Expanding the Class D Conversation: How Would You Characterize Their Differences?

The way it was explained to me by Alberto at AGD is the speed of GanFETs.  Here is an explanation as to why speed matters.  

Gallium Nitride (GaN) transistors switch much faster and with fewer parasitic effects than traditional silicon MOSFETs. In a Class D amplifier, the output devices rapidly switch on and off hundreds of thousands of times per second. With silicon devices, designers must insert a small delay called dead time between switching the high-side and low-side transistors to prevent them from conducting at the same time. If that dead time is not perfectly controlled, it can introduce small nonlinearities around the zero crossing of the waveform, which may contribute to subtle distortion or “grain.” Because GaN devices switch more cleanly and have virtually no reverse-recovery charge, designers can reduce dead time and achieve more accurate switching transitions.

Additionally, Class D amplifiers use an output LC filter to remove the high-frequency switching carrier and leave only the audio signal. Faster, cleaner switching allows for higher switching frequencies and smaller filter components, which can reduce phase shift and interaction with complex speaker loads. GaN’s lower capacitance and cleaner commutation behavior also reduce ringing and high-frequency noise, which can make the amplifier’s behavior closer to the theoretical ideal.

That said, while GaN can enable cleaner switching and potentially lower certain distortion artifacts, the overall sound quality of an amplifier depends far more on the complete design—its modulation scheme, feedback topology, power supply implementation, and output filter engineering—than on the semiconductor material alone. GaN is a powerful tool, but not a guarantee of superior sound by itself.

Basically, there are advantages to GaN but a bad design is a bad design and GanFETs are not going to magically make it sound good.  

Alternativey, Bruno Putzey mastered the control of switching timing and feedback behavior so well that silicon device limitations became largely irrelevant within the audio band. If that timing is not under control, you get the cold, unpleasant sound of older Class D modules.  

What is impressive to me is that people have found multiple solutions to the same problem and all are excellent and sound amazing.  

So, Nord is a value-proposition, IF they sound good. 

Here’s some of what they say  on the Nord website:

"We source the very best Modules and components from around the world. Modules come from the market leaders Hypex, Purifi and ICE Power. But we use only the very best of their products. We then integrate them with custom Nord boards and electronics to deliver some of the very best sounding Amplifiers available at any price. 

Operating with ultra low noise floor and exquisite bass control, revealing palpable breaths, finger plucks and decaying notes delivered by the mids and treble. Levels and layers of detail, dynamics and sound stage are breathtaking, with the ability and current reserves (28A) to drive the most demanding speakers down to 2 ohms. All delivered with no hardness or harshness."

Do Nord amps sound good? Well, from what I gather, your random amp of this type will have (borrowing Nord’s language):

• ultra low noise floor
• exquisite bass control
• dynamics
• sound stage

But:

• Levels and layers of detail -- not sure if this claim will be true
• no hardness or harshness -- not sure if this claim will be true

And those last two are deal-makers or breakers. That plus tonality, which also needs to be checked.

The Nord Web site and many posters return again and again to the easy-to-quantify factor -- the "module" involved (Purifi, Hypex, Ice, GaN). The Nord company has all three -- but how are they implemented? Which to choose? I cannot tell.

But the key point was alluded to by @atmasphere (when he said how different all Class D amps sound from one another) and articulated well by @verdantaudio:

That said, while GaN can enable cleaner switching and potentially lower certain distortion artifacts, the overall sound quality of an amplifier depends far more on the complete design—its modulation scheme, feedback topology, power supply implementation, and output filter engineering—than on the semiconductor material alone. GaN is a powerful tool, but not a guarantee of superior sound by itself.

The very nicely built DIY Hypex (recent module) I tried sounded harsh with my speakers, in my room. The M23 first gen Purifi Eigentakt sounds better. An earlier version of a GaNfet amp I heard sounded bright, forward and harsh. But the AGD Audion GaN sounds quite amazing so far.

Regarding my AGD vs. NAD comparison, I have one more listening session to do to feel confident about the differences I am pretty sure I am hearing. Report to come.

Random question: does anyone know if the NAD C298 upgraded its Eigentakt to the version 2? Or does that just apply to the other NAD amplifiers, such as the M23 and its brethren?

I never thought I would learn THIS MUCH about Class D stuff. Pretty crazy and awesome!

To answer my own question, I'm pretty confident the NAD c298 does not (yet) have the Eigentakt 2. If there is an Eigentakt 3, I bet it will get the 2.