Class D - A Fork in the Road

The OP poses an interesting question and it appears most of us haven’t been able to do a direct comparison. I’ll add my my experience, but I’m not sure it’s worth more than $.02. Please bear with me for a bit of 10 year historical context before I share my experience.

In the past, I’ve owned in one system or another:

Parasound A23 - (class A/B) - connected to the P3 preamp

PS Audio Stellar S300 - (ICE module) - connected to their Stellar DAC/preamp)

NAD M22 - (Hypex module) - connected to the M12 preamp/DAC

NAD M33 (Eigentakt module) 

T+A R2000 integrated

T+A PA 3100 HV integrated

currently own Atma-Sphere Class D monos in one of my systems, and Simaudio North 761 in the other.

Of course, there were other variables involved - preamp sections, DACs. But I did have a feel for each. Each purchase was an improvement over what it replaced. The Parasound was flabby/tubby in the bass compared to the others. The PS Audio was pleasant but lacked a bit of what the reviewers call “tone color”. The M22 was a bit denser in that regard, but still not there. 

The M33 (eigentakt and relevant to this conversation) - was clearly better than the hypex module but was that the preamp? The DAC? I’m unsure. And yet, i still didn’t like it - finding it lacked emotional impact. The speakers I had at the time (Sonus Faber Olympica Nova III) sounded nice, but lacked the drive and resolution I knew they were capable of. Too polite.

I reference the T+A gear because it fully satisfied in a different system but my needs there changed (different DAC, etc.). They were class A/B and offered warmth with resolution and impact. World class? No. Very good? Yes.

@westcoastaudiophile - your comment about the significant advantage of class D being low weight and high horsepower is only partially true, but also misleading. Yes,  those can be advantages, but class D, implemented well, has no sonic disadvantages against other topologies whatsoever. Indeed, it can be superior.

The Atma-Sphere Class D monos are as good as anything I’ve auditioned as a standalone amp. I now have them fronted by a very good preamp and DAC and that combo drives the SF speakers spectacularly. Yes - I changed a lot of gear in there, but the amps showed no limitations in that setup. 

I’m in the process of setting my 2nd system up in a new room (hopefully done w/in the year - we’ll see). I decided to upgrade the speakers when my SF dealer offered me a trade up I couldn’t refuse so I’ll be using the Class D’s to drive SF Il Cremoneses. Some would say a mismatch. I’m guessing not - but we’ll see. 

I’ve directly compared the Class Ds to the Sim North 761. They’re different sounding, but closer than you might expect - and I find the Sim North to be (to my ears) significantly better than previous Simaudio gear and vastly better than many of the $20k amps I considered.

My conclusions: implementation matters greatly - more than topology. Based on my limited sample set, GAN has more upside than Eigentakt, ICE and Hypex and probably has the potential to be as good as anything out there. 

Maybe that was all worth $.03.

Thanks for sharing your experiences. I would say that was $2.00 worth.

@mgrif104 +1 Correct, design implementation does matter, to push performance to theoretical limits! The same applies to any amp class design, AB/A/SET/D/H/G/..! Good amp design includes perfect topology, components, signal/power/thermal/mechanical integrity, component aging considerations, validation/tests, and final auditioning. You may try to find precise microphone with extended up to 100kHz range and check if no ultrasound is generated around amp., dogs typically don’t like it and walk away when “bad” D class amp is turned ON. 

I used 80kW total output power of D-class amps in live shows, with active PA signal processing (freq./phase/delay/hall-acoustic correction). Such PA system setup isn’t possible using AB class amps, class A is even worse. 

@westcoastaudiophile 

I have a dog with seemingly very acute hearing. He lays down in front of both of my systems (class A/B and class D amps).

May I respectfully suggest you reconsider your biases against the topology. I stand by my contention that class D “can” be competitive with anything else and we needn’t relegate it to pro-audio or for the initiated. 

I used to be the assistant engineer at a large performing arts venue in the 80s doing live shows and recording the symphony and smaller shows that came through. I learned a great deal though there was so much more for me to learn before I left. Regardless, I know well of what you speak but it really doesn’t apply here. No signal processing going on in the amps we’re discussing. No digitization or room correction going on either. 

@mgrif104 "suggest you reconsider your biases against the topology"

what “biases”??? I never said class D couldn’t perform, if it was designed well!