Testing the Cabinet, Measurements, Conclusions

By mark — July 04, 2005

Test Cabinet Construction
Having assessed the effectiveness of the free-layer damping approach when applied to test samples, it was time to build the test enclosures.

Each enclosure was designed to have 4 solid cross braces, oriented horizontally, fixed to the interior side of the cabinet's side & back panels, with the front edge, close to, but not touching the inside of the faceplate. An FEA simulation done prior to construction indicated 4 cross braces would produce acceptable results.

To gauge the effectiveness of the bracing & damping approaches chosen, a series of measurements were made at different points in the construction process using the ACH-01 accelerometer, along with a calibrated microphone for capturing acoustic measurements, useful in subsequent analysis.

Figure 8

The enclosure sported double-thickness (38mm) front, top & bottom panels, with the sides & back being single thickness (19mm) panels. Figure 8 shows one of the enclosures, midway through construction. In addition to the 4 horizontal, shelf-style braces, the front panel was cross-braced as well, with a thin shelf style 5th brace running from the inside center point of the faceplate to the back of the enclosure. The use of extensive bracing and damping added to the already substantial weight of the cabinets. The completed enclosures weighed in at over 100 kg (220 lbs) each!

In Figure 9 can be seen an heroically busy collection of plots, each sweep indicating the vibrational behavior of the panel to which the accelerometer had been affixed (the center of the left or top panel in Figure 8) at various stages of construction. The net before/after result of all of the bracing, damping & stuffing can be seen in Figure 10.

Figure 9

In Figure 10, a general downward shift in amplitude, accompanied by an upward shift in frequency is noted when comparing the "after" curve with the "before" curve. Keep in mind these are not dB SPL curves, but the relative response of an accelerometer responding to the cabinet panel's acceleration at each frequency along the sweep.

Figure 10

Figure 11 is the "before" and "after" vibrational response to impulse stimulation. (In this particular graph, "before" is the response plotted on the right, the "after" plotted on the left). The improvement in settling time is striking.

Figure 11

Thus far, we've seen what the accelerometer has had to say. To see what the microphone had to say, I employed the noble and familiar knuckle-rap test. I removed the accelerometer, rapped away at the spot formerly occupied by it, with the microphone positioned nearby. In Figure 12, the purple "before" curve presents a generally higher relative amplitude response than the green "after" curve. The difference, however, sounded much greater than it appears in this graph.

Figure 12 Spectral Analysis, Before/After Knuckle-rap test, Cabinet 2

Note: Figures 13 & 14 are the time-domain versions of the data presented in Figure 12.

Figure 13 "Before" Knuckle-rap test, Cabinet 2

Figure 14 "After" Knuckle-rap test, Cabinet 2

Subjective Assessment & Closing Comments

Measurement indicated that the various measures taken to reduce panel vibration were indeed effective in lowering the mechanical noise floor of the system. However, objective assessment tells only part of the story and a subjective assessment is required to complete a picture of any system's performance.

In Figs. 15 (right) we see the test system, at an earlier stage of construction, awaiting completion of the other modules making up the architecture of the full system.

Though the cabinets are attractive in their unfinished state, with an appearance reminiscent of, for example, some of Axiom's current line of products, they will eventually have a finish applied. The choice of finish, though, is left to my wife - an effective approach for increasing WAF or SAF. Now there's a tip for DIY Audioholics everywhere!

Two listening tests were conducted. Each lasted about an hour. A variety of material with which I and the 4 other listeners who had taken the time to audition the system with me were very familiar with was used. The upstream test gear used in both instances was the same (Hafler DH-500 power amp, etc).

At the conclusion of both tests, results were tabulated. From the raw data collected, I threw out everything people thought they heard. Owing to the very informal, very unscientific nature of the test trials, I was interested in certainties only. I also threw out those items for which no consensus could be reached. In the end a few things did emerge.

First, resolution of low-level detail was dramatically improved. This was not some at-the-fringe, barely discernable effect; it was a substantial improvement. On this point, all listeners were unanimous. I heard things in recordings I have listened to for decades that I had not heard before.

Second, overall mid-bass clarity improved, resulting in, among other things, improved sound staging. Individual instruments, such as cellos in classical recordings were now discernible as individual instruments, where before they tended to sound more like massed string choruses, made up of not clearly identifiable individuals. The tonal differences between different pianos, such as Steinway, Yamaha, Blüthner, etc was now clearly discernible

It appeared that the "before" mechanical noise floor had, in addition to obscuring the low-level mid-bass detail beneath the murk of mechanical noise, imparted a euphonic "sameness" to everything played through the system. Now that "sameness" had been largely, though not entirely, eliminated. Overall, the system's presentation had become more realistic & musical and that resulted in a more satisfying listening experience. Not bad when considering how old a technology the KEF B139 drivers used represent.

About the drivers …
I first became acquainted with the KEF B139 drivers back in the late 1970's. In those days I was working in a recording studio "programming" a Moog synthesizer. ("Programming" back then meant plugging the Moog's patch panels full of patch cords). I remember carrying around a 3-ring binder full of preferred patches. Ancient history!

The synthesizer fed its output, via a Dynaco power amp, into a Pair of IMF transmission line loudspeakers that used the B-139 as the woofer. The bass those things could crank! My teenage ears were impressed.

Outside of live music situations, I had (up to that point) never heard such low bass at such volumes. (Remember, we're talking late 70s here - monster subs were not as commonly available as they are today). Years later I would see B139's used in Wilson Audio's $100k + US WAMM loudspeaker system.

Over the years I acquired half a dozen of the units for my driver collection, used mainly for research purposes. As lately they had just been lying around gathering dust, when it came time to build a general-use system for the house, it was an easy choice to design the mid-bass modules for the system around them.

Though technology has certainly provided for many substantial advances in driver quality since the B139s were first introduced over 30 years ago, there are some things they do well, to this day.

Figure 1 7 B139 Spec sheet, ca. 1980's

Resources

2. McIntire, M. E., et. al.: " The Influence of Geometry on Linear Damping ", Acustica, Vol. 30, #4, 1978

3. Caldersmith, Graham: " Vibration Theory and Wood Properties ", Journal of the Catgut Acoustical Society, #42, November, 1984

4. No Author listed " Measuring Vibration ", Brüel & Kjær, Nærum , Denmark , September, 1982

6. Brown, Alastair, W. M.: " The Ensemble Statistics of the Response of Structural Components with Uncertain Properties ", Doctoral Dissertation, University of Cambridge , March 2003

11. Iverson, J. K.: "The Theory of Loudspeaker cabinet Resonances", Journal of the Audio Engineering Society, Vol. 21, #3, April 1973

12. 11. Tappen, P. W.: "Loudspeaker Enclosure Walls", Journal of the Audio Engineering Society, Vol. 10, #3, July 1962

14. McIntire, M. E., et. al.: " On Measuring Wood Properties, Part I ", Journal of the Catgut Acoustical Society, #42, November, 1984

← Previous Page

Sections

Testing the Cabinet, Measurements, Conclusions