Class D Technology

To the O.P.... Whether or not class D amps are inherently better or worse than any other topology is a false problem.

Rather, with class D amps, like with any other topologies, you are bound to find amps that are of your liking, and others that you will hate with passionate abandon.

The key is... Set your budget, and then... Use your own ears to evaluate and determine what you like!

Ignore the pronouncements of any manufacturer who is committed to any particular topology -- Hello Atma, how doing?

Ignore those who attempt to denigrate a class of operation as a whole, be it D, A/B, A, or tubed, particularly when they are basing their pronouncements on measurements, graphs, and specs.

Like any other technology, class D amps range in price... From less than $1K, to at least $60K... Some amps -- -- might sound a little non-denominational, and might sound even sterile, without much sublety. Other ones, even modestly priced, are amazingly subtle and musical.

One thing to bear in mind, is that class D amps are bears to break in... The wonderful old Bel Canto Ref 1000 Mk.2 took a relatively modest 600 hours to deliver its best, but my NCore-based Rowland M925 monos took about 1500 hours to achieve thier top magic.

I adore my class D Rowland M925 monos without reservations for everything they do... In my opinion, their resolution and seductive grainless musicality is in the same league of the big Solution class A/B amps and the ARC 800 tubed series.

I have no interest in "upgrading" to anything else at this time, as what I heard this far would be, at most, a lateral move... On the other hand, at some future, I might indeed "upgrade". But whether this process will take me to a class D device or a class A/B device remains an open ended question.... I will evaluate, and will decide which gives me "magic" close to my heart, and which manufacturer offers me confidence.

But I admit... in spite of class A and big tubes being as capable to deliver sound that I would adore, There will probably be no pure class A amps in my life... I have no symphaty for inefficient heat generators... Nor will there be big tubed amps, because I have no intereest in tendeng and swapping sagging or burnt-out tubes.

   Saluti, Guido

Hi seanheis1, amplifier designers modifying class D modules is largely an urban legend.... One of the few amp designers who has done this is Bruno Putzeys, who has made some minor modifications to his own Hypex NCore NC1200 for integration into the Mola Mola Kaluga amplifier that he has also designed.

Rather, Designers who utilize class D amplifier use an incremental design approach to their amp implementations.

For some entry level implementations, it is often possible to limit oneself to housing the power conversion module in a chassis and wiring it to output terminals.... If the module is of very high quality, such as the NC1200, or its NC600 younger brother, the results can still be remarkably good.

Some other manufacturers, such as Merrill Audio, do a thorough "passive" component design around the modules... They implement highly dampened chassis, highest quality internal wiring (e.g. Cardas), and I/O connectors: see for example the Merrill Audio Veritas monos that I have reviewed for PFO a couple years ago. In some cases, like in the entry level Merril Teranis stereo amp, they design custom input stages.

Finally, some manufacturers, like Rowland, utilize the Hypex NCore NC1200 models as component parts of sophisticated amps, where all but the power conversion module is a custom-designed component.... In the M925 monoblock, the power supply is a 2500W DC multi-regulated SMPS unit fed by a power factor corrected rectifier (PFC) where the generated DC is further "whashed" through Jensen 4-pole capacitors to eliminate any residual ripples. The inputs are coupled to very large Lundahl transformers to maximize common mode noise wrejection.... There is a lot more to the technology of the Rowland M925 amps in my own system, but I do not pretend to know all technical details.

Suffice to say that the resulting sound and music are too die for.... The amps are as quiet as can be, and the output is harmonically articulate, with no trace of grain, and certainly without any treble intermodulation that I would otherwise easily detect as harshness in split high string and high brass parts.

Granted, in general NCore amps, the higher end Pascal amps, and also the better ICEpower implementations are not likely to fit the requirements of the lover of triode designs who prefers a slight bloom in the midrange, a warm mid bass, and a gently tapered off treble.

Rather, these amps, once they are well broken in, tend to yield what I like to call a "goldielockian" musicality. In other words, a sonic environment that yields an even harmonic treatement to the entire audible spectrum.... I love it, but admittedly not everyone does.

Regards, Guido

Thank you grannyring, yours is an excellent case in point... You and I happened to have adopted class D amps because -- using our own ears *Grins!* -- we fell in love with their sound... And are still in love with what our devices continue to do for us.

Saluti, Guido

Hi Charles, I agree with you, in the following way:

From a sonic point of view, there is really no absolute advantage in principle between one topology and another one.

Rather, there are users that might find a particular amplifier of one topology preferable to another amplifier.... Be the second amplifier of the same topology, or of another one.

Not all class D amplifiers are sonically desirable to my ears... There is a broad range.... From devices that make me cringe, to those that make me swoon.

The same is for amps that are built around other topologies... Thus my love for Solution, ARC amps, and some VTLs.

While there are practical considerations why I am unlikely ever to own VTL or ARC amps to power my Die Muzik speakers (heat dissipation and tube maintenance), the reason why I do not own a Solution amp is pure happenstance.

I also heard a few triod amps that I truly enjoyed, although they are unlikely to be able to drive my speakers to my satisfaction.

So, if I were to use the language of formal logic, I would say that.... For all amps of a given class I have found there exists at least one amp that I have enjoyed.

Finally, as we are all different, it is self evident that what makes me swoon, is likely to leave some other suffering audiophrene totally non-plussed.

As for nixing a class on a couple specs.... Now, that is beyond me to fathom.

Saluti, G.

Fascinating... Looks like my last post has dematerialized. I was not aware that what I write is so controversial. Oh well...

To Charles, Mola Mola is a very fine amp, but it might not constitute the "last word" in the application of NCore technology to an amp. In the current Rowland lineup, for example, I have found some amps that I prefer over Mola Mola, and at least one that I like less. E.g. My M925 monos and the M825 stereo are in the top range of my preferred amps, together with Solution and ARC reference amplifiers. Yes, I do prefer them to Mola Mola.

When I listened to Mola Mola, the amp was not fully broken in, so it exhibited some excesses in transient response, and some transient behavior that might have been traces of intermodulation.... Difficult to say if the issues are inherent to the device, or they were a product of incomplete break-in.

Hence, I currently prefer the Pascal M-Pro2-based Rowland Continuum S2 over Mola Mola.... An integrated that I find infinitely graceful.

On the other hand, the Rowland M525 mono based on Pascal S-Pro2 might not be as subtle as Mola Mola.

Note that I have not listened yet to the Rowland M125 bridgeable amp, nor to the Daemon super-integrated... I have heard wonderful things about them, but have no direct experience.

Saluti,

seanheis1, while most newer class D amps are incredibly quiet, -- as in having no background noise -- foreshortened decay or "ead and opaque leadden silence around the note is not the sign of any good amp, unless a track was created synthetically and not from a live recording, in which case, no amp can ;be imputed with wrong doing.

Rather, a good amp, and so much so a good class D amp, is expected to give you what is sometimes called the "sound of the living Silence", that is the natural decay of notes, and the unavoidable ambient cues characteristic of all venues, except for anecoic chambers.

Saluti, G.

Adding some more info about the Bel Canto REF1000 Mk.2. John Stronczer of Bel Canto added a custom input stage which elevated the input impedance from 8K Ohms to 100K Ohms, to make the amp compatible with those tubed pres that exhibit high output impedance.... As a side benefit, the amp also became much sweeter, immune from audible intermodulation in the treble, and... Quieter to boot.

Regards,    to   

Mapman & Al... I'm afraid the your effort to open some solidly shut minds -- if just a smidjin -- might be on the neutral side of... Futile. To quote our unforgettable Rodman9999:

"Just don't fall into the category of those whose minds, like concrete, are thoroughly mixed up and permanently set"

As for me, I finally moved into my new house two days ago.... The system is in place, if not yet connected... As soon as I find my wires and PCs, it will be making amazing music again... Through my class D Rowland M925 "mid-fi" mono amps *Grins!*

G.

You are right Mapman, if I ever change amps, the replacement of my monos is going to be an integrated+DAC.... Probably the Rowland Daemon, which is also running in class D.

Congratulations geekboy!

Be patient with your new BC Ref600S before evaluating them seriously.... Unless they were completely broken in before you got them, it will take anywhere from a few hundred hours to about 1000 hours for them to reach their full potential.... Try leaving them powered up 24/7 and let them crunch a signal at medium lowish volume... E.g. the interstation hash from an FM radio tuner.... That should clear their musical throat quite nicely. Full break-in on class D amps is essential... The difference is quite staggering.

Once they have stabilized, the careful selection of PCs will bring them to even newer hights. For PCs that are not super-expensive, I would suggest Shunyata Alpha

Regards, Guido

Hi Mapman, when I had BC REF1000, I had best results in ascending order with Furutech Evo II, Shunyata CX series, and Shunyata Z-tron... By Extrapolation, I conjecture that the Shunyata Alpha and Sigma series might yield even more desirable results.

Saluti, Guido

The concerns of the OP and of George are correct, at least from a historical point of view.... When I first started looking at class D amps, in the middle of the last decade, there were many examples of class D amps which were either sounding dark, or limited in the harmonics of the upper frequencies, or harsh, or high-fiish, or a combination of several of the above undesirable characteristics.... For Example:

* Flying Mole -- All the delicacy of music reproduced with 150grit sand paper!

NuForce Reference 9 SE V2 monoblocks -- Clean but quite climical, even after 1K hours break-in... And treble was clinky and limited.

Red Dragon monos -- Both harsh and dark at the same time... Quite an achievement!

Wyred4Sound monos -- Listened extensively at RMAF several years in a row... Harsh and fatiguing to the point of non-listenability.

Rowland M501 monos Powerful, but lacking subtlety.

Bel Canto Ref1000 Mik.1 -- Better than the above, but still somewhat rough in the treble... Note that the problem simply does not exist on the Mk.2 version.

Rowland M302 stereo -- Sweet and delicate, but will little macrodynamics... And just slightly cold.

But... The World has moved along... A dozen years have gone by, and modules and device designs have evolved enormously in the land of class D amplification, starting with the Rowland M312 stereo, and the Bel Canto Ref1000 Mk.2 monos: products that simply... Make real music.

Today, amps like the Bel Canto REF600 monos and the Merrill Teranis stereo based on Hypex NCore NC500 modules, once appropriately broken in -- yes I insist it is needed -- make extremely fine music for their quasi entry level pricepoints, without pretending  to achieve the ultimate subtlety of reference level amps.

But if one were keen in hearing what musical refinement has been achieved by current reference-level class D amplification without audible compromises, please have a listen at least to some of the amps that have captured my heart over the last few years:

* Merrill Veritas monos based on NCore NC1200.

* Rowland Continuum S2 integrated -- Based on Pascal M-Pro2.

* Rowland M825 Stereo -- based on NC1200.

* Rowland M925 monos, based on NC1200.

* Bel Canto Black -- based on customized NCore modules.

And there are more marvellous class D devices, which I have not had the fortune to audition at all price points.

Are all current class D amps today superb music makers? That is unlikely.... For example, last time I heard W4S -- admittedly this was 3 years ago -- the sound of my music camples: piano, string sextet, Diana Krall, orchestral, and vocal + sax was still so harsh to give me an "ear bleed".... But by now, W4S may very well have matured as well.

Please note that exactly like proclaiming that all class D amplifiers are inherently flawed constitutes a logical fallacy, so would be the assertion of the opposite... Reality is, that with all topologies, some amps will meet our particular taste in music reproduction, some will miss by a country mile, and some will be somewhere in between.

G.

 --

Thank you Tim and nyame.... I try to report what class D amps I have experienced in the last decade in as unadorned a way as I can.

One more class D amp worth looking at below $10K is the Theta Prometheus... I believe it is an NCore NC500 implementation which uses custom toroidal transformers instead of the standard NCore SMPs or a custom SMPS...

Perhaps you Statman can tell us more about your unit.... How long has break-in taken to complete... What amp is it replacing.... And what audible characteristics are most salient on Prometheus?

Regards, Guido

Hi seanheis1, as far as I know, preamplifiers cannot be implemented in class D... I have been told that their topology is the equivalent of class A. However, Preamps with regulated SMPS do exist... ANd the ones I have listened to are rather wonderful: e.g. Rowland Corus and Criterion, Rowland Capri and Capri S2, and the preamplifier section in the Continuum S2 integrated.

G.

WOops... You are correct George: Theta Prometheus is indeed based on full NCore1200, not on the lower end NC500 module... My bad!!!!

G. 

Hi all, in olden days, with the NuForce and M312, RF leaks were of some concern... My old FM tuner did pick up hash from the amplifiers if when held less than 2 feet from the power supply of the amps. 

However, newer amps like the Merrill Veritas and my current Rowland M925 do not radiate at all.... AM and FM bands are not affected at all, not even if the tuner is placed on top and is in direct contact with the amps.

Without making a broad generalization, I suspect that FM leakages might today be less common than in the past. Perhaps more of historical interest than anything else.

Worth checking out though, particularly when testing older class D amps.

Regards, G.

 on the

Hi keeponrockin, it is fabulous that you had an exceptional experience with the Leviathan.

Admittedly, I audition Leviathan several times at length, using my standard test tracks of piano, chamber, jazz/Krall, orchestra, and sax/vocalists, but only at RMAF.... What often happens at shows is that manufacturers bring along brand new devices fresh off the factory floor, with little or no break-in.... So, it is possible that I was experiencing the limitations of units which were not properly broken in.

Regards, Guido

Hi RSA, you are safe with Bel Canto... Like Merrill Audio, Rowland, Theta, and D-Sonic, they are reported to make make wonderful amps. Several years ago I review the now withdrawn Bel Canto REF1000 Mk.2 monos for PFO.... Here is the article:

http://www.positive-feedback.com/Issue43/bel_canto_ref1000.htm 

Hi Richard, actually what the performance comparison between CS2 and the M625S2+Corus combo highlights is that the latter com bo was designed to a price point Three times higher than Continuum S2, and that it contains circuit sophistications not possible to meet CS2’s pricepoint.

Were you to compare the performance of the class A/B M625 S2 to the class D M925 monos flagship, which costsover three times as much as M625 S2, you might be tempted to draw an equally unwarranted conclusion that class D is inherently superior to class A/B... Which of course would be a fallacy as well.

Hi George, I am amongst those who cannot hear any displeasing high freequency artifacts or limitations in some -- and by no means not all -- class D amps in current production. I would include at least:

Rowland M925  and 825 (both based on NC1200), Continuum S2 (Pascal M2-Pro).

Bel Canto Black (NC1200 derivative).

Merrill Veritas (NC1200).

Conversely, a number of older amps, like the old Rowland M501 and 201, and the NuForce did  display modrate to severe audible limitations.

It is also worth mentioning that even the best class D amps can sound limited, oscillating  between screetchy and tubby until break-in is complete, which unfortunately can range between 600 and 2000 hours, depending on power conversion module and amplifier design... THus, breaking in a class D amp can be an exercise in extreme patience and anxiety control *Rolls eyes!* 

Yet, As you are mentioning the switching frequency as a probable cause of limitations that you perceive, would is your estimate of the switching frequency one would require to push any such artifacts into the inaudible stratosphere for good?

Saluti, Guido

If it weren't for efficiency & power, would there be any reason to own a Class D amp? Do they beat Class A in any other categories that count for sound
quality?"

"

Let me try to respond to the question posed by the OP in the crispiest way I can:

1. I am deeply dissatisfied with at least 90% of amps on the market that I had the fortune, or the misfortune, to audition, regardless of class of operation.

2. I have selected my class D amps because I fell in love with their total audible performance, for every parameter.

3. Their power efficiency is important to me, but I would have overlooked it, if I had found an amp that was clearly audibly preferable for my allocated budget and musical taste.

4. I found less than a handful of amps that enthrolled me as much, or almost as much... The leading contender was made by Solution... Enchanting class A/B amp, which I would put more or less at par with the classs D Rowland M925

As for which class is inherently audibly superior .... It is an absurd question, which has no meaningful answer.

On the other hand, I grant you that when class D misses the musicality mark, either because the amp is not properly broken in, or because of inherent design flaws/limitations more common on older designs, it has the ability of outschreetching and outtubbing most everything on the market... The most unendurable listening experience for me was in fact that of a lower cost class D amp... Closely followed by that of an uber-stratospheric SS class A/B system... Conversely, even the worse uber-expensive tubed amp I heard, was only marginally worse than totally uninspiring.

G.

.

Hi Mitch2, my PFO article on Veritas details all the technical information that I have gathered on Veritas, based on conversation with Merrill, and by reading the publicly available technical literature on the amp. See:

http://positive-feedback.com/Issue68/merrill_audio.htm 

Unfortunately, I have not heard the ATSAH, so I really cannot comment on internal differences, nor sonic differences.

In itself, Merrill Veritas is exceptional value for its $12K list price. In a more traditional SS design, At the time of my review a couple years ago, I would have expected such sound to be more in the capabilities of an excellent amp in the lower $20K range.

Although there are obvious commonalities between the behavior of M925 and Veritas, there are quantitative and qualitative audible differences, which are however less than the price difference would suggest.... Famous laws of diminishing returns that audiophiles are acquiented to *grins!*

Thus, were I to ignore any budget considerations, my preference would comfortably go to M925 for the following parameters: 

* Power reserve

* authority.

* Macro and micro dynamics.

* Harmonic exposure.

* Linearity.

* Frequency extension.

Transparency.

* Sweetness and complexity of treble.

* stage size and concreteness.

* Imaging concreteness.

* Aliveness of "black" background... Also called the sound of the silence between the notes... Or the reproduction of ultra-low level information.

* Immunity to electronic background noise.

*  Immunity to intermodulation artifacts in the treble region.

Yet again.... let us Keep things into perspective.... the rowland M925 was engineered with a sophistication commensurate with its pricepoint just shy of $60K, several times the price of Veritas.

Saluti, Guido

Given the power rating, the module of the D-Sonic 3/3000 is probably the Pascal X-pro2. There is also a mono version of this amp, called 3/1500M, probably based on Pascal X-Pro1. See:

http://www.d-sonic.net/products/mono/

Cost of D-Sonic 3/1500M is $1300 per chassis.... And $2600 for a pair of 1500W Pascal amps is a pretty good deal.

Per my old notes, Dennis at D-Soic can be reached by phone at:

800-862-7998

Guido

Hi Tim, I will verify with Dennis if the M3/1500 monos are based on Pascal X-pro1 or X-Pro2.

About Rowland,

Continuum S2 integrated contains Pascal M-Pro2.

The flagships M825 stereo and M925 monos contain NCore NC1200.

Daemon integrated is the only Rowland amp containing Pascal X-Pro2.

Saluti, Guido

Hi UberWaltz, Bel Canto Ref500 is not one of my favorite amps... Even when it was current technology, it lacked the sweetness and transparency of its more powerful REF1000 Mk.2 brother.

On the other hand, it is worth mentioning that ICEpower-based amps have traditionally been bears to break-in.... And for at least a couple hundred hours out of the box they tend to sound dry and uninvolving.

In your $5K price range, I suggest you look into the current Bel Canto REF600 monoblocks, which are based on the excellent NCore NC1200 modules.... All NC1200 implementatios I have heard this far have been sweet and musical.... And you should hear some decent music out of them within about 200 hours.... But complete break-in hovers around the 1000 hours boundary.

One more mono amp...  Below $3K is the D-Sonic M3/1500. It is based on the excellent Pascal X-Pro2 module:

http://www.d-sonic.net/products/mono/ 

Regards, Guido

Hi seanheis1, I am not sure what circuit components are most prone to evolving during break-in.... I suspect that it is just about everything, from resistors, caps, boards, conducting traces, grounding surfaces, connectors, dyelectrics.... According to Jeff ROwland, capacitors and transformers are particularly slow at breaking in.... And am not sure what is the electro-magnetic phenomenon at play either.   I only know that I invariably experience significant audible evolution during break-in.... And it is not a particularly linear phenomenon either..... there are lots of dips and peaks during this period.... Gradually tapering off.

ICEpower modules seem particularly unattractive out of the box.... a more benign break-in curve can be experienced with NCore NC1200 and Pascal M-Pro2.... But do not expect miracles from any brand new class-D amp.... My experience is that patience does remain Golden *Grins!*

I cannot hear any trace of harmonic distortions on the 3rd or any other partials when I play music with my class D Rowland M925 monos or class D Rowland Daemon integrated.... Nor do they intermodulate. They are as sweet and stunning as can be.

As fashionable as it might be, generalization is a dangerous tool. Best to avoid the infimous logical trap of "1, 2, 3.... 99, 100".

Like with any other class of operation, the performance of class D amps can be evaluated only one amp at a time... And only through actual listening to boot. There is a huge variation of audible performance characteristics and tonal qualities amongst amps to make any sweeping proclamations one way or another.

 Regards, Guido

I should however like to point out that as with any other class of operation and topologies, class D amps range from the marvellous to the horrid. When I visited RMAF a few years ago, there was an equal proportion of class D amps and tube, class A, and A/B that left me befuddled for their total lack of musicality, be their tone dull and wooden, or acrid and shreaky.

Bottom line... Let’s use our ears when choosing an amp best for our music genres, and avoid preconceived abstractions. Old and trite urban legends are bound to drive us down some suboptimal path.

G.