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


This review first appeared in the April 2010 issue of hi-end hifi magazine fairaudio.de of Germany. You can also read this review of the NAD M2 in its original German version. We publish its English translation in a mutual syndication arrangement with the publishers. As is customary for our own reviews, the writer's signature at review's end shows an e-mail address should you have questions or wish to send feedback. All images contained in this review are the property of fairaudio.de or NAD. - Ed.

Reviewer: Jörg Dames
Source: Fonel Simplicité; Archos 7 with docking station
Amplification: Pre/power - Accuphase P-5000, Belles 21A, Fonel Renaissance, Myryad MXA 2150; integrated - Fonel Emotion, Abacus Ampino Rieder, Denon PMA 2010AE
Loudspeaker: Thiel CS 3.7, Sehring S 703 SE
Cables: Low-level - Straight Wire Virtuoso; high-level - HMS Fortissimo, Reson LSC 350, Straight Wire Rhapsody, Ortofon SPK 500; digital - Aqvox COAX-75
Review component retail: €5.990

Contrary to usual preference, I’ll segue into this review with some verbal definitions. Most of you already know that class D stands for neither class Digital nor means that this class of operation is digital per se. While the basic operation of this very energy-efficient approach does resemble digital by consisting of just two switch modes of the output devices—on and off like 1 and 0—the more common pulse width modulation technique is actually defined by the duration between switch changes. And in theory, this duration—very non-digital at that—can assume endless intermediate values to not be limited to given samples. "PWM is definitely analog" is how Nelson Pass once put it succinctly.

This hasn’t prevented many from equating class D with digital. What confuses matters further is that class D amps are usually fed analog signals even though a few exceptions are driven by digital signals. In the latter case, PCM signal of digital sources like CD or network players connects directly to the amplifier input. Analog sources then have to undergo A/D conversion the moment they enter the amp. This gets us to the NAD M2. It belongs to the still rarer species of digitally driven class D amplifier which process the signal in the digital domain. After filtering out the switching frequency components, the outputs of course pass analog signal. In that sense, the M2 is a ‘true’ digital amplifier.


For advantages, NAD cites "the elimination of analog distortion, noise and hum in the preamplification and power processing stages" and also describes the M2 as fundamentally an over-dimensioned D/A converter (at 200wpc into 8 ohms the M2 is stoutly potent even for an integrated amplifier) which has practical consequences.


The concept begs for a bevy of digital inputs and NAD offers five: 2 x coax, 2 x optical and 1 x AES/EBU all of which accept 16/44.1 to 24/192 signal. There are also coaxial and Toslink digital outputs. The M2 further accepts analog inputs on RCA or XLR which, naturally, encounter immediate A/D conversion via TI PCM4222 chips that are selectably upsampled to 48, 96 or 192kHz but always at a fixed 24-bit word length. This creates an additional format conversion layer and NAD admits that the digital-direct inputs have a sonic advantage over going in analog.


While digital signal is processed up to 24 bits and analog automatically converted to 24 bits, NAD actually talks of 35-bit math. At first glance this makes no sense. It's simply about mathematical headroom however. Envision a basic pocket calculator. If it accepts 24-digit figures, subsequent calculations can easily exceed those digits. A calculator with 35 digits will produce more accurate results.


This internal processor density also ties to NAD’s DDFA™ direct digital feedback amplifier scheme which the development team seems particularly proud of. It allows additional corrective math just ahead of the speaker outputs. In collaboration with Manchester-based semiconductor specialist Zetex, NAD has developed a circuit topology which "corrects even the smallest pulse deviation" by comparing the PCM input signal against an internally generated PWM reference signal. Possible errors include pulse amplitude (incurred by remaining power supply rippled), pulse width and a softening of the pulse edges – all factors with are directly relevant to the quality of the music signal.

A common challenge for class D amps is the treble not least because the output filter handling the switching frequency (a high 108MHz with the NAD) is influenced by a loudspeaker’s variable impedance to affect HF linearity. The M2 runs a special adaptive filter whose impedance matching can be altered under signal (selectable with the volume control in the relevant menu layer from 2 to 8 ohm in 1-ohm increments and one final >8 setting).


On the user interface, there are NAD’s obligatory soft clipping (here that's in the digital domain of course and adjustable on the rear to ‘sharp’) and absolute phase inversion but also sensitivity adjustments for the analog inputs in four levels between -9 to 0dB, with a fifth fixed setting. About this desirable feature, the owner’s manual casually advised to "choose it when the M2 connects to the outputs of a preamp to operate as power amp. Then its outputs are fixed and its volume control bypassed."

While playing with the 0-9dB range under signal, an unintentional click too far and all hell breaks loose. Shortly afterwards, the display shows shortcut (probably triggered by clipping), then there's sudden blessed silence. The - um, protection circuitry of NAD works reliably. This described an actual occurrence in my digs accompanied by very accelerated heart beat. And no, I hadn’t yet studied the manual. While the latter deserves praise, this unprotected feature with no secondary enable function to avoid such mishaps seems somewhat ill-considered.

Before getting into sonics, two more housekeeping items. The M2 draws a lot of current which theoretically could challenge your residential circuit breaker system. My 16A AC line created no bottle neck however. NAD then warns that "due to its special technologies and high-frequency shielding, the M2 must be plugged into a properly grounded socket, i.e. the ground connection must be properly earthed." For most home applications, this should be a given. The M2 has a snazzy display whose text could be a few picas larger however to remain visible from the couch without strain.


NAD recommends a minimum 1-hour warm-up which, during the M2’s multi-week stay, got circumvented by never powering it down at all. I measured ca. 80 watt idle consumption, not exactly cause for a green laurel wreath but compared to many non class D amps still relatively civilized. While on power and AC, the M2 reacts audibly to correct power polarity. Schuko users should experiment with our bidirectional EU plugs to be sure. Also the M2 was far from happy when sharing a duplex with an idling Fonel Emotion while playing itself. The sound lost luster and drive. Hence the M2 got its own power strip and all A/B amps were kept in standby in the adjacent room. Coddling and all that.
Enlarge!