[RonN5]

There has been a pretty healthy debate raging on over at Audiogon regarding the sonic, technical and now the reliability of class d vs class a and a/b.

Here is just one of the threads  https://forum.audiogon.com/discussions/class-d-is-affordable-and-sounds-as-good-or-better-the-ss-valve-why-buy

Today, a posting was made asserting that class d amplifiers are not reliable and can not be fixed if anything goes wrong.  https://forum.audiogon.com/discussions/the-hifi-trajectory-of-class-d-amplifiers   (the quote is below)

"Class D amplifiers are similar to the switch mode power supplies in computers and therefore non user serviceable.   Failures are simply dealt with by warranty replacement or purchasing a new amplifier.   The technology of professional high power Class D amplifiers is so complex and fragile that reliable operation of after 7 years should not be expected".

I'm hoping that Tommy and other experts can shed some light on what seems to be a misrepresentation of the actual situation.

[Freo-1]

I don't see much misrepresentation on the thread. I do see a varied set of mostly opinions regarding the benefits and limitations of Class D amps.  Hadn't thought much about the longevity of the Class D modules, but that could be an issue.  Back in the mid 2000's, Yamaha released a high end Class d amp (MX-D1) that sounded very good, but it was prone to failure after a period of time.   I've had several Class D amps over the years (including the Yamaha amp mentioned), and did not experience any reliability issues with them.  The issue regarding getting them fixed may be valid. in that it would likely be a whole module replacement if something went wrong. 


The post from eddiet on your linked thread has some good information.  Not sure it means a lot to the average user for daily use of one's system. 


A lot of the earlier Class D amps did sound rather dreadful.  They certainly have improved remarkably over time.  Haven't had the pleasure of listening to the Cherry amps, but have no reason to think that they wouldn't sound very good, and likely sound excellent.  I had a set of Primare Class D amps that sounded very good for a year or so. 


IMHO, the Devialet amps are the best sounding switching amps I've heard, and are among the best sounding amps of any given topology available at any price.  They are a hybrid design, running voltage in Class A, and provide the current via a Class D module.  It's a bit mysterious how they do it, but it works very well. 

[OzarkTom]

Four years owning Tommy's Cherry Mono amps without any problems.

[Tyson]

I've heard the Davialet at RMAF a couple times now.  IMO, Tommy's are better sounding.  And that's from a tube guy! 

[AmpDesigner333]

Guys, first, thanks for the compliments!  I started Digital Amp Co in 1996, and that was after about a decade of amplifier design, including Class-A, Class-AB, Class-G, Class-H, and several variations of Class-D (analog in, digital in, analog modulation, digital modulation, open loop, closed loop).  I’d like to think the experience, hard work, and sound (pun intended) engineering practices has produced what will someday be referred to as the best sounding amplifiers in the world.

Second, modern Class-D designs vary wildly in topology and implementation. Unfortunately, the biggest market segment for Class-D has been cost centric.  Some of the sonic problems with the “old” Class-D carried over from the old stuff to the new stuff, namely modulation techniques, control techniques, and general topologies.

Third, reliability wise, Cherry Amps have been out there for about 12 years now, and our reliability is amazing!  There are hundreds of Cherry Amps still powering systems reliably after more than a decade of use.

Fourth, we do things differently.   There’s a reason that most Class-D amplifiers on the market are pre-fab module based.  Class-D design for super high performance is NOT easy. Engineers with decades of analog amplifier design experience have tried Class-D designs and failed miserably.  I hope to add more to this thread later.  Thanks, Ron, for starting this very interesting topic (:

[Wayner]

Pretty difficult to work on a circuit board that is loaded with SMDs (surface mounted devices) that really only machines can accurately and efficiently place. Some of the components are so small and so fragile, the only practical fix is to remove and replace the entire board, and that in itself could be challenging.

I own the S.M.S.L 98E amp and it is very tiny but spec'd at 160 watts per channel @ 4 ohms. While the the distortion levels do increase with power output, the "real" usable wattage is extremely clean and detailed, ideal for smaller listening environments.

If the unit were to ever die, one would shell out another $110 and get a new one. Much cheaper then a repair bill.

[AmpDesigner333]

A few reasons for the bum rap that Class-D gets (will elaborate later)....

1. Chi-Fi and chip amps.  While it’s very involved to design a quality Class-D amp from scratch, it’s easier than ever to throw an amp together with an off-the-shelf Class-D chip and a few external parts. These chips are riddled with issues and not intended for quality audio systems. They are fine for boom boxes.  Typical implementation is just a reference design copied verbatim.

2. Attempting to force fit Class-D amps with the same control mechanisms as old analog amps.  This is a topic difficult to summarize without getting into details.  Bottom line is that almost all Class-D amps have an output filter that adds serious issues to feedback circuit design.

3. Design for efficiency.  The original motivation behind Class-D was purely conservation of power.  Class-D is theoretically capable of 100% efficiency, meaning that no power is wasted (dissipated) by the amp itself, and all the power consumed goes into the load (speaker). In practice, efficiency over 90% is easily attainable, but this can be at the expense of audio performance.

4. Switching amp design is related to switching power supply design, so it was common for audio companies to hire power supply designers to develop switching amps. However, these specialized engineers typically had no experience designing audio circuits/systems. The result was a bunch of awful Class-D amps on the market.

5. Single ended Class-D suffers from a problem called the “power supply pump effect”.

6. Class-D suffers from market acceptance for the same reasons any new technology is initially shunned by a dug-in specialized market segment such as audio.

[Freo-1]

My issue with Class D amps is that I've never been all that enamored with the treble performance.  I  suspect that is because that all Class D amps employ brick wall filtering to get rid of unwanted ultrasonic noise.


The one switching amp design that does not have to employ this filtering is Devialet.  That is because the Devialet Analog Digital Hybrid (ADH) is not Class D,  I think this is one of the major reasons they sound as good as they do. 

[RonN5]

As the owner of a new 2Cherry... the amp has a beautiful, sensuous, liquid, and yet still very detailed top end without a hint harshness.  If you haven't cared for class d up to now, this one may well change your mind!!

[jonbee]

I've used about 20 quality Class D amps (out of ~100 total, of all topologies), beginning with a Bel Canto EVO nearly 20 years ago. All of the older class D iterations had sonic high points and flaws, some quite obvious, but the improvement trend has been clear throughout.
To me, the N-Core was where the Class D started coming out of the shadows of very good non-class d amps, and IMO the Cherries now stand comparison with any real-world priced (< ~$20K) older tech amps. I like the Nuprime amps quite well too.
The Cherries are not tube amps, SE or PP, nor class A room heaters, etc., and with synergy being king in this hobby there are systems, rooms and listeners where preferences go to other topologies.
That will always be the case, and those choices are completely, 100% valid based on listener satisfaction, the only real yardstick in the end.
The technical arguments for me are a not very relevant sideshow- all competing technologies continue to have technical issues and limitations which have been argued back and forth for decades, far longer than the 50+ years I've been in the hobby. How a designer faces them and how the listeners weigh them is personal choice.  Techies sneer at tiny SE triodes, etc., for their obvious engineering shortcomings, but if they've heard one in a proper system the jokes are over.
So for me, Class D is now a truly viable option in the finest systems, but far from the only option, of course.
In the end, truth to the music trumps all.

[8T_BoCO]

I've only heard good things about the Cherry amps, but they are out of my price range.

My first experience with high-power Class D is the IcePower IceEdge 1200AS2 and I'm quite impressed, especially after it has broken in over 5 months.  I came from a 15w/ch Push Pull EL84 Mapleshade-modified Heathkit with lovely mids & highs, with a bit of a soft bottom end.
My worries about high frequency harshness in the 1200AS2 were unfounded.  I've not experienced such authoritative control over LF drivers before, so much so that I no longer use a powered subwoofer.  For the same (limited) budget, I do not think it's possible to acquire similarly capable tube or SS amps. 

The Class D detractors on the AG thread cited by the OP seem to assume people can afford to spend kilobucks on the finest SS or tube Class A, AB.

[RonN5]

One of the things that comes across in everything you read about the Digital Amplifier Companies is with all of their Cherry amplifiers, the sonic signature is pretty similar.  If you follow this link...here is one of their amps for $1200...and I wouldn't be surprised if it was on the summer sale at a more attractive price....and if you watch for the next kickstarter...there are always deals there as well.

https://www.cherryamp.com/product-page/stereo-maraschino-stm

[AmpDesigner333]

One of the things that comes across in everything you read about the Digital Amplifier Companies is with all of their Cherry amplifiers, the sonic signature is pretty similar.  If you follow this link...here is one of their amps for $1200...and I wouldn't be surprised if it was on the summer sale at a more attractive price....and if you watch for the next kickstarter...there are always deals there as well.

https://www.cherryamp.com/product-page/stereo-maraschino-stm

Details on our current Pre-Summer Sale by emailing Support@DigitalAmp.com.  LOTS of demos, too (:

[barrows]

Ummm...  As to Devialet, and the notion that it is somehow unique in its topology, really?  Most class D amplifiers of which I am aware are all hybrids: that is, they use a traditional class A (or perhaps A/B) linear input stage, mated to a class D, switching, output stage.  Now how they divide up the gain amongst these two stages may vary, but I keep hearing how Devialet is "different" but it really does not appear to be so.  And indeed, as Devialet uses a switching output stage, they must have an output filter to remove most of the switching noise, although describing class D amplifier output filters as "brick wall" is a little harsh, most of them are much gentler than anything I would consider a "brick wall".  Hypex publishes frequency response curves for their products, and the high frequency roll off can be seen there.

[AmpDesigner333]

Details on our current Pre-Summer Sale by emailing Support@DigitalAmp.com.  LOTS of demos, too (:

We currently have a Stereo Maraschino (STM) demo for under $800!

Power supply can be 30V, 36V, or 48V.

[Freo-1]

Ummm...  As to Devialet, and the notion that it is somehow unique in its topology, really?  Most class D amplifiers of which I am aware are all hybrids: that is, they use a traditional class A (or perhaps A/B) linear input stage, mated to a class D, switching, output stage.  Now how they divide up the gain amongst these two stages may vary, but I keep hearing how Devialet is "different" but it really does not appear to be so.  And indeed, as Devialet uses a switching output stage, they must have an output filter to remove most of the switching noise, although describing class D amplifier output filters as "brick wall" is a little harsh, most of them are much gentler than anything I would consider a "brick wall".  Hypex publishes frequency response curves for their products, and the high frequency roll off can be seen there.

The Devialet is a unique topology.  Here is link to the white paper which explains how it works:


From the Stereophile review of the D-Premier:



Conventional class-D amplifiers suffer from high levels of ultrasonic switching noise riding on their outputs, which mandate use of a hefty low-pass filter between the output stage and the speaker terminals. In the D-Premier, there is no LC filter on the class-D amplifier's output; instead, the analog amplifier provides a very wide-bandwidth correction signal that cancels the ultrasonic switching noise that would otherwise be present.
Read more at https://www.stereophile.com/content/devialet-d-premier-da-integrated-amplifier#itODRXCtmGpqomRY.99


http://www.dynamicaudio.jp/file/110125/201101DevialetWhitePaper.pdf




Here is the summary: 


Key points • This is not a class D! but a genuine analog amplifier + integrated DAC • Original design process from the blank slate & extensive innovation (5 patents) for such a different outcome • ADH Analog Digital Hybrid amplification technology: • a class A and a class D amplifiers connected in parallel to the speaker output • class A is the master of the output: it sets the sound transparency • class D is slaved to the class A to provide most of the current • class A is improved by the assistance of class D (yet unexpected): ADH is more accurate than class A • Very special innovative class A: ultra linear, low output impedance, low power consumption yet infinite available output current in class A • Very innovative & ultimate DAC with direct high voltage output (no gain stage), located in the core of the class A: shortest possible signal path ever • Smart, PFC switched-mode power supply using a planar transformer, much less noisy than traditional switched and so-called linear supplies • High quality analog inputs, including an advanced versatile phono input, matching the quality standard of the amplifier.

[AmpDesigner333]

Cherry is the only Class-D with reduced feedback, faster switching, and wide band high damping.

Sonically, there hasn’t been any significant competition.