This review page is supported in part by the sponsors whose ad banners are displayed below

Arguments in favor of Ncore [excerpted from Bruno's Is class D finally ready to compete with class A talk presented at HighEnd 2012 in Munich]. "Manufacturers of class D amplifiers often oversell their wares. And serious listeners have often said 'but you're not yet there'. This has become something of a default position. When people ask what's so special about class D, I always have to disappoint them. There's only one area in which class D is fundamentally better than other amplifiers - power efficiency. When a class D amplifier sounds good to boot, it's in spite of operating in class D, not because. So why bother?


"The industry needs them and it's something I'm said to be good at. But what you get in return for power efficiency is a raft of serious technological issues that need to be addressed. Contrary to what people think, bandwidth isn't one of them. If 192kHz sampling is good enough for digital, the 400kHz switching frequency of class D amps certainly won't get in the way. It only complicates the design of control circuits which usually like a lot of bandwidth to work in. Let's get to the first problem, distortion. Early class D amplifiers had distortion in the 1% range. That high the precise distortion spectrum makes all the difference between nice sound and unacceptable. The second problem is possibly even more severe - output impedance. Today many class D amps have an output impedance that reaches into the 10s of ohms in the audio band. The response of such an amp will be all over the place depending on what speaker you have. So people have good reason to be suspicious when someone claims that the definitive class D amplifier has finally arrived.
"But consider this. Reviews are unanimous. No two class A amplifiers regardless of price sound the same. That's the same as saying they all change the signal as it comes through. Technically speaking the biggest problem for linear amplifiers is distortion at high frequencies. Intermodulation tests at Stereophile show best-case scenarios of 80dB down when the best D/A converters go to 110dB or more. Most linear amplifiers have an inductive output network to keep them stable, hence the frequency extension of most class A amplifiers is load sensitive. That's not to say these problems are inherent in class A amplifiers, only that the ones today still exhibit them. The areas where class D amplifiers encounter their greatest challenge are the same areas where class A amplifier still have some work to do.
FFT of NC400, 1000Hz test tone, 2nd harmonic down 105dB, 3rd and higher even lower- taken from here
"Reviewers often reference amplifiers like the Mark Levinson N°.33 from the 90s to suggest that the state of the art in linear amplifiers hasn't really moved forward since the last decade and a half. The biggest advantage for class D amplifiers is that linear amplifiers have stopped improving in any substantial way. In the meantime class D amplifiers improve noticeably year after year.

"Ncore was designed under some reasonable presumptions, namely that you want good HF extension beyond the audio band and very low distortion inside the audio band; and that load sensitivity should be minimal. Distortion for Ncore remains below the noise floor until about 50 watts. It's a whopping 110dB below the signal. There are very few D/A converters that can match this. Until the onset of clipping it remains better than 100dB down. The ultimate test for amplifier performance is intermodulation distortion between two large sine waves at the very end of the audible band. I've not seen any published measurements anywhere approaching what Ncore delivers - class A or otherwise. Frequency response is wide enough to enjoy the benefits of high-resolution audio but not unnecessarily wide. Output impedance is completely off the scale. Those used to thinking in terms of damping factor, this is 11.000 over most the audio range. A switching amplifier has finally broken the class barrier. But why has it taken this long when class D was first described in 1938?
"One reason I mentioned already. Most earlier class D guys weren't critical enough of their own stuff. But class D isn't just another technology. It's a huge order of magnitude harder. Those people who poo-poo class D for its limited bandwidth have a point. The filter gets in the way. They're right. RF interference too is perfectly valid. Designing a class D amplifier, you start with an enormous handicap because it's trying to do everything but amplify audio. Where people get it utterly wrong is believing that these problems can't be solved. It simply takes much more dedication, theoretical study and attention to detail than it takes to making a good-sounding class A amplifier.
Mola-Mola display at its HighEnd 2012 launch

"With an amplifier like Ncore, getting the placement of a part or trace wrong by one millimeter is the difference between top performance and not getting the thing to work at all. If you want good performance, you can't use industry-standard Mosfet driver chips. You have to roll your own discrete drivers. The same for comparators which are either too slow or too noisy. Integrated op-amps are a favorite topic amongst audiophiles but they distort more than this entire class D amplifier. The truth is, a background in class A design is unhelpful for class D. The claim to fame of having an amplifier designed entirely in-house thus becomes a limitation for most manufacturers who attempt class D amplifiers. The only rational choice for them is to stay with class A/AB or partner with people who breathe and live class D. I enjoy the obvious luxury of being able to sell into the high-end market a class D amplifier and claim it is entirely my own design. That was an opportunity I couldn't pass up. And it throws Brussels-proof power efficiency in for free."


On NC400 vs. NC1200. "The NC400 has somewhat lower distortion & output impedance than the NC1200 and those deluxe discrete input buffers. Otherwise the NC1200 has about 40% more voltage and current reserves. That's pretty much it."


On DIY vs. commercial. "Let me recap the overall plan for Ncore. UcD is already rather good and certainly adequate for practically any application. It would make no sense to rush out and flood the market with Ncore. So we're reserving Ncore for high-end companies prepared to give it the profile it deserves. The NC1200 is the platform for that so isn't sold freely. I'm saying platform because we're offering customization. One of class D's problems in the high-end market has been the perception that it's just the same module in different boxes. We'll make sure one company's NC1200 won't be the same as another's (although of course measured performance will be the same). For the DIY market there is the über-tweaked version called NC400."
Enlarge!