NHT SuperZero loudspeaker & SW2 subwoofer Page 7

Aunt Corey's Uneducated White Trash DIY High-Pass Filter
But I knew there had to be a way to mate these two terrific products into one awe-inspiring $880 loudspeaker system. I tried running the SuperZeros full-range with the Muse while sending my preamp's output via a Radio Shack Y-adapter to the MA-1's line-level inputs—so the SW2 could fill out the bottom end, even though the SuperZeros weren't high-pass-filtered at all.

Now I was getting somewhere! This sounded much better than using the MA-1's crossover. The SuperZeros remained clean and clear, and the SW2—the MA-1's 80Hz setting gave the best blend in my room—really fleshed out the sound to make for a truly full-range speaker system with all the SuperZeros' amazing attributes, but coupled with a real bottom end.

Still, the transition between the SW2 and the SuperZero wasn't as smooth and transparent as I'd hoped for. And driving the SuperZeros full-range did nothing to improve their dynamic headroom and midrange purity—reasons to use a subwoofer which are just as important as merely adding the bottom octaves.

Like Martin Colloms, I can't help messing with stuff. Martin bi-amps his Wilson WATTs/Puppies by inserting a passive first-order RC filter between his preamp and the amp that drives the WATTs, so I thought I'd try the same thing with the NHT system. I made up a couple of passive first-order 6dB/octave line-level filters: simple RC networks consisting of a series 0.1µF "Kimber Kap" polypropylene capacitor shunted to ground with a 15k ohm Resista metal-film resistor (fig.1). In conjunction with the Muse's 51k input impedance, this gave a passive high-pass filter -3dB down at approximately 137Hz (footnote 7) (fig.2)—high enough in frequency to roll off enough bass in the SuperZeros to give them more headroom, but low enough to ensure a seamless transition with the SW2. I soldered these components directly to an unmounted RCA jack at the filter's input, and to an RCA plug at the filter's output. These filter modules were plugged directly into the Muse amplifier's input jacks, and the interconnects split from the Y-adapter at the preamp output plugged directly into the filters.

Fig.1 Simple DIY passive high-pass filter for use with satellite speakers.

Fig.2 Amplitude & phase response of fig.1 filter when loaded with 51k ohms amplifier input impedance (2dB/vertical div.). Note polarity inversion of satellite feed due to series capacitor (footnote 8).

This was the best-sounding configuration of all. With the high-pass filters providing a degree of bass attenuation in the SuperZeros, and the SW2's own crossover set for 80Hz, the transition between the sub and the SuperZeros was much more seamless than when driving the SuperZeros full-range. And the system remained clean and clear at higher levels, too—much higher levels, in fact, than either the Spica Angelus or the ProAc Response Two, both of which cost a good deal more than the combined price of the NHT system, and neither of which approaches the full NHT system's bass extension. I was surprised at how much this setup reminded me of the sound I enjoyed from my previous reference speaker system—the $3000/pair ProAc Response Twos mated with the $2750 Muse Model 18 subwoofer. In terms of finesse, musicality, and top-to-bottom coherence, the $880 NHT system came very close to the kind of sound I used to hear from that $5750 reference combo.

Depending on the room I used them in, the polarity of the subwoofer connection varied with the SuperZeros/SW2 system. In my He-Man listening room, I achieved the flattest measured response and smoothest blend between the sub and sats with the woofer connected in the same polarity as the SuperZeros. However, in my Real World listening room—my mild-mannered living room by day—I found that reversing the polarity of the connection between the MA-1 amplifier and SW-2 subwoofer gave the flattest measured response and the better integration with the SuperZeros. I determined the optimum subwoofer polarity in both rooms with the bass warble tones on track 16 of Stereophile's Test CD 2 and a $30 Radio Shack sound-level meter set for "C" weightin' and "slow" needlin'.

NHT's Ken Kantor agrees that there's really no "right" subwoofer polarity, and that the user should try both in order to find the polarity that makes for the smoothest transition and the least excitation of room modes. This was certainly true of my living room, where a positive subwoofer polarity interacted with the pretty large bass hump in the 80-100Hz region. Reversing the subwoofer's polarity nulled out this hump and gave a much flatter response through the whole bass range.

Bottom line: Try both polarities with various male vocal recordings (the Fairfield Four were ideal for this, as were Richard Lehnert's spoken intros on track 1 of Test CD 2) to see which sounds more natural. The difference won't be subtle, believe me.

I said before that if the $230/pair NHT SuperZeros had been around back when I bought my $550/pair Spica TC-50s, I'd've chosen the NHTs. Well, if the $880 NHT SuperZero/SW2P combo had been available when I bought my $1275/pair Spica Angeluses, I'd've gone with the NHT system. The Angelus is an amazing-sounding speaker, but it just couldn't compete with the SuperZero/SW2P system in terms of bass extension or dynamic ceiling. And while the Spica has one of the most holographic soundscapes at any price, I don't think the NHT system (hooked up the way I finally had it) lags too far behind. As with the Harbeth LS3/5As vs the subwooferless SuperZeros: If I had to choose between the Angelus and the SuperZero/SW2P combo as a reference speaker system for my reviews, I'd choose the NHT.

Footnote 7: To calculate the approximate -3dB point for a first-order filter like this, use the formula f = 1/(2PiRC), where Pi = 3.142, C is the value of the series capacitor in farads, and R is the combined resistance in ohms of the amplifier's input resistance and the shunt resistor used. For the example CG gives: a 15k resistor in parallel with a 51k resistor gives a combined resistance of 11.6k ohms; the capacitor has a value of 0.1 x 10E-6 farads; the formula predicts a -3dB point of 137.1Hz.—John Atkinson

Footnote 8: In John Atkinson's comments on the NHT SuperZero, he noted an apparent phase inversion imparted by a simple high-pass RC filter (fig.2, p.149). In fact, the capacitor cannot invert the phase—it causes a phase shift which only approaches 90 degrees at the limit and only shifts by 45 degrees at the nominal crossover point, 137.1Hz. On the graph, a 180 degrees phase inversion is shown, due solely to measurement presentation. Allowing for this, the graph does show the predicted 45 degrees lead at crossover. The corresponding low-pass network for the woofer could well lag by 45-90 degrees, according to the order of the crossover. The resulting system phase match is indeterminate, neither in nor out, and practical experiment with either phase connection is worthwhile to see which suits you and the listening room best.—Martin Colloms

537 Stone Road, Suite E
Benicia, CA 94510
(800) NHT-9993