should keep the committed audiophile warm and happy. Sympathetic friends and relatives would be able to gaze in wonderment at the big valves through the small perspex viewing window. Notice how the padded cell also solves the room interaction problem.

Alan

[Argus25]

Originally Posted by kalee20

Irrespective of being able to change the bias and the load, interesting that they had not cottoned on to the cancellation of even harmonic distortion (and cancelling of DC) with pure Class A! That alone would make PP worth the trouble of getting the extra drive voltage, which back then was I daresay with a transformer with turns ratio 1 : 3+3 for PP (rather than just 1:3 for SE) and later by an extra valve as phase inverter.

There is more than meets the eye too.

Of course, as you know, there is no net core magnetization in push pull as the standing bias & plate current cancel in each half of the primary.

So everybody got excited that this was a wonderful feature, much smaller (and cheaper) and lighter weight iron cores for output transformers, that didn't need to be butt stacked either, all like a salesman's dream of a good deal if you throw in the increased efficiency, making push pull (valve or transistor audio opt stages) even more enticing, especially for battery operated and portable radios.

But of course, there might be something lost. There is no such thing, most of the time, as all advantages and no disadvantages with a circuit configuration design change.

It gets forgotten a lot of the time that the high standing (zero signal current) and power dissipation in class A has one very big advantage (if you have an output transformer). It is not all wasted. This energizes the core of the output transformer with magnetic field energy, which is returned to the speaker on half cycles where the anode voltage rises above the power supply voltage and transfers energy from the magnetized core, to the load.

In push-pull with no net magnetization of the core, this half cycle of energy comes from one of the valve's increasing plate current. Iron laminations and copper wire are less complex than a vacuum tube are they not ?

So what I am getting at, despite the obvious advantages of push pull in the energy efficiency and low iron mass department, it doesn't take away some of the interesting features of class A, where if efficiency & mass is not an issue, it might be better.

[bikerhifinut]

Originally Posted by ajgriff

Yes and one of the most significant factors in the creation of speaker distortion is the difficulty associated with passive crossover design. Unfortunately, for practical reasons, the mid to high crossover point is usually in the middle of the frequency range most important to human auditory perception. The solution is to use active crossovers. I would therefore propose a three way active stereo system using six single ended valve amplifiers. That's four big output valves plus two even bigger ones for the bass drivers. This set up should keep the committed audiophile warm and happy. Sympathetic friends and relatives would be able to gaze in wonderment at the big valves through the small perspex viewing window. Notice how the padded cell also solves the room interaction problem.

Alan

And makes even more sense of the high efficiency loaded horn speaker with a single full(ish) range driver? I will happily concede that these single driver speakers are compromised in many areas but explains why they go well with low power SE amps?
Once in the dark past, most speakers were quite large and efficient which meant 10watts from a pair of EL84's was plenty! (Actually to my ears it still is in a normal UK sized lounge).

It may be a bit on the fringe but if the amplifiers make a pleasing sound to the listeners lugoles, I really don't care how the watts got there.
Sometimes I think we forget that music is an emotional experience, (and if it isn't what are you listening to it for I wonder?) so if a certain recipe of gear tickles that part of the synapses, then does it matter that it's not "Hi Fi". Many people wouldn't really like total accuracy in a domestic environment, who wants the LSO at full chat in front of the sofa? Or the visceral bass slam of a modern rock band (or an older one even). I don't think thats Audiofoolery as most of us I think would buy the domestic stereo on how it sounds to us as individuals rather than a stack of bare performance statistics alone. I agree with criticism of the more mendacious marketing, but most of us are capable of detecting the smell of the farmyard i'd hope.

Andy.

[G8HQP Dave]

Pursuing something other than hi-fi only matters if you falsely believe that you are pursuing hi-fi, and so believe that your chosen 'other' has better fidelity when in fact it has worse fidelity.

[kalee20]

Originally Posted by Argus25

In push-pull with no net magnetization of the core, this half cycle of energy comes from one of the valve's increasing plate current. Iron laminations and copper wire are less complex than a vacuum tube are they not ?

Personally I'd say that a vacuum with a single kind of fundamental particle whizzing through it is much simpler than a complex matrix of fundamental particles, some of which are paired and some aren't, giving rise to a net magnetic effect - which itself is partly thwarted by imperfections in the spatial arrangement of that matrix.

In other words, a winding with a core of ferrous alloy, grain-oriented but with dislocations, giving a high permeability but with both non-linearities and hysteresis, is a lot harder to understand than a valve!

In push-pull it's easier to make the effects of the OP transformer negligible.