[Construct]

This is about NON diy class D... that is what I am after.

[roymail]

http://www.audiocircle.com/index.php?topic=76400.0

[jtwrace]

Have fun reading the thread....

[srb]

"Class D"  refers to how much of the 360 degree cycle the amp device is biased on.   Class D pumps out tons of power with little heat compared to class A/B.

Class D amplifiers use a high frequency digitally controlled switching circuit, often Pulse Wave Modulation, to modulate the analog wave.  Since the output devices are either on or off, there would theoretically be no bias current.
 

http://www.audiocircle.com/index.php?topic=76400.0

The link above is to a topic on DIY ampflier modules by a company called "Class D Audio", who use either International Rectifier or Yamaha IC's to control the switching of FET output devices.  They also sell linear power supplies for use with the amplifier boards.
 

Q:  So what class D amps are considered audiophile grade-and what makes them different from the hard sounding PA amps?  Or is this issue related to a PREAMP more than power amp section?  All inputs appreciated!

There are some manufacturers that utilize their own proprietary class D circuits, such as NuForce, but the majority use modules from Hypex (UcD) or Bang & Olufson (ICEpower).  Some manufacturers use the modules in stock form, while others do their own modifications, often to the input stage.
 
Channel Islands Audio uses modified Hypex UcD modules, while Bel Canto, Jeff Roland, Rotel, Virtue Audio (ICEBlock M5001) and Wyred 4 Sound use the B&O ICEpower modules.  There are economy as well as higher-end modules available, and the use of the higher-end modules along with proprietary modifications can differentiate themselves as "audiophile grade".
 
I haven't seen class D employed in preamplification (it's possible it exists somewhere), and the general application is for power amplifiers.
 
Steve

[Kevin Haskins]

Class D amplifiers use a high frequency digitally controlled switching circuit, often Pulse Wave Modulation, to modulate the analog wave.  Since the output devices are either on or off, there would theoretically be no bias current.
 
The link above is to a topic on DIY ampflier modules by a company called "Class D Audio", who use either International Rectifier or Yamaha IC's to control the switching of FET output devices.  They also sell linear power supplies for use with the amplifier boards.
 
There are some manufacturers that utilize their own proprietary class D circuits, such as NuForce, but the majority use modules from Hypex (UcD) or Bang & Olufson (ICEpower).  Some manufacturers use the modules in stock form, while others do their own modifications, often to the input stage.
 
Channel Islands Audio uses modified Hypex UcD modules, while Bel Canto, Jeff Roland, Rotel, Virtue Audio (ICEBlock M5001) and Wyred 4 Sound use the B&O ICEpower modules.  There are economy as well as higher-end modules available, and the use of the higher-end modules along with proprietary modifications can differentiate themselves as "audiophile grade".
 
I haven't seen class D employed in preamplification (it's possible it exists somewhere), and the general application is for power amplifiers.
 
Steve

There are some of the chip companies that have developed them for or mobile applications that would work as preamps but there is no benefit for high-end audio.    They are mainly for saving on battery power in mobile phones, iPods and other mobile gizmos.   

There is also several amplifiers that I'd consider high-end that are not Hypex or Icepower.    TI has some reference designs that are pretty good.   The main issue with Class D designs is dealing with the possible noise issues and layout, grounding and general design of the PCB is much more critical than it is with Class A/B or Class A designs where you don't have to deal with high frequency switching noise.    For that reason many audio companies choose to use the modules from well established brands that have already fought those demons.   

[roymail]

Am I right in assuming that the amps sold by Digital Amplifier Company like the DAC4800 and Cherry amps are in the upper end of Class D amps?

[Bemopti123]

I wonder what is the up with the supposed "original" designs as are taunted by the designer of Axiom speaker and also the guy who is Hephaestus Audio.  Do they get a lone chip and go their own at designing a PCB board?  I am curious. 

[dflee]

Please keep this post going cause I am interested in class D but not into DIY at all. From what I have read in other threads, They talk about a lot of stuff that I have no idea what their talking about, or what chips do what or or or I get so confused, I want my Mommy!

Don

[srb]

Please keep this post going cause I am interested in class D but not into DIY at all. From what I have read in other threads, They talk about a lot of stuff that I have no idea what their talking about, or what chips do what or or or I get so confused, I want my Mommy!

Some have replaced their expensive high-end class A, AB and A/AB amplifiers with class D amplifiers and love the way they sound, while others find them a bit unnatural or sterile sounding.  At least one manufacturer incorporates a tube input buffer to make them sound a little less analytical.  In the end, it comes down to listening and personal preference.
 
The class D ICEpower amplifier modules frequency response tends to be more dependent on the speaker load, so synergy with your speakers is important.  If you are energy conscious they are the most efficient, and for those that must install them in a cabinet, enclosure or rack with minimal ventilation, class D amplifiers make a lot of sense.
 
Steve

[Construct]

Ok simple question.  Crown has class D stereo amps for $399.  What would make them fall short?

[lowtech]

The class D ampllifier's frequency response tends to be more dependent on the speaker load, so synergy with your speakers is important.

This is a generalization that has nothing to do with class-D topology.  The Hypex modules, for example, are completely load insensitive.

[sts9fan]

I am a big fan of the low power Tripath chips   

[srb]

This is a generalization that has nothing to do with class-D topology.  The Hypex modules, for example, are completely load insensitive.

I shouldn't have generalized, and my comment was more directed at the ICEpower amplifiers that can exhibit high frequency attenuation depending on speaker load.
 
Steve

[Niteshade]

I have found that many current Tripath & Class D based amplifiers made today do not have many shortcomings sonically. They're fast and tonally well balanced. Their projection is not typically forward, but more from left to right (more width than depth). Instruments have plenty of room around themselves and vocals are very clean and clear.

These qualities are what I have found to be generally true between three different brands. I have never listened to ICE based amplification.

They tend to have a crispness about them that many will have to get used to. Typically highs are not rolled off at all, yet accurate and I would not call it strident. I call this the garbage in-garbage out effect for the most part. Most stridency is due to lousy recording practices. Fiddles should not make your hair stand on end, for example.

Two big LIKES regarding Class D based amps: Bass management and difficult load handling. Even the little 15 watt Tripath I have does an astounding job for $35.00!

[Feanor]

Am I right in assuming that the amps sold by Digital Amplifier Company like the DAC4800 and Cherry amps are in the upper end of Class D amps?

Looks like some of the DAC equipment, e.g Cherry PLUS at $5k, would qualify as high-end at least based on price.  Another switchng amp company which is definitly high-end by both price and reputation is Spectron.  They use their own, proprietary technology I believe.

[mfsoa]

Feanor,
The Cherry series and all DAC amps are also based on proprietary tech, not UcD or ICE based.
Worth a listen, IMO.

-Mike

[mfsoa]

but more from left to right (more width than depth). Instruments have plenty of room around themselves and vocals are very clean and clear.

Very true in my experience as well.

Smaller instruments on a bigger stage vs. the more tube-like larger instruments on a smaller stage (more overlap)

-Mike

[AmpDesigner333]

There is no such thing as "completely load insensitive"...

Also, it takes wrapping feedback around the output filter to REDUCE load dependence, and in our experience, this affects the sonic quality of the amp in a negative way --- no pun intended.  In fact that "too much negative feedback" sound (sometimes associated with Class-AB amps) is just what you get with all the phase margin compensation that's required to do so.  As long as you're not driving a ridiculously low impedance, a low feedback design sounds so much cleaner than one designed to do better on the bench than into your ears.  We can all agree that bench measurements don't tell the whole story.  Without writing a book on this topic here, lets just say that we started with a design that was excellent on the bench, then tweaked it over several YEARS to get it to sound the way it does.

I have been designing Class-D amps (and listening to them) for more than 20 years, and these opinions are not based on some fly-by-night theories.  There are many misconceptions regarding Class-D technology, and companies that take the risk of designing their own Class-D amps from scratch can easily fall short from inexperience.  Same goes for pre-fab module makers.  I have seen audio companies employ switching power supply designers that have never designed an amplifier in their life to build their "proprietary amp technology".  It is also important to know analog amplifier design in order to make a good "digital" amp.  Yes, I know there will be somebody out there to once again point out that Class-D doesn't mean the same thing as digital, blah, blah, blah.  I designed my first analog audio amp when I was in grade school, by the way...  Designing analog amps is a good way to learn what results in good sonic performance versus good bench measurements.  It can be considered a good lead up to digital amp design --- a prerequisite.

[lowtech]

There is no such thing as "completely load insensitive"...

Also, it takes wrapping feedback around the output filter to REDUCE load dependence, and in our experience, this affects the sonic quality of the amp in a negative way --- no pun intended.

Read the AES white-papers published by Bruno Putzeys, if you haven't alread done so.  IMO (and a lot of others) he's the undisputed expert in the field of class-D amplifier design.  He also happens to build the best sounding class-D amplifiers, IMHO.   

[AmpDesigner333]

Read the AES white-papers published by Bruno Putzeys, if you haven't alread done so.  IMO (and a lot of others) he's the undisputed expert in the field of class-D amplifier design.  He also happens to build the best sounding class-D amplifiers, IMHO.   

I've known Bruno for quite a few years and probably read all his papers, too.  There's no doubt he's a very smart guy --- probably with a genius IQ.  He works very hard and is very dedicated.  Did I just use the word "very" three times???  In the past we have corresponded frequently.  We have similar philosophy when it comes to audio (and even politics, actually, but we can't discuss that here).  Bruno's ears are probably ringing right now, and not from loud music.

Notice that Bruno only has 2 posts on Audio Circle, both on my thread.  One of them is this one where he compliments my design:
http://www.audiocircle.com/index.php?topic=58526.msg520319#msg520319

Anyone worthy of such a title wouldn't be arrogant enough to claim that they are "THE undisputed expert in the field of class-D amplifier design".  Anyway, happy listening!