Sumo Samson subwoofer & Delilah crossover Page 2
The fact that both low-pass signals of either polarity are available makes it a cinch to use bridged amplifiers to drive the subwoofersone output is taken to one half of a stereo amplifier, its inverse to the other half, and the subwoofer connected across the two "hot" speaker terminals. (The amplifier in question must not have "floating" outputs.)
Getting the top cover off is an easy matter, involving undoing four Philips-head screws on the sides; getting it on again proved more difficult, for some reason known only to Murphy. Construction is to a high standard. (Sumo actually manufactures their electronic products in their Chatsworth factory.) Inside, Delilah's circuitry is laid out on one large pcb. Mains power is supplied by a toroidal transformer, with ±17V DC rails supplied by discrete regulators. Though the high-pass stage is discrete, the low-pass filter circuitry is based on LM833 dual op-amps, this a fairly recent audio-quality IC from National Semiconductor (footnote 5), with extensive use made of 1% metal-film resistorsthough not in the output stagesand Sumo's own-brand polypropylene-dielectric capacitors, these having stranded-copper lead-outs. The bass-level control is a good-quality component from Noble.
The filtered signals are buffered by discrete-transistor output amplifiers, these running in class-A and biased by LEDs. (A turned-on LED acts as a pretty constant voltage-drop reference, without the noise problems featured by Zener diodes.)
As mentioned earlier, one of the problems inherent with one or more subwoofers is integrating them into the system so that their response runs seamlessly into that of the main speakers. Delilah has been very well thought out in this respect, and the fact that it has a bypass switch, enabling the sound with subwoofer to be A/B'd against the main speakers driven full-range, coupled with the advice in the excellent handbook, enabled me quickly to zero in on set-up problem areas.
Subwoofers, in general, are large pieces of furniture, which restricts choice of positioning in an average-sized room. Sumo's advice is to keep it or them within three feet of the line between the main speaker systems, which will ensure optimum phasing. In my room, the ideal position was with the two subwoofers adjacent to the main speakers, on their inside and about a foot behind. This was domestically unacceptable, howeverit blocked the fireplace. I ended up with each Samson between two and three feet behind its respective speaker, along the line of sight from the listening chair, and quite near the rear wall.
Sumo recommends high crossover frequencies if the main speakers are two-ways or are limited in dynamic range, in order to minimize intermodulation distortion and thus maximize midrange clarity. Although both the speakers I wanted to use with Samson, Monitor Audio R952MDs and Celestion SL600s, are two-ways, neither has a dynamic-range problem and both give output to below 50Hz in my room. I would be free, therefore, to choose a lowish frequency, 50Hz or 63Hz, for both high- and low-pass sections, using Samson to cover just the one missing octave. Ordinarily, the user would make the crossover frequency the same for both sections, but in case of bass heaviness, Sumo recommends staggering of the crossover frequencies, ie, 63Hz low-pass and 80Hz high-pass. Delilah's low-pass slope can be set to either 12 or 18dB/octave. The former, of course, will give faster rejection of midrange frequencies from the subwoofer feed; 12dB/octave is preferred, as it makes possible in-phase operation at the crossover frequency of subwoofer and main system.
The phasing of the subwoofer will be dependent on a number of factors, including the positioning of the subwoofer relative to the main speakers, the crossover slope chosen, and whether the amplifiers used are both inverting, both non-inverting, or a mixture. This is where the bypass switch and the fact that Delilah has both inverting and non-inverting outputs come into their own. Start with an arbitrary choice of phasing to the subwoofers and the lowest crossover frequency. Play music with a repetitive, percussive bass content. Advance the bass-level control to the point where no difference can be heard in LF volume when the crossover is bypassed. Now reverse the polarity of the subwoofer feed and repeat the process. If, when the subwoofer is switched into circuit, there is now less bass audible, the original phasing was correct; if there is more, then the new phasing is correct.
Setting the level of the subwoofer is also critical. It is also somewhat program-dependent. It is only too easy to have the subwoofer easily audible; you then often end up with a bass "continuo" accompanying the music. The level should be set so that you are not aware of the subwoofer workingexcept when the program has true bass content!
Finally, an amplifier has to be used with the subwoofer (although a version of Samson with an integral 1kW switching amplifier designed by John Ulrick, once of Infinity, is also available for $1295). It is very tempting to use any old amplifier to drive a subwoofer. I found, however, that it does need to be of good quality if it is to provide an adequate degree of control. More on this topic later.
One Samson, driven from Delilah's summed mono output, was used first, then two Samsons were used with both pairs of main speakers. Needless to say, no other speakers were in the listening room the whole time.
Upon first being switched on, Delilah had rather a grainy, forward sound quality on its high-pass outputs. The graininess disappeared after an hour or so and the forward character diminished, but did not disappear altogether. This might be a relevant factor in a system already somewhat forward in balance.
When first connecting a subwoofer to a system, there is a great temptation to play records with "demonstration-quality" bass. I, of course, succumbed to this temptation and reached for the Telarc CD of Michael Murray playing Bach organ works from the First Congregational Church in Los Angeles (CD-80088). With one Samson driven by half of a Sumo Polaris, the Monitor Audios took on, there is no other word for it, majesty in their reproduction of the D-Minor Toccata and Fugue. This was awesome bass. I put on other organ recordings. Again, that majesty!
I reached for my acid-test rock recording, the 12"-single of Huey Lewis & The News' "The Power of Love" (Chrysalis 4V9 42889). Ah hah. Now I heard a discontinuity between the '952s and the Samson. The body of the bass-drum sound lagged a little behind the sound of the beater striking the skin. The obvious solution seemed to be to turn down the subwoofer level, but though this took down the overhang, it was still not fully integrated, the thunder still following the thud.
Looking at the spectrum of the sound at the listening position (with the bass-level control set correctly) using a Heathkit audio analyzer revealed the response in-room to extend to 25Hz, the 3dB point lying at 20Hztruly excellent low-frequency response. However, the 40Hz 1/3-octave band was about 6dB higher in level than the 50Hz and 63Hz bands, and the 31.5Hz band was also a little high. That this was a room problem was shown by the fact that this peak did not appear on the nearfield response, but I could not solve it by repositioning the single Samson.
Maybe driving the room acoustics in a different way, using stereo subwoofers, would tame the 40Hz problem. Two Samsons were positioned as outlined above and driven by a single stereo Polaris. Again the majesty was present on the organ recordings, but in addition, comparing the sound of the Monitor Audios full-range with that of the '952/subwoofer system revealed an unexpected effect. Switching in the Samsons, even on music which didn't have any low-bass content, both added depth to the soundstage and increased the delineation of individual images within that depth. If the David Willcocks/New Philharmonia recording of Fauré's Requiem with boy trebles (Seraphim SZ/4XG 60096) is my favorite, musically, then the more recent (1978) LP by the City of Birmingham SO and Chorus, conducted by Louis Fremaux (HMV ASD 3501), qualifies on sonic grounds. (Don't bring up the subject of the highly touted, original-version Conifer CD with John Rutter. Forward? Lacking air? It sits on your lap gasping for breath.)
Despite the Fremaux being SQ-encodedBah!it beautifully captures the ambience of Birmingham University's Great Hall. With the stereo Samsons handling what really could only be the lowest notes of the string basses and what English writer Rex Baldock used to refer to as the "sound of the floor" most of the time, the soundstage opened up even further. Solo voices somehow became "smaller," more natural-sized.
This freeing of the imaging was also heard to happen with the Telarc Bach CD mentioned above. This spaced-omni recording has the three mikes pretty close to the instrument, to judge by the dryness of the presentation. Again, stereo Samsons inserted depth between the layers of pipes in the image, to the benefit of the music. Adding that final octave of bass is perhaps not so important in terms of frequency response, but in the way it allows the stereo soundstage to breathe. It operates as a perspective enhancer.
At this point, it seemed appropriate to replace the Monitors with the Celestions (which excel in this area) to extend the system's capabilities further in the same direction.
Space! More space! (This, Gordon, is why the SL600s are in Class B of Stereophile's "Recommended Components.") With the Samsons in circuit, even further space was added, with a superb degree of magnificence to the sound. The system was now capable of reproducing organ and orchestral music par excellence. I put on my live recording of Elgar's The Dream of Gerontius. It was the best I have heard it reproduced: the Celestions' midrange seemed unrestrained; the Samsons enabled such spatial detail as the disposition of the layers in the choir, and the decay of the organ chord at the end of the climax in Part Two to be effectively resolved. They also recreating the chest-shaking feeling that the Ely Cathedral organ has in real life.
Footnote 5: Walt Jung suggests the Signetics NE5535 as a worthwhile upgrade in the Four/87 issue of The Audio Amateur.