Round #1 (Formula, Lara, nCore 500, Codex). In our hardware/ear context, when I fit the socketed Sonic Imagery 994 not Sparkos Labs SS3602 discrete opamps to Nord's input buffer with discrete voltage regs, our class D monos emphasize subjective detail density but also set tonality to cool and cadence to somewhat clipped. Yet even with our essentially passive albeit actively buffered Wyred preamp, these qualities can work a treat into very meaty fiercely dynamic but hard-to-drive boxes like Mark & Daniel's faux-marble Maximus II monitors. Those rely on ultimate control to launch fully. They thrive on ultra-low output Ω and take class D's massive negative feedback in stride without drying out. Yet on our Audio Physic 4-ways, the same combo gets too lean and pedantic.


However; when I'd replaced our usual Aqua discrete R2R DAC with hq audio's HQ9038—2x12AU7 I/V stage, 2x6922 balanced outputs—the powerful valve signature of the Vietnamese/Danish DAC had prompted a course correction from too cool to straight down the middle. This expertly avoided any thick or fuzzy tube-dominant outcome. Given that happy precedent, I expected that back to our Aqua converter, injecting tubular action into the preamp stage instead should net a similarly happy nCore/Codex marriage of amp to 89dB speaker. And so it was. Despite hitting desired levels at 7:00 where clients with similar conditions would surely specify attenuators with a far shallower taper, I had neither channel imbalance nor sonic signs of a mismatch. This included noise; or rather, the blessed absence thereof. Without signal but ear on tweeter, the first onset of subliminal power-supply hum occurred at ~2:00 which even at full tilt remained very minor. Considering the wildly excessive gain of my first chain, this suggested excellent circuit design.

Unlike the late Terry Cain who happily tolerated how his favourite best-sounding tube amps all hummed into his hi-eff widebanders, I find noise unacceptable. Whether mechanically buzzing transformers or electrical power-supply surf, noise is the enemy of fine detail. As such it should have no place in costly high-end hifi. At the volume settings I actually used and well beyond them, Lara was dead quiet. With her one need make no excuses for listening to last-century tube tech.


This segues straight into the roguish accomplishment of my first round. Rogue Audio's Mark O'Brien has long since and very successfully packaged 12AX7 tubes with Hypex class D and linear power supplies. Consider his well-regarded Sphinx and Pharaoh [left] integrateds; or Medusa and Hydra stereo amps. For even more power and monaural class D, the same can be accomplished across three chassis. The requirement is simply a properly quiet valve preamp; and preferably one with true balanced circuitry since the majority of audiophile class D amps run balanced input buffers.


The functional advantage of such a combo is far higher power delivery than most all class A or AB amps can muster for the same coin. The sonic advantage is something closer to class A fullness than pure class D tends to. Finally, today's best class D offers 120dB or higher signal-to-noise ratios which might maximize whatever resolution advantages so-called hi-res files promise. The hybrid path of tubes+classD is thus an attractive solution to marrying retro with current tech. From my two recent experiments along those lines, I would highlight in particular excellent low-level listening clarity. Pretty much all competent hifi sounds impressive played loud enough as long as room and ears don't protest. At low levels meanwhile—the sort courtesy demands at 6:00 AM or 23:00 hours whilst neighbours or loved ones still or already sleep—most hifi systems are stuck in 'park'.