Trying to build dipole sub - a la Celestion 6000

The reason you don't see much of this is because there's not many crazy enough to try. "Dipole subwoofer" is about as close to an audio oxymoron as an unbiased opinion.

NgJockey,

Interesting point you raise there...

I am amazed at the results the Celstion 6000 subs have been able to produce.

I'll have to look into the links provided above to research this better.

I built Siegfried Linkwitz's Orion design, have run them in three different rooms so far, and have heard a handful of other Orion setups and other dipole bass/sub-bass setups in different places. They work great.

I've also heard Martin Logan's attempts to integrate monopole woofers with dipole main panels and that doesn't work well.

Ngjockey writes
>The reason you don't see much of this is because there's not many crazy enough to try. "Dipole subwoofer" is about as close to an audio oxymoron as an unbiased opinion.

As long as we're being relaxed about defining "sub woofer" like the original poster that's extending the headroom and extension of his electrostatic main panels or the consumer market definition which extends to 80Hz or beyond:

Compared to monopoles there are significant perceptual and measurable differences in how dipole bass couples to room modes (with adjustments possible through source rotation), has no gain below the space's fundamental resonance (why you can't put "big" conventional speakers in small rooms), has a different energy time curve, and has a more accurate modulation transfer function resulting from the speaker/room interface.

People don't do it because it's expensive. In a domestically friendly 14" deep enclosure you need four times the displacement of monopoles for the same SPL at 40Hz and 8X at 20Hz. For music the 40+ Hz output is the issue since even "bass heavy" music has last octave peaks 10-20dB down from the rest of the spectrum with the IEC musical power spectrum approximation specifying a second order high-pass at 40Hz. For mixed home theater use you also need a separate monopole sub-bass system for the lowest frequencies and probably LFE unless you take the Monte Kay approach using 24 15" drivers split into two 9' tall W-frame stacks with a total 1400 pound weight (That does yield 103dB @ 2.83V / 1 meter sensitivity at 18Hz with a 1% second harmonic and 1.5% third harmonic).

That's not very viable commercially in a world where designers lament that $8000 per pair MSRPs only allow for $80 midrange drivers and consumers would need to buy both dipole woofers and monopole sub-woofers for home theater.

Drew,

Thanks for sharing your experience with the Linkwitz Orion.

I like the effects I heard with the Celestion but I'm not confident I can reproduce the same results by just mocking up the design. I rather play it safe and go with a tested design.

I can't tell how significant it is the Celestion Subs have their drivers facing each outer is. I don't know how much of a factor it is in creating that dramatic mid-bass cleanness and punch. I figure I should be able to replicate the punch if I also use 4x12" drivers and perhaps use one of the designs that don't have them firing at opposing ends.

The physics of having drivers firing directly at each other seems complicated to me. Whether they will cancel, reinforce or interfere with frequency output. I can't tell if the gap distance in that configuration matters. I guess, I can experiment by building a rig to have two woofers facing each other and play various tunes through it as I manipulate the gap separation.

My room is only 22'x24'x8' and it is pretty well damped as it's right next to my master bedroom and I had to insulated well to allow my wife to sleep during my late listening sessions. But I can't imagine having 24 15" bass drivers firing off in my music room. No amount of sound insulation is going to hold that back :-)

I think I looking for more dynamic impact and fullness in the sound than just deep seismic bass. I'm a jazz/pop/light rock kinda guy, so I'm not a total bass freak.

The Sanders Sound panels covers a wide freq range pretty well on it's own. It covers 173hz-> over 20Khz. So I'm really just looking for the bass rig to cover 20hz->173. I would like to find dipole drivers that can serve this range well and I am concern about the ringing of the woofers as it will be box-less design. It will have to be something that provides servo/braking to keep the bass notes true.

>I like the effects I heard with the Celestion but I'm not confident I can reproduce the same results by just mocking up the design. I rather play it safe and go with a tested design.

You can do better. For instance mounting one woofer magnet in and the other magnet out results in a ~15dB reduction in even order harmonic distortion.

If you lack the gear (calibrated measurement microphone, USB interface microphone preamp, software like ARTA) capabilities or motivation (take what you built to an open parking lot and make ground plane measurements) to measure and adjust I'd build a well cooked Linkwitz design, probably one of the H-frame options (Seas drivers per Orion 3.4 with purportedly better large signal performance) since you'll avoid a cavity resonance near your passband. Even with the gear I'd be inclined to share Siegfried Linkwitz's or John Krevosky's driver choices because they've taken the opportunity to listen to and measure a number of drivers' performance in open baffle installations.

>I am concern about the ringing of the woofers as it will be box-less design. It will have to be something that provides servo/braking to keep the bass notes true.

I don't know what your technical concern is here.

People (the Carver Amazing dipole comes to mind) like to use high-Q woofers to compensate for the 6dB/octave dipole roll-off although drivers are minimum-phase devices so where you have flat amplitude you can't have time domain problems.

With a box you get inductance from the air spring in the box in parallel with the suspension compliance which increases both the resonant frequency and Q. Without you just have the driver Thiele-Small parameters and needn't have higher Q.

For sub-woofers the metal cone breakup is so far out of the driver's passband it doesn't matter.

Aerodynamic noises are a bigger issue since you no longer have a box to attenuate them. Pole or cone vents are a fine idea. Voice coil tinsel leads can make noise too but that's at a low enough level that although audible with sine wave test signals it will be masked with music.