[Freo-1]

Where did the 90s design for the Gan400 come from?

[Freo-1]

The two use the same board- they have the same lack of post filter feedback. They are thus both load dependent. This isn't an opinion, it's fact based on their design. All modern class d amps I am familiar with have evolved and have flat frequency response with load because they use a feedback arrangement that puts the output filter (the source of most of the distortion in a class d amp) inside the loop. Do some homework.

No disagreement is not insulting. Calling a well respected and accomplished engineer like Putzkeys a guru is, in my opinion.

I'm aware of this issue.  I believe the design was as intended.   The superior speed and thermal characteristics of the GanFET is being utilized.  The Merrill amp has zero feedback (not needed) .  Apples to oranges comparing GanFET to MOSFETS.


Bruno being referred to as a guru is a sign of respect.   In Australia,  he would be referred to as a boffiin .   Methinks someone is too sensitive.

[mofunz]

Orchard Audio has the most accomplished GaN class d amp currently available

No pun intended?

[goryu]

I'm aware of this issue.  I believe the design was as intended.   The superior speed and thermal characteristics of the GanFET is being utilized.  The Merrill amp has zero feedback (not needed) .  Apples to oranges comparing GanFET to MOSFETS.

The "superior speed and thermal characteristics" do nothing to decrease the output coil distortion or the load dependency, issues all class d amps, regardless of mosfet or GaN have to deal with. While GaN can be run as high as 1.4Mhz, you don't see that in this amp module. I believe it runs around 800kHz, not that much higher than some mosfet class d amps. It's mostly a marketing tool at this point to add some product differentiation. Perhaps in the future designers will find a way to utilize the potential of GaN, but until then, there are better designed products available.

By the way, the Merrill GaN amp has a switching frequency of 400kHz, the Purifi, 500kHz. So much for exploiting faster GaN switching frequencies.

I'm done here. You have demonstrated a lack of understanding of the issues here such that further time discussing this is better spent on other things.  Enjoy your amp. Peace out....

[mresseguie]

 

Well, Freo-1, I am looking forward to your observations and impressions of this new amp. 

[Tyson]

Same class d module used in both it seems...the 90's called- they want they class d amp back...

Here's a subjective review for you from another audio forum:

"If turns out that both the LSA Voyager GAN 350 and the Peachtree GaN400 amplifiers use the same modules from/designed by Elegant Audio Solutions. While both have what seems to be different power supplies, the main amplifier board looks identical. The board uses the GaN Systems HEMT and Elegant Audio Systems is listed as one of their partners. If you go to the GaN Systems website, they have reference boards and detailed information about their reference designs. If you look at the pictures in the reference designs, you will see that those also are from Elegant Audio Solutions.

I had a chance to listen to the Peachtree GaN400 version of this and it at least worked as you would expect out of the box. I (subjectively) had the opportunity to compare the NAD 2200 (upgraded components), GaN400, McIntosh MA252 and the Purifi Eval1. Being an engineer myself and a believer in numbers I have to say that I have bought several components including the NAD 2200 after seeing the great measurements on ASR and was firmly in the camp of that it would be impossible for me to hear a difference in real life with any of these components. I have to say that I was wrong. In my listening tests I found that consistently of the 4, I picked the NAD 2200 as the worst sounding and the Purifi Eval1 as the best sounding with the GaN400/MA252 somewhere in between consistently. My general observations are:

1. The Purifi amp has a level of clarity, detail and high frequency response that I have never heard on any other amplifier including the Benchmark ABH2. The LSA wasn't bad, but didn't really compare to the Purifi. At about 1/3rd the price, you could buy 3 of the Eval1's for the price of the LSA.

2. You need speakers that can resolve the detail. I found that you need speakers that can resolve the details such as Magnepan LRS, Magico S5 vs. a couple of other speakers that I tried that I just couldn't hear any differences.

Last week a couple of friends and I did some blind testing using the Purifi, MA252 and the ABH2 connected to Topping D90SE and Magico S5MKII. One of these days I will post the results of what we did and how, but if you were going to buy something and you aren't able to try out a bunch of amplifiers, buy the Purifi, it is unlike any other amplifier that I have heard and it is the cheapest of the bunch.

Corrected: MA252 not MC252."

Wait, did you just post a subjective opinion?  Isn't that a bit hypocritical? 

Haha, and it's not even your direct experience.  It's someone else's experience.  So a second hand subjective experience, but I guess you feel it's OK when you do it, because it plays in to your pre-existing notions. 

[Stercom]

This whole conversation is so typical of "audio" forums today. Sad really. Music is about emotion and how it makes you feel.....which means it's SUBJECTIVE. If someone buys a piece of audio gear based on measurements I thinks its pretty obvious they don't understand music and believe it's just about making sounds.   

[goryu]

Wait, did you just post a subjective opinion?  Isn't that a bit hypocritical? 

Haha, and it's not even your direct experience.  It's someone else's experience.  So a second hand subjective experience, but I guess you feel it's OK when you do it, because it plays in to your pre-existing notions.

I posted it to show you how worthless subjective opinions are, which you have confirmed by your criticism. When everyone has their own subjective opinion, no opinion really means much of anything.

What can't be argued rationally are facts, like the fact that this amp has a load dependent frequency response due to a rather primitive design by today's standards. If you like that, great. I prefer amps that have a load independent frequency response.

[goryu]

This whole conversation is so typical of "audio" forums today. Sad really. Music is about emotion and how it makes you feel.....which means it's SUBJECTIVE. If someone buys a piece of audio gear based on measurements I thinks its pretty obvious they don't understand music and believe it just about making sounds.

Music is indeed subjective. Electronic reproduction is technology, designed using rational engineering design in an effort to convey the music as accurately as possible. That's what "hi-fi" means. It's pretty obvious when people don't understand engineering and believe it is just about making sounds.

[Stercom]

When everyone has their own subjective opinion, no opinion really means much of anything.

Ok, we get it. Please go away like you said you would

[Freo-1]

The "superior speed and thermal characteristics" do nothing to decrease the output coil distortion or the load dependency, issues all class d amps, regardless of mosfet or GaN have to deal with. While GaN can be run as high as 1.4Mhz, you don't see that in this amp module. I believe it runs around 800kHz, not that much higher than some mosfet class d amps. It's mostly a marketing tool at this point to add some product differentiation. Perhaps in the future designers will find a way to utilize the potential of GaN, but until then, there are better designed products available.

You appear to overlook the fact that because of the increased speed of the GanFET over the MOSFETS,  there is less distortion to overcome.  The white paper from gansystems.com discusses board layouts to minimize induction.   Honestly,  you should challenge the designers regarding their approach.  You actually might learn something.


You should also listen to some GanFET amps with ears as well as measurements.   Measurements are important, but listening is as well.  Read this link again, especially regarding TIM and feedback.


Here are a couple of points from the link:


" Power stages using eGaN FETs come much closer to synthesizing the ideal power digital signal compared with power stages using MOSFETs because they are far superior to silicon MOSFETs in every characteristic that contributes to the distortion of the class-D power stage. The same factors that contribute to deviation from the ideal waveform are responsible for power losses in the amplifier, so the eGaN FET-based class-D audio amplifier will naturally be a much more efficient amplifier."


"A second type of distortion is Transient Intermodulation Distortion (T-IMD), which is the result of introducing frequencies not present in the original audio [1]. With power MOSFETs, higher open-loop impedance from the higher RDS(on), QGD and QGS of the switching device, as well as the longer dead-time, requires excessive feedback to improve audio performance. Excessive feedback limits bandwidth, and introduces intermodulation distortion [3]. The low on-resistance and low capacitances as well as reduced dead time of eGaN FETs enable simpler and lighter feedback providing substantial reduction in T-IMD as well as higher efficiency in Class-D systems."



https://www.eeweb.com/how-to-gan-egan-fets-in-high-performance-class-d-audio-amplifiers/

[Tyson]

I posted it to show you how worthless subjective opinions are, which you have confirmed by your criticism. When everyone has their own subjective opinion, no opinion really means much of anything.

What can't be argued rationally are facts, like the fact that this amp has a load dependent frequency response due to a rather primitive design by today's standards. If you like that, great. I prefer amps that have a load independent frequency response.

All engineering and design choices must be correlated, ultimately, to subjective experience.  I'm not saying measurements don't matter.  I am saying that measurements by themselves cannot tell you how an amp sounds.  Ultimately, subjective experience trumps measurements. 

Bruno has shown he can design a great measuring amplifier, I agree with you on that.  But can he design a great sounding amplifier?  Well, he hasn't yet. 

And, just to point out that I don't have some kind of weird grudge against Class D amps, well I've heard class D amps I like.  The DAC Cherry and Maraschinos sounded really good.  And if we expand the field a bit, I thought the GaNTube amp sounded very good too.  If it seems like I'm picking on Class D, I'm not.  Just the Hypex and Purifi amps because they sound so meh.

[fredgarvin]

This whole conversation is so typical of "audio" forums today. Sad really. Music is about emotion and how it makes you feel.....which means it's SUBJECTIVE. If someone buys a piece of audio gear based on measurements I thinks its pretty obvious they don't understand music and believe it's just about making sounds.

I'm seeing it more as one guy discussing amp design objectively and two other guys getting defensive, holding hands and making rude responses. We come here to discuss, it's not church.

[goryu]

I'm seeing it more as one guy discussing amp design objectively and two other guys getting defensive, holding hands and hurling wise a$$ comments.

Good summary. When you get comments like this:

"Bruno has shown he can design a great measuring amplifier, I agree with you on that.  But can he design a great sounding amplifier?  Well, he hasn't yet.  "

You realize you are dealing with people who have no concept of what subject opinion really means, much less how design and engineering work. When someone has the ignorance and/or ego to think that their subjective opinion actual means anything to anyone other than themselves, and they feel no shame in speaking as if their experience is truth for all, then rationale discussion is impossible. What you get instead is the above- defensiveness, straw men, wise ass comments, obfuscation, religion.

Bruno can't design a great sounding amplifier? There are tens, or perhaps hundreds, of thousands of people who would disagree with your opinion. That's why subjective opinions are meaningless- everyone has one, and for each person, their opinion is right, for others, not so much. I don't need you or anyone else to tell me what is "great sounding" any more than I need you or anyone else to tell me what wine tastes "great or how much salt tastes "best". I don't have the presumption and bad manners to try and tell people what they should prefer when it comes to subjective matters of taste, and I don't believe in pontificating about the subjective performance of audio products as if I am some grand arbiter of taste. The simple fact that there are a wide variety of different types of amps, each type with a rabid following, is proof that subjective opinions are meaningless in the main. I have addressed the technical deficiencies of the GaN amp. You have done nothing to address the technical issues. I have repeatedly told you that if it's what you like, great, enjoy it. I was once told that my opinions end at the tip of my nose. Seems no one has ever told you that.

[Tyson]

I'm seeing it more as one guy discussing amp design objectively and two other guys getting defensive, holding hands and making rude responses. We come here to discuss, it's not church.

I hope I'm not coming off as defensive, that's certainly not my intention.  Maybe I can take the discussion away from GaN and Hypex and talk more abstractly to illustrate my point. 

Let's say I'm a designer of an amp and I'm in the process of designing an amp.  I make a change and after that change the amp measures better.  How will I know if it sounds better?  The only way for me to check my work is to actually listen to it.  Listening is important because not all parts of amplifier design vs. sound quality are fully understood. 

It would be silly to design an amp and never listen to it, I'd almost certainly end up with a mediocre sounding amp. 

So my beef is not with Hypex or Purifi or Gan400's.  My issue is with people that say actually listening to the amp is not necessary or important. 

[goryu]

Who said listening is not important? Most amp designers I am familiar with listen to their amps. Most understand that they aren't selling the amp to themselves and do their best to produce something as close to a wire with gain as possible. Unless of course their market is that consumer base that prefers otherwise.

People here are calling this GaN amp a "move forward". In what way? Does it measure better objectively than the current state of the art class d? No, not even close. Does Merrill's new amp represent a "step forward" objectively? No, it has a SLOWER switching frequency and a higher noise floor. Is the "step forward" a difference in subjective sound? If yes, then that is not a step forward, it's a step in a different direction. It sounds different perhaps, but different doesn't mean "better" for all. It just means different. Some will prefer it, others won't.

Do GaN devices have the potential to actually improve the performance of class d amps? Sure. But there are engineering challenges to surmount before the actual potential of these devices can be implemented in ways that actually improve the performance in objective ways. To date, no one has produced a product that improves upon the performance of mosfet class d amps like Purifi, for the reasons given by Bruno and Lars above. Again, at this stage, it's all marketing the new flavor. It's not objectively better, just subjectively different. Some will like it, some won't. Wash, rinse, repeat.

[Freo-1]

Who said listening is not important? Most amp designers I am familiar with listen to their amps. Most understand that they aren't selling the amp to themselves and do their best to produce something as close to a wire with gain as possible. Unless of course their market is that consumer base that prefers otherwise.

You said that you didn't need to listen to am amp to decide you didn't like it based on measurements.  Well, there are no measurements  provided regarding the Gan400 to arrive at the conclusion you wouldn't like the Gan400.  The LSA measurements don't apply,  as the performance specifications listed between the two are different.


The engineering data provided regarding GanFETs is not subjective,  they are factual.  There is a legitimate ongoing debate regarding feedback.  No one engineer,  no matter how brilliant,  has all the answers.   GanFETs are gaining popularity in SMPS, for obvious reasons.  It does stand to reason they will gain popularity over time, as improvements are made to both design and components.   


I'm going to contact Peachtree to see what additional data they can provide regarding design principles and measurements.  The specs they do provide are comparable with other various high end amps.   The subjective sound from the Gan400 exceeded my expectations by quite a bit.


The design of the driver boards with the Gan400 are different  because the behavior of the output devices exhibit significant differences.  It stands to reason that the focus of the designs would be modified to best take advantage of the GanFET performance.  The white papers posted clearly outlines the inherent advantages.   They obviously have addressed the challenges, or it would not sound as good as it does.

[Tyson]

Who said listening is not important? Most amp designers I am familiar with listen to their amps. Most understand that they aren't selling the amp to themselves and do their best to produce something as close to a wire with gain as possible. Unless of course their market is that consumer base that prefers otherwise.

People here are calling this GaN amp a "move forward". In what way? Does it measure better objectively than the current state of the art class d? No, not even close. Does Merrill's new amp represent a "step forward" objectively? No, it has a SLOWER switching frequency and a higher noise floor. Is the "step forward" a difference in subjective sound? If yes, then that is not a step forward, it's a step in a different direction. It sounds different perhaps, but different doesn't mean "better" for all. It just means different. Some will prefer it, others won't.

Do GaN devices have the potential to actually improve the performance of class d amps? Sure. But there are engineering challenges to surmount before the actual potential of these devices can be implemented in ways that actually improve the performance in objective ways. To date, no one has produced a product that improves upon the performance of mosfet class d amps like Purifi, for the reasons given by Bruno and Lars above. Again, at this stage, it's all marketing the new flavor. It's not objectively better, just subjectively different. Some will like it, some won't. Wash, rinse, repeat.

I agree with all of that.  Also, I'm not the person saying that it's 'a step forward'.  I am saying that the one GaN amp I've heard (the GaNTube, not the GaN400) sounds better to me than the Hypex and Purifi. 

Taking a step back, I think the point that we disagree on is what actually constitutes a neutral sound.  For you, it seems to be 'does the output match the input, just larger/louder".  For me it's "how close does it sound to live, unamplified music in a good acoustic space".  If I insert an amp and it sounds closer to that, then I know the amp is neutral.  If I insert an amp and it's not closer to that, then I know the amp is not neutral.  And that's regardless of the measured specs.

Look man, I get it.  I'm a techie person.  I work in a technical field, and I really like new technology and elegantly engineer products.  Those are all reasons I have owned digital amps in the past.  Plus factor in the incredible efficiency and the fact that they run much cooler than SS amps and I am very much biased toward liking those amps.  But the fact of the matter is, when I put them in my system, things get worse.  A simple, repeatable experiment. 

And that simple experiment is something that no Class D amp has passed (yet).  But based on initial impressions of the GaNTube amp, that one might. 

But I feel like we're talking in circles, because your next point will be about how my repeatable experiment is not blind (let alone double-blind) and that I should not trust me ears unless I can do it double blind, etc.  So lets pretend you said that and that I replied that blind tests are needed for the elimination of bias, but in this case it wouldn't make sense because my initial bias was in favor of the amps not against them, and then you'll say that nothing else matters except being able to pick out which amp is better while blind and then I'll say that if that's true I'd like to see YOU pick out a Purifi amp from a run of the mill class d amp and that you can't, can you?  And if YOU can't pick out the purifi amp blind, then what's the point anyway? 

So, let's just pretend that all that happened because I'm really tired of these redundant unsophisticated arguments. 

[goryu]

You said that you didn't need to listen to am amp to decide you didn't like it based on measurements.  Well, there are no measurements  provided regarding the Gan400 to arrive at the conclusion you wouldn't like the Gan400.  The LSA measurements don't apply,  as the performance specifications listed between the two are different.

As I have said repeatedly, I have been told that the LSA and Peachtree amps use the same amp modules. They have the same lack of post filter feedback. This leads to a frequency response dependent on load regardless of devices, switching frequency, etc. I don't like frequency dependent designs. We have moved passed this in class d.

[goryu]

Taking a step back, I think the point that we disagree on is what actually constitutes a neutral sound.  For you, it seems to be 'does the output match the input, just larger/louder".  For me it's "how close does it sound to live, unamplified music in a good acoustic space".  If I insert an amp and it sounds closer to that, then I know the amp is neutral.  If I insert an amp and it's not closer to that, then I know the amp is not neutral.  And that's regardless of the measured specs.

Your requirement is logically flawed and impossible to satisfy. An amp can't create  "the sound of live, unamplified music in a good acoustic space". All an amp can do is take an input and make it larger, as accurate to the source as possible. The recording, source, the room acoustics, and all the other electronics in the chain play a part in modifying and converting that signal before it gets to you ears, and from there, your own perceptions further interpret the signal. It's an interdependent system. If an amp does anything to modify the signal, it isn't an amp, it's an effects box. The best we can do is to take the source and make it bigger. If the result isn't what you consider "the sound of live, unamplified music in a good acoustic space", there are innumerable possibilities why that is so, from the source material to room acoustics, to distortion in the signal chain. Today's amplifiers produce so little distortion, phase shift, etc., that the problem is almost always somewhere else in the chain, either because of interactions or where distortions are many magnitudes greater.