The Radial Loudspeaker Project
2001 by Steve Deckert
Well it's been three years since I wrote part I of this project and it's nice to finally have it completed. There were developmental problems during this period but the reason for the three year gap actually had nothing to do with the speaker. It was because just prior to writing part one I had launched our little Zen Triode amplifier. Time has been scarce ever since.
As I indicated in part one, the potential of this design was disturbingly high. I half expected (as I typically do) to pick the project up 6 months or a year later only to find I must have been stoned and that it actually sounds like crap. Well, not the case - in fact I've toyed with it very little since then and now that I've picked it up again the sound is just as impressive if not more so.
The developmental problems that did occur were specifically related to the 8 inch driver itself. I made 5 revisions of the driver over the 3 year period with hopes of finding something I could produce. The original driver, gone over in part one, was difficult to replicate. Revision A, Revision B & Revision C were all inferior to the original, lacking both low frequency extension, and in efficiency.
At one point earlier this year I made some profound progress in the design of the driver and was able to duplicate predictable results in several different frames. That means that should my parts source for the current model dry up, it will be painless to move on to a slightly different substitute.
This drastic design change addressed three issues - bass response, power handling, ease of manufacturing. Those issues would have been the main flaws of the original design. While the original design was warm in signature, there was little low bass to speak of suggesting the use of a subwoofer for fully satisfactory sound. Even though this is still a fragile driver, the original would easily damage itself with excessive power. I suspect that if the original driver had lower bass response it would have become even more delicate.
The change involved using two cones vs. just one. This is something I've dinked with since the original concept. The problem is the standing waves between the two cones and resulting frequency response. Apparently persistence pays off, because I got the idea to relieve the pressure (energy) standing between the two cones by ventilating the spacer that joins the cones together. I eventually found the correct hole diameter and pattern to do this without decoupling the cones. By that I mean I wanted the transient response to be communicated to the top cone without any delays.
By using two cones in this configuration it is possible to accurately couple low frequencies to the top cone. The top cone has a mass of less than 8 grams which is not strong enough when driven by a singular point to handle dynamic excursions. The Q-Tip that transfers the energy from the motor assembly to the top cone is still the main conduit between the two. The bottom cone creates a controlled air space under the top cone giving it terrific support across the lower octaves. This also creates a barrier between the top cone and the sound inside the enclosure commonly called "backwaves."
The current driver has an fs of around 35 Hz, but retains it's original stiffness. The efficiency ends up being around 90dB with a flat response to 2 kHz with a natural 6dB per octave taper out to 10 kHz. A tweeter is used like in the original design to offer a point source from 2 kHz on up.
In part one I got in depth about basket resonance's. The same type of resonance control is important in every component of the speaker, including the cone and the enclosure. In the current driver, the frame is cast with a heavy high quality motor assembly. It is relatively free of resonance's so no additional damping is used. The cones on the current driver have been tragically dampened whereas the original drivers were not. This is largely because of the increase in power handling offered by the new design. The speaker can now play relatively loudly bringing into play new modal resonance's in the top cone.
To damp these, since each driver has to be hand made anyway, I sweep it with a frequency generator and find the exact spots on the cone that ring. There are three inherent modes, one at around 600 Hz, one around 1200 Hz and one around 3 kHz. There is a tiny one at around 5kHz that is also addressed. The results of this damping creates a memorable pattern on the cone. Further testing and fine tuning IN THE ENCLOSURE necessitated one final phase of damping that short circuits ringing across the entire cone, turning the energy into heat. This was only necessary when pushing the speaker to the wood with larger amounts of power, but didn't hurt things at lower listening levels either.
The enclosure is unchanged from the original prototype in basic design. I spent the better part of 3 years in the past working with cylinder shaped enclosures so I didn't feel the need to try much else. The open bottom design is based on a heavy flange that refracts energy back up into the enclosure while allowing it to breath naturally just above the floor. Changing the surface from wood to carpet, as well as adjusting the height of the gap between the floor and the bottom of the cabinet are ways to effect both midrange signature and bass response.
To maintain a coherency between the driver and the tweeter some special considerations were established. These included small physical size so as not to create diffraction problems and extraordinary speed. I choose a small 3/8" balanced drive dome tweeter that uses a titanium spattered poly dome. Listening tests confirm this was the right approach.
No crossover is used on the driver, and only a single high quality poly cap is used on the tweeter. The overlap is seamlessly cancled between the two units acoustically by moving the output of the tweeter back in time 180 degrees. This is done by reversing the polarity of the tweeter before the capacitor.
The enclosure is constructed of a wood composite high in fiber and at a thickness of 0.25 inches the radius give similar strength and damping performance to a 24" x 24" inch panel over 3 inches thick. It is also specifically damped internally to further diminish standing waves in the mid band frequencies.
After completing the first production pair, solidly built and finished in real cherry wood and on our best Zen Triode Gear I have to say the final results not only lived up to the promising results of part one, but easily surpassed them. The biggest difference being the bass. It is wonderful, clean, fast and deep. A speaker that makes use of a subwoofer totally optional - probably a good thing since most subwoofers are slower than catsup and would never be coherent with such a fast speaker.
The imaging is far superior to anything I've heard, able to take a small room and bloom it out to a point of being almost ridiculous. As much as I like my Magnapans, these are in a different league. They have more spank, are faster, and far more forgiving of room placement and sitting position. They also have a touch more extension in the bass, a touch more of it, and are easily as quick and neutral.
So finally after all this time, we have the perfect speakers for a Zen Triode amp, not to mention lots of other lower powered amplifiers. They are so room friendly that holographic imaging is inevitable in almost any listening room.