The Truth About Digital (Class D) Amplifiers

[krabapple]

You can drive two different cars on a highway at 60mph and not notice much of a difference until you are faced with external influences such as weather conditions, road obstacles requiring quick maneuvers that the inferior car may not be up to task to handle.

My point is, alot of it depends on load conditions, quality of source material played back, quality of listening space, quality of speakers, discernability of your ears.

I personally purchased the older version of the Panasonic Class D receiver b/c I felt the amp section was so bad that it was a worthy piece of gear to add to my reference system for comparative purposes. I will be referencing this in up and coming listening tests once the primary theater room in the Showcase home is completed.

I do hope these listening comparisons will be blind, or better yet double-blind. Because that matters *at least as much if not more than* load conditions, quality of source material, speakers, etc. The Audioholics site is exemplary in so many ways, but a continued use of an inherently flawed subjective evaluation paradigm -- sighted evaluation of *sound* -- isn't one of them. If sighted evaluation is used as a basis to report that a Class D amp sounds worse (or better) than other topologies, that would be doing a disservice to readers.

 

[gene]

but a continued use of an inherently flawed subjective evaluation paradigm -- sighted evaluation of *sound* -- isn't one of them. If sighted evaluation is used as a basis to report that a Class D amp sounds worse (or better) than other topologies, that would be doing a disservice to readers.

And why would you assume we would NOT do a true DBT? I promise to give all of our listening panels black eyes before we begin our sessions

 

[Sleestack]

Here's what Bruno had to say regarding digitally controlled amps like the Tact:



I read this article in a journal, though what I have posted (and more) is available on Bryston's website. Here's the link:

http://www.bryston.ca/newsletters/74_files/vol7is4.html

Thanks. Although his spostion should apply to the newer TACT products, the product he is referring to is their 1st generation product.

 

[gene]

Thanks. Although his spostion should apply to the newer TACT products, the product he is referring to is their 1st generation product.

It seems like Bryston took quotes from an old Editorial Bruno wrote for us to post on their site as arguments against Class D amps for full range applications. I don't recall giving them permission for this and I can see why this has steamed so many folks on our forums. I will look into this matter.

[Edit: 6/28/06] We just verified Bryston did have Bruno's consent on posting his comments.

 

[Sleestack]

Yes I agree which is why its important to NOT put words in peoples mouths or misquote them when they write articles

I should have been more specific. I do realize the article pointed out that it was based on professional opinion and your own limited test sample. I just felt that without looking at a broad range of current products and more parameters, the article was premature and did little other than to create negative sentiment toward digital amps. Nevertheless, I don't think it was your intent. That's just how I felt after reading the article.

I do hope that you do look at a broad range of current products available and would even consider letting you examine some of the TACT gear I have if you don't have it readily availabe.

 

[Bruno Putzeys]

I'm just back from vacation and saw Gene pointed me to this thread. I'll keep my answer short, touching on a few points I saw mentioned during the one minute (sorry short on time!!!) spent reviewing the thread.

The point about the 60mph car is exactly made. With more or less ohmic speakers, a well designed open loop amplifier with a stable supply sounds fine. If it floats your boat... Point is, one can get equal or better results more cheaply and reliably (ie also in difficult loads) if one allows oneself to apply feedback. It removes the requirement for a stabilised supply (effectively reducing cost and increasing power efficiency), allows the power stage to be optimised for efficiency and EMI rather than distortion and potentially (if the feedback is taken after the output filter) guarantees a flat frequency respons.

The IcePower modules are all analogue and use feedback (so far only partially after the output filter). Although sometimes the company calls the amplifiers digital (popular verbage, goes well with the clientele), the designer makes no bones about the fact that the amplifiers are analogue and that this design method is preferable on technical grounds. There is total agreement between myself and KN on this.

Like myself, the designer of the TacT Millennium, Lars Risbo (now mainly active for TI), likes toying with all kinds of topologies and prefers to draw his conclusions based on experience and fact, not ideology. He's recently designed a new TacT -oops- Lungdorf amplifier which, although sold as "digital" for commercial reasons, is highly analogue and uses a most innovative feedback and modulator design that is bound to place this amp on the map in a serious way. The design is outlined in his excellent AES paper (preprint 6693, see http://www.aes.org/events/120/papers/session.cfm?code=P8). Ironically, just after the designer of world's first commercial "all-digital" amp showed how to improve modulator linearity in an analogue circuit, I went on to present a *digital* modulator (preprint 6694) with improved behaviour in the same respect. Much to our delight, we both explained the same problem, arrived at the same conclusions, but got round it in very different ways (my proposed solution is equally amenable to analogue implementation btw).

This particular paper session started with a completely rewritten re-run (preprint 6690) of the argument made in the audioholics article that is the object of this thread. I hope it explains the rationale behind some of my assertions with sufficient clarity as to make obvious that the arguments are technical, not ideological in nature.

For copyright reasons (the AES needs to get its funds from somewhere) I cannot repost the preprint in its exact AES format. A reformatted (but otherwise unedited) version can be found at http://www.hypex.nl/docs/allamps hypex layout.pdf . Note the co-authorship between competitors, as another illustration that the class D world is way too small for gripes among colleagues.

Likewise, several designers of class D products and chips who have all invested in "digital" designs have privately stated their agreement with this paper, but you will understand that they have the right to publicly defend their investment in the way they see fit. Note though how several manufacturers are now slowly introducing partial feedback measures into their designs. Feedback (and hence analogue processing) is gaining ground, but those who previously jumped on the "all digital" bandwagon with much ado will do so more quietly than those who have started off analogue in the first place. In general, it's always better to define or change one's opinions to reflect reality rather than to try doing the converse, but often pre-existing commercial interests make it hard to do about-turns just like that. In a world riddled with preconceptions (and the insistent belief that all thinkable opinions are equal in value) it can cost a company dearly to do so too obviously.

In closing I'd like to clarify my professional status. I worked for Philips (division changes name constantly) from 1995 to early 2005. During this period I designed several class D circuits, both analogue and digital, including the UcD circuit for which the company holds the patent. I'm now an employee at Hypex. I don't do consulting work. If there's one way to land into an IP quagmire it's consultancy.

I've still some 200 emails to work through so I'm leaving it here. I might check the thread later but I'll only reply to questions to which the answer hasn't been given in this reply or in the referenced papers.

 

[Sleestack]

Bruno

Thanks for chiming in. Although many of the technical aspects were a bit over my head, I really enjoyed reading your AES article as it helped me better understand the context of the audiohlics article. This topic is obviously of great interest to me as I do use the TACT BOZ amps.

 

[gene]

I should have been more specific. I do realize the article pointed out that it was based on professional opinion and your own limited test sample. I just felt that without looking at a broad range of current products and more parameters, the article was premature and did little other than to create negative sentiment toward digital amps. Nevertheless, I don't think it was your intent. That's just how I felt after reading the article.

I do hope that you do look at a broad range of current products available and would even consider letting you examine some of the TACT gear I have if you don't have it readily availabe.

Though we haven't yet written much on this topic, we have been studying many of the newer Class D amps. Many of them are poorly designed or laid out. I don't have much experience with TACT and I appreciate your offer. I may contact them in the near future to arrange for a product review.

 

[Dan Banquer]

Class D amps

I would like to add a few things to this thread if I may. After all I am sort of responsible for one of these articles.

First and foremost, the majority of the articles I attempt to write are to inform people. Before the article was published I'll guess that most of you were not aware of what was really coming out of some of these units and why. Now you may have some idea and quite possibly you may well be better educated for it.

I don't have a "political" agenda and nor do I have a business agenda. If there is any agenda at Audioholics it's to explore engineering issues with FACTS, not snake oil. I do like from time to time to write articles that either lay waste to some of the gross misinformation that goes on in this wonderful business, or articles of a more technical nature that inform readers of issues not readily talked about in the consumer audio press.

Please note that the audio magazine I write for is Audioholics, and why is that? It's because Gene is interested in getting these type of articles out to you and archived on this web site. He also appears to be the only person in consumer audio that wants these type of articles, which says alot about consumer audio and Gene DellaSalla. If I had more spare time I would write more.
d.b.

 

[Sleestack]

First and foremost, the majority of the articles I attempt to write are to inform people. Before the article was published I'll guess that most of you were not aware of what was really coming out of some of these units and why. Now you may have some idea and quite possibly you may well be better educated for it.
I don't have a "political" agenda and nor do I have a business agenda. If there is any agenda at Audioholics it's to explore engineering issues with FACTS, not snake oil. I do like from time to time to write articles that either lay waste to some of the gross misinformation that goes on in this wonderful business, or articles of a more technical nature that inform readers of issues not readily talked about in the consumer audio press.

I'm the one who mentioned an agenda in this thread and it was in reference to Bruno, although I do believe his agenda is inherent in his work... he wants to build a better amp. I hope you didn't misunderstand. I simply felt the overall tone of the Audioholics article was negative. Furthermore, while it provided an interesting read on some of the issues, it was not very useful in that only a limited number of products were looked at. Hopefully more informative tests and reports are forthcoming.

 

[Bruno Putzeys]

I like how you put my "agenda". Indeed, I want to build better amps. Apart from that I also want to see better amps built by other people. The latter is why in some publications or forums I sometimes divulge more technical secrets than is healthy for my own commercial interest. Call me a commie.

The reason why the article sounds the way it sounds is that it was edited from a usenet post I made in answer to a question concerning digital amps. It would probably have made sense to rewrite it before publishing it as a proper article. If I had to rewrite the message of the article in a more positive tone, it'd go something like this: Class D is the future of audio amplification. If we all think twice and avoid falling into the trap of thinking they're digital (and designing accordingly) we can really get this revolution up to speed.

 

[ar-t]

As a manufacturer that uses Class D modules that use Bruno's topology, as well as others, I see nothing to take issue with in his postings. Even those that seem to have been taken out of context.

Pat

 

[Sleestack]

I like how you put my "agenda". Indeed, I want to build better amps. Apart from that I also want to see better amps built by other people. The latter is why in some publications or forums I sometimes divulge more technical secrets than is healthy for my own commercial interest. Call me a commie.

The reason why the article sounds the way it sounds is that it was edited from a usenet post I made in answer to a question concerning digital amps. It would probably have made sense to rewrite it before publishing it as a proper article. If I had to rewrite the message of the article in a more positive tone, it'd go something like this: Class D is the future of audio amplification. If we all think twice and avoid falling into the trap of thinking they're digital (and designing accordingly) we can really get this revolution up to speed.

Bruno,

Thanks again for posting. I applaud your agenda and hope that get a chance to hear some of your work some day.

Your explanation of the context of the article also helped me understand why it left me with the impression that it did.

I am curious if you have had a chance to hear the new TACT Boz amps. You must know Boz. As he uses a no-feedback design, I would assume they have the same shortcomings you have touched upon. That being said, I have been very happy using them to bi-amp each of my Andra IIs and wonder if you have any comment on how they sound to you. I understand they are slightly different than the older Millenium.

 

[bgrounds]

PS-Audio

Any information on PS-Audio G Series amps?

 

[silversurfer]

Any information on PS-Audio G Series amps?

Looks like they use ICEpower modules.

 

[Sleestack]

I'll be picking up some of the new BelCanto Ref.1000 monoblocks, which use the ICEpower platform. The specs on the amps are quite impressive, so I'm looking forwar to hearing how they sound compared to the 216/2000 setup.

 

[silversurfer]

It is my understanding that the Bel Cantos use standard ICEpower modules with nothing added before or after the module. Many of the "high-end" companies do the exact same thing. Plug and Play.

I will be very interested in your thoughts.

 

[ar-t]

Yes, many do. But I can assure that it is not all.

Pat

 

[silversurfer]

Yes, many do. But I can assure that it is not all.

Pat

Understood(didn't I type "many"? ). I would love to know who does what.

 

[Sleestack]

It is my understanding that the Bel Cantos use standard ICEpower modules with nothing added before or after the module. Many of the "high-end" companies do the exact same thing. Plug and Play.

I will be very interested in your thoughts.

What do you mean by nothing? It almost sounds like you are saying that other than the form factor, the ICEpower module incorporates all the necessary parts for input, amplification and output. I don't think that is the case, but I really don't know. In any case, at 500W at 8ohms, 4 monoblocks are sounding nice right now.