Why collector/drain output stages are so rare???

[lineup]

Originally posted by Eva
ZEN V5 is a really funny circuit, it will amplify power supply ripple nearly as much as the input signal,
and output stage bias current is not regulated and completely dependent on the magnitude of the supply rails instead.

Such a circuit is only to be powered by attaching the PCB directly to the output binding posts of a high quality bench power supply

Originally posted by Nelson Pass
http://www.passdiy.com/projects/zenv5-7.htm

Amplifiers: Zen Variations - Part 5 by Nelson Pass (c) 2003
The Complementary Zen

Quiet, Please
The ZV5 is much more sensitive to supply noise than other amplifiers, and this is because the input AC signal is referenced not to ground, but to the AC value of the supply rails. To the extent that the supply ripple noise is symmetric, there is some cancellation, but what remains is still amplified by the gain of the amplifier. Typically you will find that the value of the noise on the supply is comparable to the noise that appears at the output.

This means that the Complementary Zen requires a very quiet supply.
Given a quiet supply however, the noise figures are astonishingly good on our samples; on the order of 18 uV over the audio band.

As a rule, commercial manufacturers are happy with 10 times that amount.

Mister Pass have not often really cheated
when he has published his amplifiers.
Whether they are single end Class A or like here Complementary Class A
or X or whatever
he publish the recommendations of supply for his Amplifiers.


It is not like we can see some of us others do.
Just publish an interesting amp schematic with little or nothing else.
Some even do not tell the level of V+ positive and V- negative.
Which is one essential information.


have no fear about Zen v5
If you read the rather long and detailed article by Pass
you will know how to build a very good amplifier:
Amplifiers: Zen Variations - Part 5 by Nelson Pass (c) 2003
The Complementary Zen


This is the final verdict from the constructor.
The Conclusion.
And I know Nelson is not a man often to
say what he does not really mean or say too much.

The Complementary Zen

Conclusion
This is a very good sounding amp. It sounds most like ZV2, but with a bit more power and a little more punch on the bottom. Whether it is actually better or not is a matter of taste. At higher bias currents, it definitely delivers more into lower impedance loads. As of this writing I’m driving a set of Jordan 92’s mounted into rear loaded horns (the upcoming J-Horn project).
I believe I could live with this for a couple of years.

Me, lineup say:
I believe many of us could live with one Zen-v5 amplifier for very many years.

[tiefbassuebertr]

Originally Posted by Leolabs

Is it because of stability issue???

A good question. Here very few examples:
1) Moxtone Lab
Simple 2 stage PP fet amplifier (post #1)
2) ABACUS (theoretical background, advantages in opposite to power followers, only in German)
http://www.abacus-electronics.de/fil...no-prinzip.pdf
and here additional URLs:http://www.abacus-electronics.de/files/spiegeltest.pdf
http://www.abacus-electronics.de/files/info-ampino.pdf
http://www.abacus-electronics.de/files/info-60-120b.pdf
ABACUS electronics | ABACUS
Abacus rieder 60-120 amplifier
3) "A new Class AB design"
from the authors Wim de Jager,
van Tuy & van der Ven (developed at University of Twente)
Electronics World, December 1999, Page 982; only schematic:
http://francis.audio2.pagesperso-ora.../C31_Jager.gif
4) LEF (Carlos Candeias, CEC/SANYO) detailled schematic wanted - who can post this ??
C.E.C AMP3300R (3300 R Amp 3300R Amp3300) and AMP5300R LEF Class A schematic wanted
LEF (Load Effect Free) Amplifier technology
5) Linn power amplifier "Klout" (power stage) schema see attachement:

[ghaudio]

Originally Posted by Christer

But nobody so far has given a clue what the supposed advantages are, except from the obvious one that we get gain also in the output stage.

From my point of view, the main advantage is that such an output stage is a current source as opposed to the traditional source/emitter-follower stages being voltage sources. The reason is as follows:

For a current source, the power supplied to the speaker is P = R x Isquared
For a voltage source, the power supplied to the speaker is P = Usquared/R

Now the resistance of the speaker is never constant. When the resistance changes, the power supplied to it changes as well. In the case of the current source, the power changes propertionally to the resistance. In the case of the voltage source the power changes reziprocally to the resistance. Thus the voltage source incurs a lot more distortion than the current source.

The current source is less susceptible to the phase shift incurred by the speaker as it is located in the voltage build up by the speakers resistance. This voltage, however, has little effect on the output stage. A voltage source, in contrast, has to supply the blind currents as well.

To sum it up: a current source has much better control over the speaker than a voltage source.

Another advantage of the current source is that one does not have to provide so much power. In the case of the voltage source the current increases when the speaker resistance drops, which is why those amplifiers need a lot of current (and thus: power) reserve.

I think the 'warm' sound of the tube amplifiers has to be mostly credited to the fact that they are current sources, and less to the even harmonics distortion.

Regards,

gh

[DF96]

Thus the voltage source incurs a lot more distortion than the current source.

Really? So changes in R cause less 'distortion' than changes in 1/R?

To sum it up: a current source has much better control over the speaker than a voltage source.

Really? It has no control over bass unit damping, whereas a voltage source has some control.

Another advantage of the current source is that one does not have to provide so much power.

Really? At a speaker impedance peak, a current source has to provide more power and may have headroom problems as it will still be limited by the DC voltage supply.

I suppose it might be possible to design a speaker to run from a current source, but most people design for a voltage source so the amp should approximate this. Most amps do this by sampling the output voltage for negative feedback. For a current source amp you would need a collector/drain output and sample the current, not the voltage. I have never seen this done for audio, although someone somewhere may have tried it.

[davidsrsb]

The old Musical Fidelity A1 was a good example. Almost full Class A operation avoided any cross conduction problems. It had a 700mA quiescent and current limited at about 2A due to limited base drive, without any current limit circuit. The low voltage losses in the output allowed 20V rails and made the efficiency just about OK.

[ghaudio]

Originally Posted by DF96

Really? So changes in R cause less 'distortion' than changes in 1/R?

I might be mistaken, but AFAIK speakers are current sinks, at least the dynamic ones. The magnetic field is built by the current running through the coil, not the voltage. Thus, if the current is constant, there is no change in the force to the membrane, no matter what the impedance is, and thus no distortion. With the voltage source, the current changes, and thus the force on the membrane, and thus there is distortion.

Originally Posted by DF96

Really? It has no control over bass unit damping, whereas a voltage source has some control.

Please explain bass unit damping, I'm not familiar with this issue.

Originally Posted by DF96

Really? At a speaker impedance peak, a current source has to provide more power and may have headroom problems as it will still be limited by the DC voltage supply.

Sure, but how far can such an impedance go up? Going from 4Ohm to 6Ohm is much less critical for the current source than a going from 4Ohm to 2Ohm is for the voltage source.

Regards,

gh

[ghaudio]

Originally Posted by DF96

For a current source amp you would need a collector/drain output and sample the current, not the voltage. I have never seen this done for audio, although someone somewhere may have tried it.

How sad. I think this is the only correct way to design a power amp.

Regards,

gh

[DF96]

Originally Posted by ghaudio

if the current is constant, there is no change in the force to the membrane, no matter what the impedance is, and thus no distortion.

There is more to speaker linearity than voice coil force. The force is not fully linear with current, as it also depends on where in the gap the coil happens to be - the magnetic field is not completely uniform. Then the suspension is not completely uniform. And on top of this, the actual cone movement depends on things like resonances, so a big cone movement can result from a small current at the right frequency.

Please explain bass unit damping, I'm not familiar with this issue

As the voice coil moves in the magnetic field it gets a voltage induced in it, sometimes known as back emf. This creates a current which opposes the driving current. This damps down the speaker resonance. The magnitude of this effect depends partly on amplifier output impedance. Most, but not all, speaker designs rely on this damping. It is completely absent in a current drive amplifier, so this will only work with speakers specifically designed for current drive. Otherwise you will get a huge bass peak at the resonance, which may sound fine for reggae but is not hi-fi.

Sure, but how far can such an impedance go up?

At a resonance the speaker impedance can easily increase by a factor of 3 or 4, maybe more. So you get 3 or 4 times as much power fed in at just the point you don't want it, because the speaker output will be peaking at the resonance too.

How sad. I think this is the only correct way to design a power amp.

Thinking that is your privilege, but I think you will find that almost everyone takes exactly the opposite view for the reasons I have outlined above. I think you need to find out some more about speakers, mechanical and electrical resonance, electromechanical systems etc. Then you might change your mind.

[tiefbassuebertr]

Here was a thread with follow topic, if I recall right
(the exactly phrasing was slightly different):
Bob Cordell Interview: Emitter follower power stages vs. CFP (Sziklai) power stages
Who can post the appropriate URL?

[DF96]

Emitter follower and CFP are both followers, even though the output comes from the collector in CFP.