Here we go again. It is becoming painfully obvious that once you get past a certain price point, more money will not necessarily buy you better sound. Especially not in the delirious world of high-end, high tech speaker systems, six of which are reviewed here.

This is in effect a dipole speaker survey. All of the speakers discussed below are open-backed dipole radiators, except the Dahlquist DQ-20 (which tries to look like one but is not). Dipoles are more difficult to measure than sealed or vented boxes; for one thing, the highly accurate Keele method of measuring the low-frequency response with the microphone almost touching the diaphragm does not work with an open baffle because the opposite-phase front and rear waves cannot be summed at that point. The room also becomes a more intimate part of an open-baffle system than of a box speaker, the distance from the back wall being a particularly significant variable, and thus the repeatability of the measurements is inevitably compromised.

Our laboratory tests of these speakers should therefore be considered exploratory or diagnostic, seeking measurable causes of audible effects, rather than an attempt to obtain certifiable performance figures. Where there is a significant glitch or design flaw, you can be reasonably sure that we have nailed it, and we do believe that our procedures are thoroughly sound and realistic, but we are not the National Bureau of Standards and our numbers cannot be interpreted as "official." Apogee Scintilla Apogee Acoustics, Inc., 35 York Industrial Park, Randolph, MA 02368. Scintilla three-way full-range ribbon loudspeaker, $3995.00 the pair. Tested samples on loan from owner.

To begin with, this very beautiful-sounding but crazily amplifier-dependent speaker system is not all-ribbon, as claimed. Nor is any other Apogee model, old or new. The woofer design in these speakers is more or less the same as in the familiar Magneplanar systems, appropriately called "planar-magnetic driver" by Magnepan and fictitiously yclept "a bona fide ribbon" by Apogee, probably because they feel that only a 100% ribbon system has the required high-tech image. We are not particularly upset by the false nomenclature, since a planar-magnetic woofer by any other name would sound as sweet (sorry, Will)-if it were a sweet design concept to begin with, which it is not. More about that in a moment.

The heart of the Apogee Scintilla, and its indisputable claim to fame, is the combined midrange/tweeter assembly with five loosely suspended four-foot ribbons, one wide and four narrow, within an integrated magnet structure. It is a true ribbon transducer, although a little on the idiosyncratic side, or shall we say an engineering think piece? One grows quite fond of it, watching those ribbons flop this way and that, with more degrees of freedom than would seem correct, but producing delicious sounds nonetheless. The two narrow ribbons in front move forward in response to a positive signal, while the two in the back move rearward, so that the tweeter is an approximation of a pulsating cylinder. The wider midrange ribbon, sandwiched in between, moves in phase with the front tweeter ribbons and with the woofer panel. The overlap in flat response capability from woofer to midrange to tweeter permits the use of first-order (6 dB per octave) slopes in the passive crossover with impunity.

The woofer is another story. The thin plastic sheet to which the conductive foil grid is bonded (with a horizontal orientation of the elements, in contrast to the vertical pat tern of the Magneplanars) is in the shape of a right-angled trapezoid, four feet tall, and is clamped around its perimeter.

We never liked this "drumhead" approach to bass reproduction (see our 1978 and 1979 Magneplanar reviews), since it tends to leave the plastic diaphragm grossly underdamped, regardless of the latter's shape and without the possibility of a cure. The resulting high-Q peak in the response is equalized out to some degree by the slope of the open-baffle bass cancellation; in fact, that is the basic principle which Bob Carver quantified and harnessed so cleverly in his speaker (see the Carver review below), but in the Scintilla and other Apogee designs this natural phenomenon appears to be hap-

hazard and the dominant peak remains, always followed by a characteristic dip (analogously to the familiar peak-dip filter). What the trapezoidal taper of the diaphragm achieves is a widening of the effective passband through the gradual staggering of resonant modes; however, the excursion of the diaphragm is in resonant segments as a result, so that only part of the total surface is available to move the air at any given frequency. In other words, the woofer is not as big as it looks.

We could go into further details on all this (see also the Apogee Diva review below), but the truth is that we have limited patience with plastic-sheet woofers-ribbon, planar-magnetic, electrostatic or what have you. As our erstwhile associate, Bruce Zayde (now up to his golden ears in computers at Hewlett-Packard), used to say when the talk turned to the drawbacks of conventional electrodynamic speakers, "the woofer is not the bad guy." Cone-type woofers, in correctly designed enclosures or baffles, are very hard to beat and still at the head of the class in our school. We find the conceptual appeal of "all ribbon" or "all electrostatic" quite specious, especially when there is a crossover net work in the system. It came as no surprise, therefore, that the low-frequency performance of the Scintilla is not its long suit. First of all, the woofer panel buzzes and breaks up all over the place when swept sinusoidally-at all levels.

Then there is a huge peak in the 1-meter response at around 32 Hz, some 14 or 15 dB above reference level, varying slightly with the exact microphone location. It is followed by smaller peaks, of the order of 4 to 6 dB, centering on 100 Hz and 220 Hz. In listening quality the bass is not as flawed as the lab tests would indicate; it is merely unremarkable. Reviewers who raved about it must have been accustomed to Celestion SL600's.

By contrast, the 1-meter response above 400 Hz or so is impressively flat over almost five octaves up to 10 kHz, after which there is a dip of 8 or 9 dB to 15 kHz, followed by a return to reference level at around 19 kHz. The dip is the result of phase cancellation, confirmed by tone bursts and clearly due to the reversed polarity of the two tweeter ribbons in the rear. Elsewhere in the audio range, the quintuple ribbon assembly reproduces tone bursts very well and the woofer panel not too badly, with just a bit of ringing.

Pulse replication is quite coherent, since the woofer panel, the midrange ribbon and the forward-firing tweeter ribbons are all in phase. The audible outcome of all this is absolutely gorgeous sound from the lower midrange upward- uncolored, sweet, liquescent, beautifully detailed and un trammeled in dynamics-demonstrating the low distortion, high power handling and superior wave launch of a well designed ribbon. The stereo image and dimensionality are somewhat dependent on room placement and the tilt of the baffles, as can be expected, but leave nothing to be desired after a little massaging.

Now, cancel every bit of the positive commentary above, forget the Scintillas altogether, unless you have the equipment to drive them with. Here we come to the crux of the matter. The nominal impedance of the speaker is 1 ohm; we measured 1 to 1.4 ohms up to 2 kHz, after which there is a smooth rise to a plateau of 2 ohms that extend from 5 to 20 kHz. A ridiculous load for normal amplifiers, even some of the best. At the same time, the efficiency is miser ably low; an input of 1 watt yields 73 dB SPL at 1 meter according to our measurements. With an input of 2.83 volts (representing the sensitivity reference of 1 watt into 8 ohms) the SPL at 1 meter is 82 dB. Thus, to obtain 103 dB SPL in mono and 106 dB in stereo, which is roughly a Wagnerian climax as heard in the first few rows-and let us not even bring up rock concert levels-you need an amplifier that can deliver 1000 clean watts into 1 ohm. The Krell KMA-200 or the even costlier KRS-200 can certainly do that; some of the lesser Krells come close but no cigar; the Carver "Silver Seven" vacuum-tube power amp (see the review in this issue) was the most suitable we had on hand for these tests and it, too, comes close but quits in the high hundreds (with a 1-ohmload, that is). There may be a few other options for the faithful, but the point is that adequate amplification will in all probability cost even more than the speakers. We find that hard to approve of.

The main reason for this review of a more than three year old product-just when the Diva is being touted as the latest-and-greatest super speaker from Apogee-is that the Scintilla is still, in our opinion, designer Leo Spiegel's best effort to date. After the original Apogee Full-Range speaker (now $10,000.00 the pair), which was a kind of lab oratory of ideas in which first-time shortfalls were almost inevitable, it was the Scintilla that emerged as the highly creative albeit somewhat impractical solution to the defined complexities of the problem. Subsequent models seem to reflect the pressures and constraints of the marketplace and possibly subjective second-guessing by dilettantes with a say-so in product development-that is our intuition, not our information. The speaker that makes us respect Apogee and Leo Spiegel is the Scintilla, warts and all.

Apogee Caliper

Apogee Acoustics, Inc., 35 York Industrial Park, Randolph, MA 02368. Caliper two-way ribbon loudspeaker, $1995.00 the pair. Tested samples on loan from owner.

To our mind, this is a poorly conceived product. The Apogee format does not translate well into a reduced-size speaker system. Just because the marketing people need an entry-level model to be able to offer their dealers a full line, the inherent engineering problems cannot be expected to go away. The Caliper is in effect a small-signal loudspeaker that uses an unnecessarily complex and costly technology to achieve its limited performance. Although not nearly as large, intricate and expensive as the other Apogee speakers, it is still too much of all that for what it does.

The woofer panel buzzes at numerous frequencies with a sinusoidal input of only 6 watts or thereabouts; we wish we could say that such is not the case when the signal is music, but the fact is that heavy symphonic or rock bass quite regularly excites the resonant modes to the point of audible buzzing. The fundamental resonance of the panel is at 36 Hz, where there is an 8 dB peak. That defines the low frequency cutoff of the system. All general comments about the design of Apogee woofers in the Scintilla review above apply here as well.

The single-ribbon midrange/tweeter of the Caliper is, by contrast, quite well-behaved when swept and sounds fairly smooth, possibly a little hot in the 2.5 to 4 kHz region but not offensively so. It moves forward in response to a positive signal, whereas the woofer moves rearward. Wiring the drivers out of phase in a system with 6 dB per octave crossover slopes is strictly an amplitude-response tailoring expedient and not very sound practice, since it destroys coherence. The overall response curve of the Caliper, taken at the best "sweet spot" we found for the microphone after considerable hunting, shows a downward-sloping trend from 300 Hz to 1.5 kHz, dropping a total of 8 dB, followed by a generally flat portion from 1.5 to 20 kHz, except for the slight "presence" peak already noted. Between the bottom end and 300 Hz, the peak-dip filter profile of all the Apogee woofers is evident.

The resulting sound, if we absolutely had to check one of two boxes labeled "good" and "bad," would still have to be classified as "good," since most speakers sound worse.

That, however, is no endorsement. On complex program material played at a room-filling level, the Caliper always sounds a little raucous; the bass, as we said, goes to pieces; the top end, though smooth, tends toward brittleness. With one driver pushing while the other is pulling, and the consequent discontinuity fairly high up in the midrange, it should come as no major surprise that the speaker is not particularly coherent-sounding, either.

As for amplifier compatibility-no problem. The nominal impedance of the Caliper is 3 ohms, and the SPL we measured at 1 meter with 2.83 volts input was 85 dB.

Thus a nominal 100-watt amplifier as recommended by Apogee should be more than adequate, since the limiting condition will be the breakup point of the speaker.

Apogee Diva Apogee Acoustics, Inc., 35 York Industrial Park, Randolph, MA 02368. Diva three-way full-range ribbon loudspeaker, $7000.00 the pair. Tested samples on loan from owner.

"The Diva will redefine the state of the art," says the Apogee blurb. Obviously, this speaker system is Apogee's everything-we-know-is-in-here statement as of late 1987.

The overall dimensions and basic "architecture" of the Diva are comparable to those of the half-decade old and defiantly impractical Full-Range flagship model, which will presumably be retired now, and everything about the new speaker is clearly intended to proclaim ecce machina! We must regret fully note, therefore, that we are not buying any of that.

It is indeed sad, bordering on the tragic, that the Diva is not as good as the considerably older and less expensive Scintilla, except for being much easier to drive. The older model proves that Leo Spiegel knows how to design a large and highly sophisticated ribbon loudspeaker whose output greatly resembles its input (disregarding for the moment the shortcomings of the plastic-sheet woofer, which result strictly from the "all-ribbon" constraint-see above). The Diva, on the other hand, appears to have a formatted output, as if someone had decided how it must sound, regardless of how it measures. All we can say is, if you are going to do that, you had better have exquisite taste.

The frequency response of the Diva can be manipulated to some degree by using a variety of adjustable settings on the externally connected passive crossover. Small trims of the order of +2 dB may thus be applied to four different segments of the response curve. Since the formatting of the overall response profile is much grosser than that, such a feature is largely a red herring. The most obnoxious departure from a smooth (let us not even talk about flat) response is a broad hump centering on 400 Hz and extending over two octaves from approximately 250 Hz to 1 kHz. It goes up 6 to 7 dB and comes down 12 or 13 dB with respect to its surrounding valleys. No matter where the speaker is placed in the room and regardless of the microphone distance or location, this bulge of excess energy is always apparent.

Since the older Apogee models are considerably more neutral (though not perfectly so) over the same two octaves, it would appear that such a midrange boost is somebody's idea of an improvement. Its audible effect is an unmistakable and unmusical honk whenever the program material has a fairly complex texture. It is the Diva's clubfoot.

The woofer panel of the Diva exhibits the same family traits that we faulted in the Scintilla and Caliper woofers, only more noticeably because of its larger size. The tapered trapezoidal diaphragm, some five and a half feet high and peripherally clamped, does not move the air as a single large piston but in staggered resonant segments. The fundamental resonance of the lowest and widest segment is 32 Hz, where the high-Q peak defines the bottom limit of the speaker. At that frequency, a sinusoidal input of 30 watts makes the panel rattle quite badly. Maximum excursion of the diaphragm is +/- 1/6" (1/3" total). The characteristic and inevitable dip in the upper bass centers on 100 Hz.

Above 1 kHz, all the way up to 10 kHz, the 1-meter response of the Diva is reasonably flat, though not quite as flat and smooth as that of the Scintilla. In the 10 to 20 kHz octave there is some decline in output, with a profile dependent on microphone location, but nothing objectionable. A dominant influence on the overall trend of the speaker's frequency response is the polarity of the drivers as connected to the crossover network. With a positive-going signal as the reference, the woofer panel is negative, the midrange ribbon positive and the tweeter ribbon again negative in polarity.

In a system with 6 dB per octave crossover slopes, this seems utterly perverse-the equivalent of a Band-Aid in frequency-response doctoring. As a direct result, wave-front coherence is shot to hell; there is no coherent pulse shape obtainable out of the Diva at any microphone location, regardless of pulse width. Nor can tone bursts of any frequency be reproduced without ringing; everything seems to be a little underdamped.

Lest anyone should jump to the conclusion that these characteristics result in catastrophic sound, let us hasten to declare that the total impact of the Diva is that of a very clean-sounding loudspeaker. That 400-Hz honk is extremely disturbing; the compromises in coherence are easily audible; the bass is not nearly as impressive as that of, say, the Carver speaker; the frequency balance is subject to instant improvement by means of ordinary tone controls-still, in a world populated largely by bad speakers, this is a good speaker. It is disappointing only because its size, price and marketing promise true greatness.

As far as suitable amplifiers are concerned, those who ignore our advice and buy the speaker will probably opt for a biamped or tri-amped configuration-and why not?- although the Apogee literature specifies a single 100-watt amplifier (presumably rated at 8 ohms) as sufficient. To us that seems like a bit of wishful thinking. The nominal impedance of the Diva is 3 ohms; we obtained a gently undulating impedance curve that stays closer to 4 ohms over most of the audio range, never drops below 3 ohms until it reaches 7 kHz and finally declines to 2 ohms at 20 kHz. The SPL we measured at 1 meter with 1 watt input (efficiency) was 76.5 dB, with 2.83 volts input (sensitivity) 80.5 dB.

Thus, the Krell KMA-200 or KRS-200, rated at 200 watts into 8 ohms and capable of putting a steady 40 volts into just about any load, will drive the Diva only to 103.5 dB (106.5 dB in stereo), which is adequate but far from awe some. We used the Carver "Silver Seven," which can drive 3 or 4 ohms quite a bit harder, and were totally satisfied with the dynamic range per se. A big stereo power amp like the Boulder 500 bridged for mono would also provide sufficient drive.

The question that remains is whether the Apogee Diva could have been a great speaker with the same technology but a different engineering/marketing philosophy. Future products from Apogee may shed some light on that one.

Carver "Amazing Loudspeaker"

Carver Corporation, P.O. Box 1237, Lynnwood, WA 98046. "The Amazing Loudspeaker," $1576.00 the pair. Tested samples on loan from manufacturer.

Leave it to Bob Carver to name a product in such a way that you cannot refer to in print or in speech without glorifying it. A cunning little marketing ploy ("Bob, this is an amazing loudspeaker, it really is." "Hey, let's call it that!" "Call it what?" "The Amazing Loudspeaker." "I love it, I love it!"), but we must confess that we were initially quite turned off by the name. Now that we have tested the speaker, we are in a better position to ask whether it is indeed amazing. The answer, in a word, is-yes. In several respects, the Carver system is an embarrassment to speaker designers and manufacturers with many more years of experience than Bob, your Editor being one of them. It solves certain design problems and achieves certain sonic results with a simplicity and flair that can only be called, well, amazing.

The amazement begins as soon as you unpack the speaker from its shipping carton. How can this sell for $1576.00 the pair? The flat, trapezoidal, open baffle is five and a half feet tall and finished in hard black lacquer, almost like a Steinway; there are four 12" woofers per side in a vertical array, flanked by a very high-tech-looking five-foot ribbon-if the manufacturer were, say, Infinity, the price could be easily three times as high before the question of sound quality even came up.

Unfortunately, the Carver speaker does not sound so amazing right out of the box. The drivers must be broken in before they settle down to their proper operating parameters.

The fundamental resonances of the woofers and the ribbon are too high before break-in, and there are also stress modes in the clamped ribbon that must be homogenized out, so to speak, by exercising it. Carver recommends approximately 50 hours of use before critical listening, and our experience confirms that number. Without break-in the upper bass and lower midrange are a bit thick and woolly, and there is a somewhat glary or electronic quality from the upper mid range on up to the lower highs. These faults are gradually, and in the end totally, wiped out as the speaker is played, but they account for the generally poor impression made by the Amazing at trade shows and also in many dealers' show rooms. No trade show is open to the public for 50 hours, and not many dealers log 50 hours on a demo pair before letting their customers listen. This is not a trivial matter; we consider it to be a serious shortcoming and fervently hope that Bob Carver will do something about it.

Once the speaker is working properly, the bass is the next source of amazement. Everybody knows that open baffle speakers without electronic equalization have no bass, or at best just boomy upper bass with no real sock down below-right? How could it be otherwise with the self cancellation of the opposite-phase wave fronts launched by the dipole? That is why we Thiele/Small nerds have always insisted on box speakers, despite all the wall-deadening, bracing, stuffing, duct-tuning, etc. pains that come with the territory. Now Bob Carver, the upstart speaker designer (with nothing but the dreadful Phase Linear Andromeda III of 1976 under his belt), has come up with a devastatingly simple and sweepingly efficacious insight into open-baffle woofer behavior that changes the ground rules and results in flat response down to the limits of audibility. Frankly, we are jealous as all hell. Somebody should have thought of this elementary solution long ago.

Here is how it works. We have all seen, in a book or an article, the classic curve showing the effect of open-baffle bass cancellation. When the woofer is mounted asymmetrically (i.e., when the front-to-back paths are not all the same length), there is a smooth 6 dB per octave slope with a turn over frequency corresponding to the half wavelength defined by the average front-to-back path. This curve prevails as long as the free-air resonance of the woofer is at a much lower frequency than the turnover. Now let us recall another diagram on a different page of the same or perhaps another book or article. There we have seen the family of curves representing the infinite-baffle low-frequency response of a woofer with different values of Q. Starting with the rolled off response when Q = 0.5 (critically damped), progressing through the Butterworth profile with Q = 0.71 (maximally flat), just beginning to show a little ripple at Q = 1, then clearly reaching the bumped-up stage when Q = 1.41, the hypothetical woofer finally displays a heavily bass-boosted output when the Q is higher than any you can find in the vendors' catalogs. A sufficiently high Q will result in a slope of just about 6 dB per octave on the right side of the bump. Eureka! Bob Carver, on the road to Damascus or Seattle or the bathroom, made the connection between the two familiar diagrams. You cut away the infinite baffle until it becomes exactly the right open baffle, then the high-Q rise of 6 dB per octave and the cancellation drop-off of 6 dB per octave add up to flat response right down to the free-air resonance. It is a much more elegant solution than electronic equalization of the cancellation slope 4 /a Enigma or Celestion, since it does not cost additional amplifier power and demands the use of a very small magnet to achieve the high Q. The penalty paid is that the free-air resonance, the Q and the size/shape of the open baffle become extremely critical parameters, unforgiving of design laxities and production tolerances.

All theory aside, the bass of the Amazing, as we already intimated, is amazing. Its bottom limit is defined by the 22-Hz free-air resonance of the woofers after break-in; the -3 dB point is a few cycles below that. The 12" woofers have exceptionally long linear travel and can move a lot of air. Bob's taste runs to somewhat heavier bass than ours and he set the Q to create a very slightly tipped-up response at the lowest frequencies, but you can easily flatten that out with a line-level passive equalizer network (included in the package as the Sub-Bass Room Damper-there you go again, Bob), which has outputs for Q = 0.7 and Q = 0.5. In many rooms the unequalized speaker will sound just right.

One thing is certain: no subwoofer is needed! The bass is at all times completely reproduced, with excellent definition, impact and musicality, unless of course the room placement is totally unsuccessful. (A separate article on room placement, of open-baffle as well as box speakers, is scheduled for the next issue.) The ribbon is amazing for entirely different reasons.

No eureka insights here, just "blood, toil, tears and sweat" to obtain the desired characteristics at low cost. The basic structure is quite similar to that of the Strathern or the Gold 3.0 ribbon, but the design details, materials and construction techniques are simpler and better. Resonances and other anomalies have been painstakingly massaged out of the design, and whatever response irregularities remained are equalized out in the crossover network. It is a case of not accepting the limitations of an existing format and just hacking away at the faults one by one until they are gone.

Amazing perseverance. The resulting frequency response is almost ruler flat from 200 Hz to 15 kHz, except for a so called Gundree dip (apparently a BBC-derived expression) of 3 to 4 dB spread over the octave from 2.5 to 5 kHz. This is a bit of deliberate "voicing" introduced in the network to satisfy a certain school of thought (not ours); it is basically innocuous and of minor importance. From 15 to 20 kHz there is a 4 or 5 dB drop, normal in a large ribbon. From 200 Hz down the response is generally elevated a couple of dB above reference level (independently of the bump at the extreme bottom); here again we are dealing with voicing, this time for an overall "warm" balance, which is not exactly our taste but persuasive enough to be accepted as Bob's right to his own priorities. Remember, these are subtle effects, not gross formatting as in the Apogee Diva. We also disagree with Bob's decision to connect the ribbon and the woofer array with opposite polarities to the crossover network. The saving grace of this dubious frequency response trimming practice is that the crossover frequency for the 6 dB per octave slopes is 100 Hz, which is low enough to make the discontinuity in wave-front coherence inconsequential.

The nominal impedance of the Amazing is 4 ohms; actually the curve meanders all over the audio band, rising as high as 14.5 ohms at 170 Hz and gradually declining at the higher frequencies to as low as 2.3 ohms at 20 kHz, all on account of the equalization network, not the drivers. We estimate that 6 ohms would be a better figure for calculating typical power dissipation with music signals. Sensitivity is quite low: 82 dB SPL at 1 meter with 2.83 volts input, on a par with the Apogee Scintilla, but the Carver is much more efficient because of its higher impedance. With 1 watt input as referenced to 6 ohms, the SPL at 1 meter comes out at 80.5 dB, still a far cry from efficient box speakers but adequate for use with amplifiers rated at 200 watts or better.

The Amazing will put the biggest amplifiers on their mettle; it just gets louder and louder without distortion as you turn up the volume and never buzzes or breaks up at any frequency. You need 200 watts but you could use 1000.

All told, what kind of musical sound do these amazing and not so amazing features add up to? Excellent sound, on the whole preferable to that of the Scintilla, with far better bass and of course saner amplifier options. Both speakers sound equally clean, sweet transparent and musical; the Carver is perhaps a bit more reticent or "polite" in the upper midrange and lower treble because of the somewhat warm balance preferred by Bob but makes up for it in full-range power handling. We would not hesitate to choose the Amazing over an Apogee ribbon speaker even if price were not the issue, as it usually is, simply because we consider bass quality, freedom from breakup and general practicality more important than nuances of tonal balance. If Bob could fix the three boo-boos of the speaker-the need for lengthy break-in, the relative inefficiency and the opposite-phase driver connection-our enthusiasm would know no bounds.

The Amazing Loudspeaker is, even so, a dyed-in-the wool high-end audio product at a medium price. Rara avis!

Dahlquist DQ-20

Dahlquist, Inc., 601 Old Willets Path, Hauppauge, NY 11788. Model DQ-20 Phased Array loudspeaker system, $1800.00 the pair. Tested samples on loan from dealer.

This is clearly the successor to the classic DQ-10 of the mid-1970's, the speaker that made Jon Dahlquist one of the lares et penates of the audio salons. We have known Jon for about eighteen years and have always respected him both as an engineer and as an audio philosopher; however, as early as the winter of 1977-78 we began to feel that he had put too much faith in the permanence of the DQ-10 design and that it behooved him to come up with an improvement.

That did not happen until about three years ago, when the DQ-20 began to be phased in, and the DQ-10 was not dropped from the catalog for another year or two, after a longer run than some of the most successful Broadway musicals. It all happened so gradually that we do not even recall seeing a review of the DQ-20.

The most immediately obvious thing about the 42" high, floor-standing DQ-20 is that it tries to look like an open-backed dipole speaker but is not, thereby continuing the tradition established by the DQ-10, which was an unabashed early-Quad look-alike. The DQ-20 consists of a 10" woofer in a sealed box of 1.4 cubic feet internal volume (our estimate), a 5" cone midrange driver in a small tube vented pod and a 3/4" dome tweeter in one of those little Scandinavian quasi-horns. The three separate units sit one on top of the other in a reflection/diffraction-reducing formation (d la B&W); the Quad-like screen enfolds them with so much air to spare that the speaker looks twice as big as the equivalent normal cabinet would-and thus more deserving of its high price tag. A mildly deceptive bit of packaging with a plausible engineering alibi (we can just hear it): you must keep the frame far away from the midrange and tweeter to avoid reflections. Very good, Jon.

The specifications place the crossover frequencies at 400 Hz and 3.5 kHz; our measurements provided no clue as to the exact configuration of the crossover network but, as we shall see, it must be fairly complex. The venting of the midrange enclosure may conceivably be an inexpensive but perfectly good way to synthesize a third- or fourth-order high-pass filter characteristic without L's and C's. The bass enclosure appears to be perfectly matched to the woofer, the Q being 0.71 or thereabouts, the bass response nice and flat, and the --3 dB point at the box frequency of 43 Hz. There are ways to get deeper bass than that out of a 10" woofer, but what there is seems very accurate.

The overall frequency response as measured on the midrange/tweeter axis 75 cm from the grille is perhaps the flattest we have ever seen in a three-way electrodynamic speaker system. The upper limit is a hair short of 20 kHz; there is an unmistakable dip, say 5 dB, at around 1.8 kHz but (get. this!) it disappears 30° off axis where the actual stereo listening takes place. The off-axis response is almost frighteningly flat, actually flatter than the axial response except on the extreme top end, leading to interesting speculations regarding Jon's methods and priorities. The plot thickens as one discovers that the DQ-20 Phased Array is not phased. Notwithstanding the staggered setbacks of the midrange and tweeter, it is not possible to recover a square pulse of any width from the speaker, not even vaguely or raggedly, not at any distance, not at any height. With respect to a positive-going pulse, the driver polarities are woofer plus, midrange minus, tweeter minus-what kind of network is this? The impedance curve provides a partial clue; it is a roller coaster with very abrupt transitions, for example from the 4-ohm minimum at 1.1 kHz to 18 ohms at 2.5 kHz, and with excessively large phase angles throughout. Not an easy load for an amplifier; something like the Boulder 500, rated at 250 watts into 4 ohms and rock stable, would be a good recommendation although not quite compatible in price. The efficiency of the speaker, while perfectly adequate, is not so high that a run-of-the mill 100-watter would be a good choice. (We forgot to take SPL readings at 1 meter before returning the speakers, alas.) We have a theory about this mixed bag of engineering data. Jon Dahlquist is definitely a purist. We are quite sure he was aiming for perfection in both the frequency and the time domain, i.e. for a flat and coherent output. It is also our understanding that he is set up to make instantaneous and unlimited LCR changes in a crossover network while listening. He must have found, given the available drivers and the general architecture of his prototype, that he could make the output either flat or coherent but not both-the usual empirical conclusion, especially with three-way systems. The only escape route out of such a dilemma is offered by computer-aided network analysis, filter synthesis and iterative optimization, all of which takes sophisticated programs and many hours of computer time. The DQ-20 shows no evidence of having been subjected to such procedures, which could probably have solved the problem. What most likely happened was that the time came to go into production, further delays were unacceptable, and a command decision had to be made whether the speaker should be flat or coherent (i.e., phased). Jon must have opted for flat but was apparently unwilling to give up the Phased Array appellation. Only a theory, as we said.

We must admit that this dead flat but not really phased array sounds extremely neutral, uncolored, balanced and unexaggerated in every way. If we had to steer Bob Carver, for example, toward a model of "voicing," this would be a good one. The relative freedom from transient ringing also adds to the just-right feeling in this respect; our tone-burst tests showed each driver to be well behaved, the midrange perhaps least so. On the other hand, the DQ-20 lacks the solid dimensionality, palpable airiness and spatial delineation evident in the sound of more coherent speaker systems.. There is something aesthetically flat about that acoustically homogenized flat response; perhaps the very excellence of the output in the frequency domain makes one long for greater excellence in the time domain. We were left vaguely dissatisfied though far from unimpressed. To be more specific, we preferred the frequency balance of the Dahlquist DQ-20 to that of our own Fourier 8e (now a defunct design but still our standard in a medium-sized three way) but definitely heard a bit more information through the more coherent Fourier.

A very good showing, everything considered, but we cannot imagine anyone not choosing the Carver "Amazing Loudspeaker" for $224 less the pair, unless vertical size is the decisive factor. The Carver is the more exciting audio component and the more complete speaker.

Martin-Logan CLS

Martin-Logan, Ltd., P.O. Box 741, 2001 Delaware, Lawrence, KS 66046. The CLS full-range electrostatic loudspeaker, $2490.00 the pair. Tested samples on loan from owner.

Gayle Sanders may very well be the most gifted of the self-taught audio designers who started out as dealers. Never mind that the genre abounds in charlatans and jackasses; Gayle is the rare prince among them, a genuine whiz, the exception that proves the rule if indeed there is one (let us hope not). Here he has created a speaker so beautiful to look at, so intriguing in design details and generally so attractive as an idea that the reviewer begins to root for it before the first note is heard and just wants it to pass all tests with flying colors. Such emotions are of course an invitation to disappointment, which in this case is happily not too big.

The CLS is a full-range, crossover-less, electrostatic dipole loudspeaker. Its active element is a four-and-a-half foot high, monolithic, seamless, gently curved, push-pull electrostatic "sandwich" of unique construction and appearance. For one thing, you can see right through it, so it looks smaller than it is because it does not block you line of vision. For another, it makes you wonder how Gayle managed to stretch the plastic diaphragm so beautifully between the two curved and perforated stationary electrodes.

(Try to bend a thin Swiss cheese sandwich into a curved shape without kinking the bread or the cheese.) That diaphragm is very special; an extremely laborious graphite treatment invented by Gayle gives it unprecedentedly high resistivity and helps eliminate arcing, the major plague of electrostatic speakers. The perfectly smooth, edgeless perforations in the stationary electrodes remove another cause of arcing; they are made possible by a dielectric paint of unique properties, also developed by Gayle. The uncovered, grille less high-voltage speaker is as safe to pat and stroke with your bare hands as the family dog, and as a final touch of virtuosity the entire structure is assembled without nuts, bolts, screws or any other hardware-just glued and force fitted. In sheer "techie" appeal the CLS is hard to beat; you want to own it even if you never play it.

It should be pointed out that Gayle Sanders has modified the CLS many times; we tested the version that was the latest-and-greatest as of the beginning of 1988. The fundamental resonance of the diaphragm is at 50 Hz, which more or less defines the low-frequency limit of the speaker.

Everything we said about plastic-sheet woofers in connection with the various Apogee models applies equally to the CLS. The driver Q per se is approximately 3.5, resulting in a huge near-field peak, but the open-baffle cancellation effect (see the Apogee and Carver reviews) synthesizes an accept ably level bass response with just a bit of the inevitable peak-dip profile. The overall frequency response of the speaker shows a general downward trend, and pretty ragged at that, but with the microphone in the farfield at seated ear level and both channels swept at the same time there is a decently flattish stretch from 200 Hz to 3 kHz. The upper limit of the speaker is around 16 kHz, with steeply falling output beyond. We have no mathematical model for the CLS geometry and are willing to be very forgiving as regards these measurements-who knows, it may all add up to reasonably correct power response. Some of the response peculiarities may be due to deliberate "voicing" by Gayle.

In the absence of a crossover and with all points on the diaphragm driven in phase, one would expect coherent reproduction of pulses and one gets it after a fashion, but the waveforms are not clean-too much ringing. Tone bursts confirm the latter conclusion, revealing modes and nodes all over the place, with some clean patches in between. There just ain't no such thing as a large, well damped plastic sheet, no how.

The remarkable thing is that this hodgepodge of so-so physical characteristics results in quite beautiful sound. Not at all volume levels, mind you; the CLS is definitely not a large-signal transducer. Nor is it a pip-squeak; let us call it medium-signal. It is not very happy with complex passages played fff. At normal, room-filling levels, however, it sounds magnificently transparent, with a warm, lush character, especially rich in the lower midrange and suggesting just a trace of electrostatic "sh" coloration. It is a somewhat formatted sound, not really accurate on an I/O basis, but exceedingly pleasant and exactly what many music lovers are looking for. If this is what Gayle Sanders likes, we are certainly not going to argue with him. The addition of a subwoofer, crossed over in the neighborhood of 100 Hz, would probably add significantly to the large-signal capability of the CLS but take away from the Bauhaus purity of the design concept.

The impedance curve of the speaker starts at around 7 ohms on the bottom, rises to 24 ohms at 250 Hz, drops a little, then skyrockets to 46 ohms at 2 kHz (equalization network?), after which the expected capacitive drop-off takes over, all the way down to 2 ohms at the highest frequencies.

It is not a particularly difficult load for an amplifier, and the sensitivity of 85 dB SPL at 1 meter with 2.83 volts input leaves the user with a good many amplifier options.

Where does the Martin-Logan CLS rank in this survey and how does it fit into our current loudspeaker pantheon? We cannot rate it quite as high as the Carver, which is just more speaker for less money. The Apogees are in effect visitors from outer space, leaving the Dahlquist as the one to beat for runner-up. Our emotional choice would be the CLS, although not as accurate as a monitor.

--------

[adapted from TAC, Issue No. 11 Winter/Spring 1988]

---------

Also see:

The Wire and Cable Scene: Facts, Fictions, and Frauds Part 1

Phono Paradise Regained, or Sao Win Rides Again: Win FET-10; Win SEC-10

Cartridge, Arm and Turntable vs. the Groove: Who's Winning? [1977]

Various audio and high-fidelity magazines