SWELL CELLS

above: Laurence Dickie, inventor of the Matrix enclosure.

Last month, in Audio's 40th anniversary issue, I pointed out that in the early days of hi-fi, some of us built massive loudspeaker enclosures with sand-filled panels that weighed more than 600 pounds. Others constructed speaker baffles with bricks and mortar! All of these extreme measures were undertaken to sup press and attenuate the boomy resonances produced by large, undamped enclosure panels. Back in those days of monophonic sound reproduction, in spite of our relatively primitive equipment, we could easily perceive that enclosure resonances greatly degraded music signals. As the hi-fi business grew, complete speaker systems with furniture finishes were introduced.

However, commercial constraints with respect to size, weight, and cost precluded any "brute force" anti-resonant construction. Thus, for many years now, commercially available speaker systems have suffered in varying degrees from the omnipresent coloration of music signals caused by enclosure resonances.

Are the deleterious effects of cabinet resonances really that much of a problem? After all, countless numbers of people listen to speaker systems that are rife with resonances, and apparently they are adequately pleased with what they hear. In fact, sad to say, many people equate the boomy resonances of their speaker systems with good bass response. Well, to each his own. (Some people prefer margarine to butter.) I suppose one could trot out handy clichés, such as "ignorance is bliss" or "what you don't know won't hurt you," but that would be unkind.

Education is a more reasonable approach. To the trained ears of a person who listens to live music, the superimposition of resonant colorations on his audio component system is anathema.

He is keenly aware that these are un natural artifacts, phenomena that do not exist at a live concert.

To avoid enclosure resonances, many people use electrostatic or planar speakers, which do not employ en closures. However, these dipole de signs trade one set of problems for another, which can include imaging difficulties, an inability to achieve high playback levels, and restricted low-frequency response.

In the monophonic era, our main concern was the amplitude of the pan el resonances which caused that un wanted boominess. With the arrival of stereophonic sound, the slow decay of panel resonances was seen to exacerbate the problem of enclosure resonance effects. This lengthened decay had always been present, of course.

With stereo reproduction, however, the differing arrival times of the direct sound from the drivers and from the panel radiation caused a "smearing" and loss of image specificity, which degraded the three-dimensional qualities of the stereo sound.

In recent years, studies have revealed the magnitude of enclosure resonances. Consider that in many typical loudspeaker systems, the total surface area of the enclosure panels may be more than 30 times the area of the driver diaphragms. In some poorly damped enclosures, panel radiation is down only 6 dB in relation to the direct radiation from the drivers. In fact, it is possible at certain discrete frequencies for panel radiation to exceed the driver output! A further problem is that panel resonances commonly occur in the range from 70 Hz to 1 kHz, which unfortunately coincides with perhaps the very most active area of the music spectrum.

Using a laser interferometer and improved accelerometers, B & W, the loudspeaker manufacturer, was able to obtain some important new data on enclosure resonances. They found that panel radiation at low frequencies is influenced by panel stiffness. High frequencies are influenced by panel mass, and intermediate frequencies are influenced by panel damping. In the time domain, it was found that the desirable rapid decay of enclosure resonances demands high damping and/or low mass.

Further, conflicting requirements in the frequency and time domains with respect to the suppression of panel resonance could not be satisfied with the typical wood particleboard enclosure. Thus, recent investigations focused on new materials for the construction of speaker enclosures.

Among these are various laminates, some using graphite or boron fibers (which are very expensive). Another is a so-called aerospace material known as Aerolam, which consists of two thin aluminum "skins" placed on either side of a core honeycomb-structure damping material. There have also been some Scandinavian speaker enclosures that return to the "brute force" philosophy of cast concrete! Tests by B & W clearly showed that the typical particleboard enclosure, internally damped with bitumen, contributed too much panel radiation. The Aerolam enclosure had significantly less panel radiation at lower frequencies, but unfortunately it also had a higher degree of coloration in the midrange. Even the concrete enclosures, which one might presume to be free of resonances, were found to have a pronounced ringing at 350 Hz caused by lack of internal damping in the concrete.

After these tests, the search for new anti-resonant construction materials didn't look very promising. Then Laurence Dickie, chief electronics engineer of B & W, came up with the Matrix anti-resonant loudspeaker enclosure. I de scribed the internal honeycomb structure used in the B & W Matrix 1, Matrix 2, and Matrix 3 loudspeakers in the September 1986 issue.

As I pointed out in my original description of the Matrix, it is a structure made of a very rigid proprietary material with high damping qualities. The Matrix is composed of a series of inter locked, perforated pieces that form a cellular honeycomb. The cell ends have a very high degree of stiffness.

The planes of the cell ends are fitted to corresponding grooves on the inner walls of the particleboard enclosure panels and then bonded in place. The many cells of the honeycomb are filled with an acoustic foam. This concentration of foam furnishes an almost anechoic condition in which the out-off-phase energy from the rear of the drivers is almost totally absorbed, having somewhat the same effect as increasing the enclosure's internal volume.

With the Matrix bonded to the interior of the enclosure, there is continuous structural integrity, with the enclosure becoming monolithic in its solidity. Although the Matrix structure does not weigh very much, it is rigid enough to support considerable weight.

Since I have been an advocate of anti-resonant speaker construction for many years, I was interested in this new Matrix development, and I asked B & W to send me detailed technical information about it. They responded by very kindly inviting me to visit their Steyning Research Center in England.

The Steyning Research Center is a very busy place, with a staff of 25 engineers and scientists headed by re search director Peter Fryer. While I was at Steyning, I had the pleasure of meeting Laurence Dickie. He told me his inspiration for the Matrix concept had come from idly staring at an open case of wine, with the bottles nestled in the cells of a protective cardboard honeycomb insert.

Made of a proprietary material with high damping qualities, these cells give the Matrix enclosures their anti-resonant rigidity.

In the laser lab at the Steyning Research Center, Dr. John Dibb (left) and Dr. Peter Fryer, the center's research director, demonstrate computerized modal analysis techniques.

The labs at Steyning are filled with advanced and esoteric equipment, much of it used in the development of the Matrix concept to measure speaker enclosures, drivers, and other components. Point accelerometers were particularly useful in measuring the amplitude of vibrations at various points on the enclosure--panel surfaces. The computer-processed accelerometer data was used for measuring the time effects of impulse excitations applied to the enclosure, among other tests.

Point measurements from the enclosure surface were also subjected to computerized modal analysis. As in the past, B & W made extensive use of their Doppler laser interferometer. This is a spot-measuring device, where the surface to be analyzed is scanned sequentially, and then the data is stored and computer-processed to yield a composite picture. The system, unfortunately, cannot be used in real time.

B & W has also put into operation an entirely new type of laser measurement tool known as an Electronic Speckle Pattern Interferometer. The ESPI consists of an optical unit with a TV cam era, an electronic processing unit, and a TV monitor, all of which is mounted on an air-suspension optical table. The ESPI employs a 10-mW helium-neon laser light to illuminate the test object.

This produces a speckle pattern, quite similar to a grainy photograph, when the object is viewed through a lens.

The ESPI's big advantage is that it operates in real time. The unit's continuous interferogram of the resonance patterns is updated 25 times per second. Sine-wave frequencies can be applied up and down the spectrum, and the change in speckle patterns can be viewed continuously and instantly on the TV monitor. The fruits of this new ESPI technique include the Matrix speaker drivers, in which the driver basket and mounting plate are made of a single casting of magnesium alloy. The rear of these driver castings is treated with a polyurethane elastomer damping compound.

B & W is also doing a great deal of sophisticated computer modeling.

Data derived from measurements and other sources, along with desired parameters, is fed into a computer whose special processing can predict the performance of a hypothetical speaker, so building an actual speaker is not necessary. I sat down at a computer terminal with the affable Dr. Fryer, and we fed into the computer some "wishful thinking" parameters that would result in an idealized Model 801 loud speaker. We derived a version with some very desirable characteristics, but as I will relate, it was trumped by an even better design.

There were many other new ideas and concepts and possible future products all deriving from B & W's sophisticated research programs. There are far too many to detail here, but I would like to tell you about several of the most advanced B & W products which will debut at the Summer CES in Chicago. First is the B & W Matrix Mini System. This little speaker measures 9 3/8 in. high by 6 3/8 in. wide by 8 5/16 in. deep and has an internal volume of 6 liters. The cabinet, molded in one piece from glass-fiber filled polyester, employs Matrix construction. The bass midrange unit is a 126-mm Kevlar driver. The tweeter is a newly developed, 26-mm metal dome unit whose diaphragm weighs just 0.33 gram; B & W says it is down only 6 dB at 40 kHz.

The tiny, 11-pound Mini has a rated frequency response of 65 Hz to 40 kHz and a sensitivity of 85 dB, and it can handle amplifiers of up to 200 watts. If you get carried away, there is B & W's Audio-Powered Overload Circuit (APOC). I heard this little giant, and it is a classic example of getting 10-gallon performance out of a pint pot! To gild the lily, there is a companion Matrix Mini Tower subwoofer. It stands 39 inches high, is 6 3/8 inches wide, and has a depth of slightly over 8 inches. It has full Matrix interior construction and uses two 130-mm bass drivers. Operating as a fourth-order vented system with an internal volume of 14 liters, it has a rated-6 dB point of 33 Hz.

Used together, the Matrix Mini and Matrix Tower are said to provide response from 33 Hz to 40 kHz at an SPL of 107 dB! Because of the extreme rigidity of this system and its lack of resonance, the sound is ultra-clean, and as for stereo imaging, the speakers simply disappear! The most choice B & W item that will debut at the SCES is the Matrix 801.

This totally updated unit now employs matrix construction with a new bass driver, and the fibrecrete (fiberglass-and steel-reinforced concrete) head sports the new metal dome tweeter. A new sixth-order Butterworth vented system provides -- would you believe? -- a -3 dB point of 19 Hz! Rated sensitivity is up to 87 dB at 1 meter for 1 watt input, and overall frequency response is rated as 20 Hz to 20 kHz, -±.2 dB. With resonances controlled, the Matrix 801 loudspeaker should provide super stereo.

(adapted from Audio magazine, June 1987; Bert Whyte)

= = = =