 |
October 1997
|
By Wes Phillips Bill Eggleston builds speakers because his father did. "My dad always told me that when he started, the only way you could get really good speakers was to build them yourself. We always had drivers and parts around, and I just began building my own so early I can’t even remember. Much more important, my father passed on his wide-ranging approach to music. He listened to everything, and he taught me to be open-minded about music." You could call the EgglestonWorks Andra loudspeaker Bill’s love-letter to his dad. Maybe even to the whole family – the name honors Eggleston’s sister. Recently, this $15,000/pair compact speaker system has been garnering a lot of praise from the press and, judging from the response at HI-FI ’97, the public as well. The word is out: The Andra is no longer a family secret. The child is father to the man EgglestonWorks was incorporated in 1992. Before that, Bill pursued speaker building as a hobby while he restored houses in Memphis, among other things. His decision to manufacture loudspeakers was based on his lifelong observation of the impact his father’s hobby had on the household. "It seemed like there was always a war between having a pile of equipment in the living-room and having a neat, normal room, so I thought there might be a market for loudspeakers with fine furniture cabinetry – we had a model called the Heppelwhite, for instance. They weren’t that well received by the industry, but a few people liked them when we showed them at a WCES. That’s where I first met Peter [McGrath, now Bill’s associate in EgglestonWorks]. "We learned a lot building those speakers. I used the same Morel midrange driver I voiced the Andra around, and I learned never subcontract important elements such as the cabinetwork. The only way to maintain strict quality control is to build them in-house." Eggleston embarked upon an ambitious project: to design a loudspeaker without constraint. "Actually, we had one constraint. The speaker was designed to be as small as possible, physically. It’s very, very important to make the speakers as unobtrusive as possible. We weren’t trying to make a design statement – we wanted a speaker that would fade out of your consciousness, leaving you free to concentrate on the music." I’m not entirely sure he succeeded. The speakers are compact, but I find them striking. They have slender baffles that flare slightly to just over 15" wide at the base, in order to accommodate the 12" woofers. This gives the Andra a wide-shouldered, slightly anthropomorphic mien – somewhat reminiscent of the icons that danced across the bottom of the screen in Space Invaders. Front, sides, and rear of the cabinet are finished in a seriously glossy black acrylic coating that seems to draw the light into it rather than throw it back. Above the "shoulders" and beneath the cambered top, the side panels are covered in 1 1/4" slabs of Italian granite whose speckled surface serves as a subtle contrast to the piano finish. Grilles are provided, but Eggleston expects most of his customers will decline to use them. The grilles are well designed and are even nice looking – no extruded foam air-conditioner filters on the fronts of these babies! The grilles are stretched across steel frames; magnets buried beneath the laminate support the grille use. This means that the baffle’s finish remains unsullied by mounting hardware or velcro strips – which may encourage even more Andra owners to play their speakers nude. Overall, the Andra has a squat form-follows-function, no-nonsense appearance, combined with an understated elegance. Does it disappear into a room? Not hardly. But, I hasten to add, neither does it dominate its environment. I’d count that a success. But don’t assume for a moment that the cabinet is all glamour. This is one seriously solid speaker box. Assembly begins when two 4’ by 8’, 5/8"-thick sheets of MDF are hand laminated together using a viseolastic-damping industrial adhesive called Swedac. Swedac has a measured effect upon the absorption of vibration – Eggleston claims a solid 1.25" sheet of MDF has merely 1/10 the vibration loss of the two laminated sheets incorporating the adhesive. The laminated MDF is then milled on CNC machinery before the hand joinery commences. This, Eggleston points out, allows the company to set tolerances to within several thousands of an inch. Driver chambers are connected using dado, rabbet and biscuit joinery for great rigidity, with screws reinforcing the joints every 2 ". The photograph showing two partially assembled cabinets illustrates some of these construction details. Laminating the acrylic finish to the cabinet adds another 1/8" of thickness to each wall, as well as another layer of dissimilar material for control of resonances. The granite, mitered to a "knife"-edge to minimize sidewall diffraction, attaches to what Eggleston refers to as "the upper torso" of the cabinet. Two different adhesive compounds are used, these chosen to combine strength and dissimilar resonant characteristics. After the cabinet is assembled, T-nuts are driven through its base. These attach to plastic cups that serve as feet. Milled cones fit into these cups, so that once the speaker is properly placed, it can be spiked firmly to the floor. Final finish involves the 12-step polishing process that gives the Andra its deep luster. A wise son maketh a glad father The Andra is such a completely personal design statement that it would be futile to try to describe it using the typically reductive drivers/loading/cross-over model. Better to eavesdrop on its designer as he muses upon its evolution and construction… "My father taught me that if the midrange isn’t right, the speaker isn’t right; so our first decision is designing the Andra was the choice of midrange driver – everything evolves from that. It’s a custom version of a 6" polypropylene midbass driver from Morel that incorporates a 3" voice-coil and a double-center magnet. "I first heard an early version of this driver about 15 years ago, when my father designed a speaker that used four of them. That speaker had about the best midrange I’d ever heard. From that point on, pretty much everything I designed utilized one incarnation or another of this driver. "Its strengths are threefold. Because it uses a 6" cone and a 3" coil, all the motive force is coming from a point equidistant from the center and the circumference of the cone. This balance results in a more pistonlike action with great rigidity. "It uses a very large winding out of heavier-gauge wire than would be used in a smaller coil. Obviously, this makes for a stronger motor and, assuming you have the right current, much greater control – greater ability to stop the cone from moving in the opposite direction. The stronger the motor, the more control you have over the extremes of excursion. "And because the coil’s so big, it can handle extremely high current output. Without that, we’d never be able to connect 6" drivers directly to the output of a power amplifier. Not even two of them – our use of two 6" drivers for the midrange was integral to achieving the midrange characteristics I demanded. I find that most speakers can clip on a raucous solo piano if they use a single 6" speaker. I wanted no limit on the dynamic capabilities of the midrange, so I knew I needed to use two drivers." But just because Eggleston knew in advance which driver he was going to employ, don’t assume that amounted to a design shortcut. "Once we’d chosen that driver, we spent nine months just developing the crossover and the midrange loading. I didn’t even look at woofers and tweeters until I had assured myself that I was getting everything out of the midrange that I could. "The simplest, purest, most uncolored approach is free-air. Any dynamic speaker with a cabinet has to manipulate the back-wave. In a sealed-box design, you’re tuning against a certain air-pressure to obtain a certain response characteristic. In a vented-box, you employ tuning to eliminate certain resonances. The only box where you have the same air pressure behind the driver as in front is the transmission line. In a transmission line design, the line is equal to a quarter-wavelength of the resonant frequency of the driver. Usually you wrap the line so that the output comes from the front, in-phase with the output of the driver, which reinforces the bass output of the driver. We weren’t interested in this characteristic – all we wanted was a free-air characteristic. And we didn’t have room for two 5’ transmission lines in the small cabinet we intended to build. "So I re-evaluated the whole approach to transmission lines. I worked with the guy who subcontracts our stuffing material and we came up with "Acousta-stuff," a polyester strand which is crimped every millimeter. This makes each strand a complex shape, capable of providing greater diffraction to the soundwaves traveling through it. The strands interlock, so they wont settle at all. This material seems to slow lower frequencies while attenuating the HFs. After a ton of trial-and-error experiments, I determined the equation of cabinet length to stuffing weight we needed for a quasi-transmission-line loading. I’m not aware of any other speaker that shares this loading system." After all the work that went into the midrange, you’d be forgiven for thinking things would get easier. You’d be half-right. "I was familiar with the [Dynaudio] Esotar tweeter and I wanted to use it. The Esotar has its own aperiodic damping chamber. It has the biggest vent on its pole-piece I’ve ever seen on a tweeter – with that vent and its large chamber, the tweeter doesn’t see any back air-pressure to speak of, so it fit readily into my driver philosophy. The tweeter’s crossover is mounted in its own separate chamber directly behind the tweeter in the cabinet." But arriving at a bass response that matched the free-air characteristic of the other drivers was no easy task. Finally, Eggleston settled on using two 12" Dynaudio drivers in a configuration he dubbed "pressure-driven," There are two parallel chambers, with on 12" driver mounted in front of the other. The inner driver, which is in a heavily ported box, acts as a servo behind the outer driver – when the outer driver moves back, it doesn’t have to compress the air in the cabinet because the inner driver moves in concert with it. The driver, even though its enclosure is sealed, gets the benefits of free-air-like operation. The woofer’s crossover consists of a heavy-gauge inductor in series with the two drivers, and there’s also an RLC network in parallel with the driver. Like any loudspeaker, the Andra has a resonance peak that manifests itself electrically as impedance magnitude. The RLC network serves as a nullifying circuit, canceling the cabinet tuning so that the 6db per octave crossover can do its job. "It’s funny," Eggleston commented, "but while 90% of the Andra’s midrange tuning was done by ear, 90% of the woofer tuning was done with nearfield measurement. Bass is so room dependent that the only way you can standardize performance is through measurement." Eggleston’s crossover is certainly unusual – the midrange drivers are run full-range, while the networks on the woofer and tweeter are vestigial. This makes it somewhat difficult to specify the precise crossover points. "Actually," Eggleston explained, "there’s quite a bit of overlap. The rolloff on the high end of the midrange is very, very gradual – it begins around 3500 Hz. The crossover point to the tweeter is about 3000 Hz. We could make it tighter, but every time we’ve tried that we’ve compromised airiness and openness. The woofers run totally flat from about 20Hz up to 120Hz. The midrange is active from about 55-60Hz and up. Again, there’s overlap, but we chose to do it this way for musical reasons – and after a lot of listening. This configuration is almost like an active system. Certainly on the midrange, you’re connected straight into the amplifier." Yet he remains modest about his unique design philosophy: "We haven’t made any technological breakthroughs. All we’ve done is put together a speaker the best way I can think of. I want to strike and emotional chord with this speaker – that’s it, basically." A son faithful and true Bill Eggleston has certainly struck and emotional chord with this listener. Let me just come out and say it: I love the Andra. Off all the speakers I’ve had in my current listening room, none has sounded better over a wide range of music material. To begin with, it played loud. I don’t know if this is important to you. Hell, I didn’t even know it was important to me until I began listening to Peter McGrath’s wonderful four-channel digital recordings using EgglestonWorks smaller Rosas as rear-channel speakers. You see, Peter likes to play stuff LOUD. He claims he’s merely listening at a realistic playback level, but it think he’s trying to achieve that sense of there being no dynamic limit that is implicit in live music-making and almost completely lacking in reproduced music. Playing the Andras at high volume did give me some of the exciting U-R-There sensation, but unlike a lot of loudspeakers, I didn’t have to crank ‘em to get the music out of ‘em. In fact, the Andras reminded me of the Quad ESL-63 in that I was required to attempt to match the output of the speaker to what an instrument or band would produce live, or ruin the sense of reality. Is this a glitch or a feature? Neither really. I think it’s a sign of how uncolored the speaker was. An acoustic guitar recording such as Enrique Coria’s Latin Touch (Acoustic ACD-23), played at a realistic volume, sounded amazingly like an acoustic guitar playing in real space. But when I played it significantly louder than an acoustic guitar, it started to sound out of kilter – overtones were no longer in proportion to fundamentals, and the instrument took on an aggressive character. You might be tempted to assume that the speaker got hard at high volumes, but a rock recording, such as Warren Haynes’ searing cover of "I’ve Been Loving You Too Long" from The Memphis Horns (Telarc CD-83344), simply has to be played louder to sound real. But outside of that envelope of loudness that might be considered accurate, this recording too lost life if played too softly; or ease, it played really offensively loudly. However, when I matched the output to a believable re-creation of the original event, I could begin to appreciate the results of Eggleston’s midrange obsession. To quote Goldilocks, it was "juuuuust right." Acoustic instruments such as Coria’s guitar, Rostropovich’s cello (Bach’s Cello Suites, EMI ZDCB 55370 2), or John Coltrane’s tenor sax (Lush Life, DCC GZS-1108) were reproduced with an immediacy and coherence that I’ve seldom heard from any speaker. Could this have been because almost all of an instrument’s range was being reproduced by the same driver? I don’t know for sure, but how could it have hurt? If you love piano, you simply have to hear your favorite keyboard discs on the Andra. This shouldn’t have surprised me – once again, Bill Eggleston’s father was responsible. "My father always said you had to use the sound of the piano as the final arbiter of tonal accuracy. He was a really good pianist, and I play some too, but we always had a lot of piano music in our house, and that’s what I use to evaluate a speaker. When Duke Ellington’s alter-ego, close friend, and lifelong collaborator Billy Strayhorn died in 1967, Ellington paid tribute to their friendship with the album And His Mother Called Him Bill (French RCA NL 89166, LP). Closing the album, Ellington played "Lotus Blossom" as a piano solo. "Lotus Blossom" is a waltz, here played with great sorrow, but lifted also by that sense of grace always present in first-tier Ellington. The piano sound is muted, but rich, vibrant, and large as life. I’ve heard this song hundreds of times – through the Andra, it was my first time all over again. If you value bass heft and swing, these speakers will seduce you with rump-thump. When an orchestral bass section really dug in, as the NYP’s does in the final movement of Mahler’s Third under Bernstien (DG 427 328-2) it had an impact and organic solidity that very few speakers I’ve heard can match. In fact, after years of listening to the disc, I heard something I’ve never noticed on it before: I used to attend Avery Fisher Hall regularly, the way a sports fan supports the home team. (I think of the Mehta years as a long slump – I’d go mostly in the hopes that my team wasn’t going to embarrass themselves.) In Fisher, you can hear the stage floorboards flex in a deep moan that you’d feel more than hear. I was never able to predict when it would happen, but some nights it could startle you out of your chair, while other nights you’d never hear it at all. It happens during the Ruhevoll near the end of the disc, and it transported me back to New York most evocatively. No love to a father’s But it would be wrong to dismiss the Andra as the latest contender in the accuracy-at-any-price sweepstakes. As uncolored as it sounds, that wasn’t the only trump it had to play. The speaker was incredibly sensitive to dynamic nuance as well. Its sensitivity to tonal shading and color was matched by its ability to re-create even the most subtle variations in loudness. We usually think of dynamic change as manifesting itself sequentially, as in crescendos or diminuendos. In fact, it goes on constantly, as in the balance between the notes in a chord, or of instruments and voices in combination. The Andra clearly revealed the constant balance being achieved during, or rather that went into creating, the musical flow. This was revelatory in a very different way from the usual "analytic" accuracy that garners most of our attention, and it went a long way toward putting flesh and blood onto the skeleton of the score. Perhaps it also accounts for what I consider to be the speaker’s greatest strength: its ability to make me experience music as communication of complete emotional information. Lets face it, we hear music as tones in time, but that’s not why we love music. We love music because it connects us to a place within ourselves where we know beyond knowing, where we experience things directly that we cannot otherwise experience. It is the communication of one soul directly to another – to many others – and it releases us from the tyranny of conscious thought. Yet we discuss music reproduction as though all of that is reducible to frequency response and crossover points. I can’t point to a place on the JA’s charts that will show you where the Andra’s ability to communicate the emotional portion of a performance lies, but when I listened to Duke Ellington play "Lotus Blossom," I didn’t just hear a real-sounding piano – I heard Duke wondering how to reconstruct a life that didn’t have Billy Strayhorn in it anymore. I can point to the descending motif in Mahler’s Third that symbolizes sobbing, but that doesn’t explain why we hear those sobs so heartfelt. But through the EgglestonWorks I experienced them as though they were torn from my own chest. I could go on all night with other examples culled from my audition. It figures. I asked Bill Eggleston what his design brief was when he conceived this speaker. He thought for a minute, then quietly said, "I get this from my father, but I believe that these is no greater than discovering new music, and then coming home and listening to it in the most emotionally evocative way possible. Passion has always been my driving force – to give the life and breath of music over to the listener." Mission most definitely accomplished. Patris est filius I’m besotted by the EgglestonWorks Andra. It is an ambitious speaker, and it succeeds brilliantly at reproducing music’s sound, fury, and ineffable spirit. I’ve seldom heard its equal when it comes to conveying the pure tonal range or the magnificent dynamics of a recorded performance. But beyond that, I can think of no other speaker I’ve heard that gets nearer to the emotional nub of a performance. Dismiss this last as quasi-mystical mumbo-jumbo at your peril – music is far greater than the nuts and bolts that define it. As superb as the Andra is at presenting those nuts and bolts, it’s even better at portraying the greater truth of music’s magic and wonder. There is no question that the Andra belongs in Class A of Stereophile’s Full-Range Loudspeaker category of "Recommended Components." In fact, I’m requesting that we purchase this pair to use as a long-term reference. There’s an inscription incised in the wall of the Loyd-Paxton Gallery in Dallas: "Love instilled into solid materials by loving craftsmanship is the only creation of mankind to defeat time." That sounds about right – and "love instilled into solid materials by loving craftsmanship" seems like a better description of the Andra than any I could come up with. Bill Eggleston’s father ought to feel mighty proud of himself. Associated Equipment: LP playback: Linn LP12 with Naim Armageddon Power Supply, Naim ARO tonearm, van den Hul Frog cartridge; VPI TNT Mk.III turntable with Immedia RPM tone-arm, Lyra Clavis D cartridge. CD playback: Mark Levinson No.39, Wadia 850 Preamplifiers: Conrad-Johnson Premier Fourteen linestage, Premier fifteen phono section; Ayre K-1. Power amplifiers: Krell FPB 600, VTL 750 monoblocks Cables: Kimber KCAG interconnects, Kimber Black Pearl speaker cables. Accessories: Audio Power Industries Power Wedge 112, MIT Z series power cables, Highwire Audio Power Wrap (on components with nonreplaceable power cables), The Shelf by Black Diamond Racing, Golden Sound DH Cones. Sound treatment: ASC Tube Traps, Studio Traps, Bass Traps; RPG Abffusors; Catillac Chat de Ville Towncat.
The EgglestonWork’s Andra is moderately sensitive, my estimated 87db/2.83V/m (B-weighted) reading agreeing with the specified figure. Its impedance plot (fig.1) suggests that the speaker appears to be an easy load for an amplifier to drive, the magnitude dropping below 10 ohms only in the high treble and below 130 Hz, and the phase angle generally remaining low. However, it does reach 4 ohms in the low bass, around the tuning of the twin ports. There is a very slight wrinkle in the impedance traces between 300 and 400 Hz, this generally indicative of a cabinet resonance of some kind. However, as might be expected from its massively constructed cabinet, the Andra is impressively inert. About the only resonant mode I could find was on the woofer enclosure wall (fig. 2) at 350 Hz. This is well down in level, however. In his interview with Wes, Bill Eggleston mentioned the acoustic overlap between the woofers and midrange units. This can be seen in Fig. 3, which shows the individual nearfield responses of the front woofer, one of the ports, and one of the midrange drivers. The latter can be seen to start to roll out below 120 Hz or so, but it still contributes to the speaker’s output and octave lower, in the midbass. The woofer’s output is the trace that peaks around 30 Hz and again between 60 Hz and 80 Hz, before gradually tapering off. It is still only 10db down from the midrange level at 500 Hz. The ports cover a broad region from 20 Hz to 70 Hz, with then an irregular rolloff. Predicting how these three nearfield responses will sum at the listening position is difficult; shown to the left of fig. 4 is my best guestimate, with the acoustic phase and path-length differences taken into account. The Andra’s output appears to peak up slightly at 70 Hz, then slowly rolls off, reaching its $6db point at a very low 22 Hz. In-room, with the usual amount of low-frequency room gain, the Andra should easily extend to 20 Hz at full level. Note that both the woofer and port have slight peaks at 200 Hz; while these result in a slight peak in this frequency in the calculated farfield response, it should be subjectively inconsequential. To the right of fig.4 is the Andra’s farfield response, measured on its tweeter axis using DRA Labs MLSSA system and averaged across a 30 degree horizontal window. The midrange trend in fig. 4 is basically evenly balanced, but note the enormous measured suckout in the upper crossover region on this axis, centered on 3khz. I must admit that I didn’t find the Andra to be as free from coloration as WP did. I noted a slight degree of hollowness that made violin and viola, for example, sound a little as if played with mutes. However, this was much milder than I would have expected from this measured response. The tweeter is 33" from the floor, which is on the low side. (Tom Norton’s research has shown that a typical listener’s ear in a typical chair is 36" high.) Fig. 5 shows the Andra’s response at different heights; it can be seen that the crossover-region suckout is worst on the tweeter axis. The broad overlap between the tweeter and twin midrange units does appear to make the speaker very sensitive to listening height. Perhaps the flattest measured response is obtained 10 degrees below the tweeter axis (the trace at the front of this graph). However, this represents a listener with his ear around 20 " from the floor. By contrast, fig. 6 shows the Andra’s measured response 10 degrees above the tweeter axis, which represents a typical listener sitting in something like a director’s chair. While there is still a lack of energy in the speaker’s upper crossover region, it is much less severe than on the tweeter axis. It must be remembered that a loudspeaker’s perceived balance doesn’t depend only on the direct sound – the which first reaches the listener’s ears – but also on the total radiated power into the room. Fig. 7 shows how the response on the tweeter axis changes to the Andra’s sides. (The data points are sparse in this graph, due to the fact that the Andra is too heavy from my automated speaker turntable.) Again, the crossover suckout is worst directly on-axis, the "horns" between 2 and 4.5kHz to the sides of this graph suggesting that the speaker’s total output into the room does not feature a lack of energy. Only in a small room, therefore, with the listener sitting close and low, will the Andra sound hollow. The larger the room and the farther away the listener, the better balanced the EgglestonWorks will sound. Despite its sloped baffle and low-order crossover, the Andra is not time-coherent on typical listening axes. (It will be time-coherent around or below 20 " from the floor, which suggests the tiltback of the baffle is too mild.) Fig. 8, for example, shows the step response on the tweeter axis. The tweeter’s output is the sharp up/down spike just before the 4ms mark, followed by the midrange units in the same acoustic polarity. As was the frequency response, the nature of the Andra’s cumulative spectral-decay plot depended very much on the measurement axis. Fig. 9 shows the waterfall plot associated with the response in fig. 6, 10 degrees above the tweeter axis. In general it is impressively clean, though there is some low-level hash present in the mid-treble. But if the microphone was lowered by 5 degrees, nearer the tweeter axis (fig. 10), a resonant made at 4.7kHz appeared, associated with a response peak at the same frequency. This and the excess of top-octave energy might tend to make the balance rather bright, everything else being equal. Note that I did not measure the speaker’s distortion performance. However, I suspect that it is superbly low. |
