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

[Argus25]

Originally Posted by kalee20

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!

I still say the transformer core and copper is less complex than a valve, which requires equally or much more complex metallurgy, multiple manufacturing steps, complex equipment for vacuum and gettering, glass work, chemicals etc, but moreover has a much more limited life than a block of iron and copper.

A transformer might cost initially more because of the much lower volume manufacturing than a valve, but mostly, they out last a vacuum tube. Also both the valve and the transformer core have non-linearities in their transfer functions, and there is no escaping that.

[G6Tanuki]

Single-ended amps can sound great! I've got a stereo amp that uses a 6V6 for each channel and it sounds good to my non-golden ears.

One thing that puzzles me is why none of the hi-fi types have ever used the idea of 'sliding bias' to get the most out of their expensive valves. At low audio-level wind the -ve grid-bias up so the valve is passing just-enough current. Then when you want more loudness, reduce the grid-bias so the anode-current rises in order to deliver the greater output.

[alternatively, if you're a heretic who is using tetrodes/pentodes, slide the screen-grid bias up and down to control anode current in the same way as 'series-gate' AM modulation worked in amateur transmitters]

[bikerhifinut]

Originally Posted by G6Tanuki

Single-ended amps can sound great! I've got a stereo amp that uses a 6V6 for each channel and it sounds good to my non-golden ears.

I think I kind of said something similar further up the thread.
I also have a stereo SE amp based on a pair of 6V6 in pentode, first stage is an ECC81 and its jolly pleasant too! It does get a bit flustered when the wick is turned up but in its defence i was driving 87dB sensitivity speakers and its really only good for about 1.5W before the distortion gets in the way. It's only running at about 240VHT though so we can forgive it! And the output transformers are at best a bit weedy I think so not surprising. But as a small domestic amp with a sensitivity of about half a volt its not half bad, with a very low component count.
It's basically an adaptation of the Alex Kitic RH84 design which uses anode feedback and I am not using it with his ideas of zeners to supply the screen grid, nor CCS on the cathode as I reckoned they did nothing good for it apart from inducing instability.
Hi Fi? Well no, unless you listen in a 8 x 6 foot cupboard, but at sensible late night levels its never less than nice. And surprisingly completely hum and hiss free which I put down to OCD heater wiring and earthing and a deliberately low sensitivity.
I liked your ideas about the biasing, but for me if I felt I needed to do all that complication I'd just get a good solid state power amp. My feet are planted in both camps and if its good its good and I dont apply religion.

Finally confession time, bless me father for I might sin...................... But I quite fancy building up an SE power amp using EL34 or KT88 types for a bit of usable power. Mainly because I am curious and the parts can be gleaned from here and there or even adapted from one of the dodgy Chinese kits but I'm dubious about their output transformers. Keeps me out of the pub I suppose....................

A.

[Argus25]

Originally Posted by bikerhifinut

Mainly because I am curious and the parts can be gleaned from here and there or even adapted from one of the dodgy Chinese kits but I'm dubious about their output transformers. Keeps me out of the pub I suppose....................

A.

If you get one of the lovely single ended Hammond output transformers stocked by Antique Electronic Supplies (AES) in the USA, you will not be disappointed.

(ps my own stereo amplifier is push pull ultralinear 6L6's)

[bikerhifinut]

Originally Posted by Argus25

If you get one of the lovely single ended Hammond output transformers stocked by Antique Electronic Supplies (AES) in the USA, you will not be disappointed.

(ps my own stereo amplifier is push pull ultralinear 6L6's)

Or Bluebell Audio here in the UK. I'll have to save up my pocket money.

[Argus25]

Originally Posted by bikerhifinut

[I'll have to save up my pocket money.

What a great thread, leading to saving up to buy a single ended output transformer. That might help answer the original question, class A is well loved.

[Synchrodyne]

Originally Posted by G8HQP Dave

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.

Ah, that’s my problem – I don’t spend any time on typical audio sites, and I gave up reading audio magazines quite some time back. So I have missed out on the corrective training needed to disavow me of the notions (1) that amplifier behaviour can be explained (and predicted) by mathematics and physics, and (2) that an amplifier’s core job is to produce a magnified, but otherwise unchanged facsimile of its input.

Originally Posted by The General

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.

That conclusion is, I think, supported by the preponderance of evidence. These days it would be difficult to find a sound engineering or economic justification for using other than solid state devices if the objective was to build a good amplifier. By that I mean a device that simply amplifies, with the minimum reasonably possible deviation from a linear output vs. input curve when working into its specified range of loads. From there one could say that the choice to use valves would need to be an initial condition, and one imposed for its own sake. That immediately makes harder, but not impossible the job of designing a hi-fi amplifier. And within that sphere, doing the job with a single-ended circuit is probably harder in most cases than with a push-pull circuit, particularly if reasonable power output is required. Thus single-ended in most cases would be an initial condition imposed for its own sake. That makes it not an elegant engineering solution but something else, such as a “fashion statement”. Still, I don’t think that we can outrule that some amplifier designers choose difficult routes “not because they are easy, but because they are hard.” The apparent results also suggest that in some cases, the objective was not a “good amplifier” as such, but a combination amplifier and “coloration machine”. Again that is another self-imposed initial condition. It may well be unreasonable to measure such devices against the standards associated with hi-fi amplifiers, given that they were never intended to approach the ideal of “a straight wire with gain”, but to modify the input signal in ways the designer saw as beneficial.

Back in era when valves were the only choice for the active devices, in general there appeared to be more engineering rationality in respect of their application. In those days, suspension of disbelief was reserved for the opera, and not extended to the mathematics and physics of amplifier design. With hi-fi amplifiers, push-pull outputs were generally used where outputs upwards of 10 watts were required, whereas the smallest units, up to around 3 watts, were usually single-ended. In between 3 and 10 watts, both types were found. Valve availability may have been a part of this. Also, a reasonable inference is that at around the 10 watt level, an EL84 push-pull pair generally offered better economics than say a single-ended KT88. There were some variations though. For example, Pye used a single-ended EL34 output in its Mozart amplifier. One may impute a possible engineering reason for this, in that it may have been the best way to fit the unit into what was, for the time, a very compact case. An EL84 push-pull pair plus an additional valve envelope for the phase-splitter might not have fit so well. This amplifier predated the availability of the ECL86, which once it was available offered another pathway to a compact amplifier, in that two valve envelopes encompassed driver, phase-splitter and output pair. By way of another example, the Heathkit MA5 used a single-ended parallel EL84 pair. This looks to have been an appropriate engineering solution for a minimum-valve count economic 5 watt amplifier, and not an initial condition. Early in the stereo era, there was some use of the orthogonal circuit. That too can be seen as an engineering solution, and was probably seen – by those who used it – as the most effective way to use two output valves where that was all that could be afforded.

An interesting case was the Lowther B5F, made from 1948 through to the early 1950s, and which used a single-ended PX25. It was apparently a derivative of the pre-WWII B5, and I’d guess that the “F” suffix indicated that it had negative feedback applied. Post-WWII Lowther was in the hi-fi business, so the B5F was almost certainly designed as a hi-fi amplifier, and not a combination amplifier and coloration machine. Its 5 watt output was seen as adequate for driving horn-loaded speakers in some situations. And given that Lowther used a PX4 push-pull pair for a contemporary 7.5 watt amplifier, and a PX25 push-pull pair for its 15 watt amplifier, one may deduce that a single-ended PX25 was seen as the logical engineering choice for the 5 watt case, and that the single-ended configuration was not an initial condition.

Single-ended transistor amplifiers are a rare species. The major application appears to have been in early hybrid car radios, where for example a single OC16, with transformer-coupled input and output, and without NFB, was used. Strangely, this configuration has yet to develop a latter-day cult following, even though it has some of the required credentials.


Cheers,

[Argus25]

Originally Posted by Synchrodyne

The apparent results also suggest that in some cases, the objective was not a “good amplifier” as such, but a combination amplifier and “coloration machine”.

Single-ended transistor amplifiers are a rare species. The major application appears to have been in early hybrid car radios, where for example a single OC16, with transformer-coupled input and output, and without NFB, was used. Strangely, this configuration has yet to develop a latter-day cult following, even though it has some of the required credentials.


Cheers,

I have a 1920's Grebe radio, where the tone control is labelled "Color Tone".

Have you heard of Synesthesia ? I don't posses it, but wish I did, I'm sure this would help with the appreciation and creation of music:

Synesthesia is a condition in which one sense (for example, hearing) is simultaneously perceived as if by one or more additional senses such as sight. Another form of synesthesia joins objects such as letters, shapes, numbers or people's names with a sensory perception such as smell, color or flavor. The word synesthesia comes from two Greek words, syn (together) and aisthesis (perception). Therefore, synesthesia literally means "joined perception."

I know somebody with this and it is quite astonishing and very difficult to understand or quantify with math and physics.
Creative and artistic folks have synesthesia often it would appear (isn't it horrible to be left out).


On the Class A OC16, I must confess to liking this arrangement as you can see from my radio:

http://worldphaco.com/uploads/THE_EF98-OC16_RADIO.pdf

[Radio Wrangler]

I enjoyed Synchrodyne's excellent post. It needed saying.

In order to be fashionable, a single-ended amplifier must not only be single-ended, it must also LOOK single-ended. So a single-ended amplifier with an active current source providing the power supply current path looks push-pull to those who don't understand circuit design, and so would be seen as unfashionable.

What is fashionable is unfathomable
What is fathomable is unfashionable

A perfect EXOR situation.

David

[Craig Sawyers]

This is an interesting topology that I came across some years ago. It is not single ended, but it is transformer coupled.

http://www.susan-parker.co.uk/zeus.htm

Craig

[kalee20]

Originally Posted by Craig Sawyers

This is an interesting topology that I came across some years ago. It is not single ended, but it is transformer coupled.

http://www.susan-parker.co.uk/zeus.htm

Isn't that just push-pull source followers? Nothing new about that. It's like the Quad II with all the output transformer primary turns in the cathode, none in the anode, and MOSFETised.

[Argus25]

Originally Posted by Craig Sawyers

This is an interesting topology that I came across some years ago. It is not single ended, but it is transformer coupled.

http://www.susan-parker.co.uk/zeus.htm

Craig

I seems to me to be a push pull circuit where the load for the mosfets is in the source rather than drain, and the voltage developed there ends up in series with the drive voltage, so its roughly equivalent to a push pull circuit with two emitter followers, so the voltage gain from the fets is no more than one, and any voltage gain will come from the transformer ratios.

It is not uncommon in early Japanese transistor radios that the emitters of the two output transistors drive the output transformer primary, but to get gain in those cases the drive voltage from the driver transformer is applied to the base-emitter circuits of each transistor and are not significantly affected by the voltage developed across the load in the emitters, so it has no effect that the load is in the emitters, it may as well be in the collector in that case.

If you are driving a base-emitter junction with a transformer, and say its a pnp transistor (in a negative earth system) its handy because you can bolt the collector to ground, avoid insulators, and put the load in the emitter instead.

['LIVEWIRE?']

Hybrid Car Radios often used a similar circuit arrangement to that in the link in Argus25s post, where an OC16(later AD140 or 149) was driven from a low-voltage operated valve.

[kalee20]

Originally Posted by G6Tanuki

One thing that puzzles me is why none of the hi-fi types have ever used the idea of 'sliding bias' to get the most out of their expensive valves. At low audio-level wind the -ve grid-bias up so the valve is passing just-enough current. Then when you want more loudness, reduce the grid-bias so the anode-current rises in order to deliver the greater output.

Because you have to move the grid bias quite slowly, or you get an artefact in the output signal. Sliding the bias from one value to another in 50msec will give a slight thump in the speakers. And even then, if you are playing dramatic music with sudden loud peaks, there will be distortion until the standing current has slewed to its new level.

You can adjust the bias very slowly to overcome the thump, but then you get distortion for longer periods. You could of course have a delay in the system so that you can change bias in anticipation of a sudden increase in signal in a few seconds time, but that's adding massive complication (audio storage, tape loops? Bucket-brigade delay lines? ADC - buffer memory - DAC? All horrors!)

[Herald1360]

Originally Posted by Radio Wrangler

What is fashionable is unfathomable
What is fathomable is unfashionable

A perfect EXOR situation.

David

Does that mean you need an exorcist to take care of audiophools?

[Argus25]

Does that mean you need an exorcist to take care of audiophools?

That is very good, but since the Audiophools operate on a faith based paradigm, I think you need a Priest and it might take both a Priest and an Exorcist together, to complete the job.

[Radio Wrangler]

I keep saying they ought to declare themselves as a religion. There are tax advantages, I understand, and at the prices of their icons, that could make a big difference.

David

[mole42uk]

Originally Posted by Radio Wrangler

Douglas Adams <snip> could have used technical explanations as practiced in hifi boutiques.

He did. The margaritas might have helped though.

[Studio263]

Originally Posted by Synchrodyne

Single-ended transistor amplifiers are a rare species.

Hardly! Every early Sony Trinitron colour TV had one up to the mid 70s and they were not alone; Hitachi used this topology too, as did many other Japanese manufacturers. BRC also used this method for the very popular for their 8000 series sets. That lot alone encompasses millions of individual pieces of equipment.

In an early transistorised colour TV set there was usually a DC supply of around 115V which was stable, so long as you didn't take random gulps of current out of it. A transformer coupled singled ended audio output stage was therefore ideal for the 1 > 2W of audio output that was expected.

[Craig Sawyers]

Meawhile, the Audio Note Kagura https://www.audionote.co.jp/en/produ.../kagura-i.html

Which will set you back £150k.

Craig