Atma-Sphere Class D Mono blocks

[Rob181]

Correct. Specs like that says very little.

Please excuse an obvious question - but how do you think the specs of an amp are calculated in the first place - one would assume the final product is measured to produce said specs & made to order examples are measured against these specs to ensure operation within tolerance (if any) or am I missing something here...

 

[Bjorn]

Please excuse an obvious question - but how do you think the specs of an amp are calculated in the first place - one would assume the final product is measured to produce said specs & made to order examples are measured against these specs to ensure operation within tolerance (if any) or am I missing something here...

While specifications are normally based on measurements, they don't show much of the whole picture and can also hide much. Distortion will vary with frequency and power. In order to see the whole picture and really know the distortion characteristics of an amplifier, complete measuement graphs needs to looked at. This is well known among those who understand measurements.

For example doesn't rated power tell us much unless we know the distortion level with that power. Power can be listed with a numerous distortion numbers. Typically used 1% THD is actually very high audible distortion and is far from decent hifi. We don't have strict guidelines of how distortion should be listed. One manufacturer might list the wattage to be with aTHD distortion of 0.1%, another 1% and another 0.01, etc. It's impossible to know unless graphs are actually posted. And seeing how the amplifier rises in distortion with power, it's important.

When someone speaks about the importance of distortion levels in regards to the audible quality and mentions weaknesses in how other amplifier measures in this regards, but fail to show any graphs of their product I personally get very skeptical. Why not post graphs to back up the claims?

 

[Rob181]

While specifications are normally based on measurements, they don't show much of the whole picture and can also hide much. Distortion will vary with frequency and power. In order to see the whole picture and really know the distortion characteristics of an amplifier, complete measuement graphs needs to looked at. This is well known among those who understand measurements.

For example doesn't rated power tell us much unless we know the distortion level with that power. Power can be listed with a numerous distortion numbers. Typically used 1% THD is actually very high audible distortion and is far from decent hifi. We don't have strict guidelines of how distortion should be listed. One manufacturer might list the wattage to be with aTHD distortion of 0.1%, another 1% and another 0.01, etc. It's impossible to know unless graphs are actually posted. And seeing how the amplifier rises in distortion with power, it's important.

When someone speaks about the importance of distortion levels in regards to the audible quality and mentions weaknesses in how other amplifier measures in this regards, but fail to show any graphs of their product I personally get very skeptical. Why not post graphs to back up the claims?

Thank you for the response - that is very much appreciated - how many manufacturers produce said graphs assessed by an independent tester (the good - the bad - the ugly) & produce them for viewing you refer to as standard business practice - haven't seen one yet in my time...

 

[analogsa]

It doesn’t use any global feedback, I think, but does use some local feedback.

No NFB for the output stage, not even local. Plenty of loop NFB in the voltage gain stage. The result is very high distortion compared to most SS amps and an ugly distortion spectrum but still musical and detailed sound. As the output transistors are not under a feedback loop, distortion does not rise with frequency.

An excellent example why harmonic distortion can easily be ignored.

 

[morricab]

No NFB for the output stage, not even local. Plenty of loop NFB in the voltage gain stage. The result is very high distortion compared to most SS amps and an ugly distortion spectrum but still musical and detailed sound. As the output transistors are not under a feedback loop, distortion does not rise with frequency.

An excellent example why harmonic distortion can easily be ignored.

I would say more an example of it not being as simple as just looking at the plots. It needs to be run through a reasonable weighting algorithm based on a human hearing model.

 

[morricab]

No NFB for the output stage, not even local. Plenty of loop NFB in the voltage gain stage. The result is very high distortion compared to most SS amps and an ugly distortion spectrum but still musical and detailed sound. As the output transistors are not under a feedback loop, distortion does not rise with frequency.

An excellent example why harmonic distortion can easily be ignored.

It also shows that transistors in Class AB are pretty nonlinear…would be interesting what this amp would measure like if it was biased fully Class A

 

[Gregadd]

We can say with a high degree of confidence that for most products the marketing department has its thumb on the scales when it comes to specifications.
Tube detractor (and often its proponents) don't think 2nd order harmonics can be ignored. To them it is the THC (active ingredient in marijuana) tube lovers crave. The detractors go far as to call tube fans distortion lovers

 

[Rob181]

I would say more an example of it not being as simple as just looking at the plots. It needs to be run through a reasonable weighting algorithm based on a human hearing model.

Really - algorithms to determine a sound of an amp - is that what AC do - you represent them so you should know.

 

[microstrip]

(...) There are times I might yearn for something a little different. But that is a lot different. (...)

Yes, the high end is an hobby of diversity. It is listener responsibility to choose what he wants in sound reproduction, accepting that there is no single "best" in stereo - the system is too dependent on user perception and experience.

 

[analogsa]

It also shows that transistors in Class AB are pretty nonlinear…would be interesting what this amp would measure like if it was biased fully Class A

Obviously at the Dartzeel's supply voltage rail it is not possible to run class A using a single output pair and adding more pairs together with unavoidable current sharing resistors will turn it into a different amp.

But it is very easy to do this in a simulator. Raising the idle current from a warmish class AB (250mA) to a more reasonable for class A 3A reduces the 2W/8ohms distortion from 0.15% to about 0.0015% and gets rid of practically all harmonics above third. 20W/8ohms 0.005%thd

This is no empty promise. I have been using a similar, slightly more sophisticated diamond buffer output stage using 4 pairs @3A and the measured performance is close to the simulated.

 

[morricab]

Obviously at the Dartzeel's supply voltage rail it is not possible to run class A using a single output pair and adding more pairs together with unavoidable current sharing resistors will turn it into a different amp.

But it is very easy to do this in a simulator. Raising the idle current from a warmish class AB (250mA) to a more reasonable for class A 3A reduces the 2W/8ohms distortion from 0.15% to about 0.0015% and gets rid of practically all harmonics above third. 20W/8ohms 0.005%thd

This is no empty promise. I have been using a similar, slightly more sophisticated diamond buffer output stage using 4 pairs @3A and the measured performance is close to the simulated.

There was a simulation paper from Boyk and Sussmann that indicated similar enhancement running Class A. I guess the DarTZeel would drop from a 100 watt amp to a 25 watt amp in the process and gain a lot of heat sinks but I bet it would sound better.

An interesting example of this is the Plinius SA103, which can run AB or A. The measured distortion and harmonics pattern are far better in Class A.

 

[Gregadd]

The amorphous term" as good as or better "are difficult to define. Perhaps I am confusing it with the term sounds identical. If they sound identical and one has 2nd order harmonic distortion and the other one doesn't, it can be reasonably assumed that distortion is not the defining characteristic.

 

[wil]

It also shows that transistors in Class AB are pretty nonlinear…would be interesting what this amp would measure like if it was biased fully Class A

Don’t some A/B amps operate in class A most, if not all, of the time anyway? — unless they are asked to produce extra-ordinary spl’s in a large room?

 

[morricab]

Don’t some A/B amps operate in class A most, if not all, of the time anyway? — unless they are asked to produce extra-ordinary spl’s in a large room?

No, marketing BS

 

[wil]

No, marketing BS

Could you elaborate on that? I’m just interested in learning.

 

[Kingrex]

Yes, the high end is an hobby of diversity. It is listener responsibility to choose what he wants in sound reproduction, accepting that there is no single "best" in stereo - the system is too dependent on user perception and experience.

And they change over time. Jusy like your car. I'm done with the Cayman S. I want the Mechan S now.

 

[Kingrex]

So, are you saying the DarTZeel doesn’t push you away but also doesn’t engage you in the music?

Measurements are available from Stereophile. It is one of the poorer measuring SS amps of recent vintage. It doesn’t use any global feedback, I think, but does use some local feedback. Overall, a low active part count, low feedback Class AB design.

I have had some highly engaged experiences with the Dartzeel. I think I posted I came to tears a couple times. Of course I don't become that engrossed every time. And there are moments I may be looking for something else. Most of the time my stereo is very satisfying.

 

[morricab]

I have had some highly engaged experiences with the Dartzeel. I think I posted I came to tears a couple times. Of course I don't become that engrossed every time. And there are moments I may be looking for something else. Most of the time my stereo is very satisfying.

So, do you consider the Dart a permanent replacement for the tube amps you have had? Out of curiosity, which SETs have you owned?

 

[Atmasphere]

No- most self-oscillating class D amps should have that characteristic.

Sorry, should have WHAT characteristic?

The characteristic of no distortion increase as frequency is increased. This was very clear in direct response to your supposition. I'm starting to think we have two different communication styles; complicating discourse.

Ralph,
What's your take on my Dartzeel NHB 108 Model 1. How are its distortion profiles. [snip]

Is the Atmasphere Class D amp on paper more clean with less higher order harmonics?

If the Stereophile measurements are correct, it has several times more distortion than our class D and higher ordered harmonics do not appear masked. So yes, on paper our amp appears more 'clean' with less higher ordered harmonics. Based on that my surmise is our amp will sound more relaxed and more detailed/transparent.

Ralph - I am already hand wringing at the current price.
(So please don‘t raise it on my account…)

How does your class Class-D differ from Purifi based ones?
(It looks like the distortion profile of the Purifi is not decreasing monotonically.)

The Purifi module is a tour de force of class D design and has the competition locked out due to a variety of patents. I've seen really variable charts of 'measurements' which do not say the same thing so I don't know who to trust. Some people have already done comparisons and preferred our amp, one stating that it was at 'an entirely different level'. That was either on this thread or on one of several on Audiogon. Execution plays an enormous role in the performance of any amplifier and the Purifi is no exception, not surprisingly with really variable results . We didn't leave that to chance in our amp.

 

[Atmasphere]

When someone speaks about the importance of distortion levels in regards to the audible quality and mentions weaknesses in how other amplifier measures in this regards, but fail to show any graphs of their product I personally get very skeptical. Why not post graphs to back up the claims?

How does a graph back up the claims?

The issue is who is doing the posting. How was the graph generated? This is the internet after all- you can put up anything you like. It seems odd to me that someone would take a graph any more seriously than text; this seems like a trust issue to me.