Ran one for some years with much happiness.
Great concept. Keeps the driver resonance at it's natural free-air point (duh) and that is likely be low enough to add a welcome whomp to the lowest bass recordings. Certainly as compared to the little boxes now in style.
Don't believe the grade-school geometry theory you hear. In a room, with some modest couple of feet of driver baffling, the rear wave isn't sneaking around and annihilating the front wave below some defined point. Not like that much. But when the baffle is small, you need to introduce a Linkwitz boost and that can be too much boost for a driver to handle without a sharp (DSP) rumble filter.
No question boxes are bad for sound quality. There would be better sound in this world of people removed the backs of their boxes. For sure, nobody would say they are any more than necessary evils and not inherently beneficial, horns aside, even if they are the best option you have. Think about it.
Ben
Great concept. Keeps the driver resonance at it's natural free-air point (duh) and that is likely be low enough to add a welcome whomp to the lowest bass recordings. Certainly as compared to the little boxes now in style.
Don't believe the grade-school geometry theory you hear. In a room, with some modest couple of feet of driver baffling, the rear wave isn't sneaking around and annihilating the front wave below some defined point. Not like that much. But when the baffle is small, you need to introduce a Linkwitz boost and that can be too much boost for a driver to handle without a sharp (DSP) rumble filter.
No question boxes are bad for sound quality. There would be better sound in this world of people removed the backs of their boxes. For sure, nobody would say they are any more than necessary evils and not inherently beneficial, horns aside, even if they are the best option you have. Think about it.
Ben
Last edited:
The magic is, IMO, in the dipole radiation pattern. Room interactions are minimal, so whatever the speaker's putting out, that's what you'll hear.
I tried a pair of nice 15" drivers (Beyma 15P1200Nd) on open baffles, and had to apply 30dB of EQ to get flat to 25Hz. Without that EQ, they sounded completely gutless. And I mean completely. Sorry to contradict you, Ben, but dipole losses are very real and need serious attention, power, and driver displacement to get useful low end.
To give you an idea, at fairly moderate listening levels at 4m distance, the drivers were getting out to 30mm p/p travel, and I was hitting each one with around 300w peaks. One driver in a tiny sealed box easily provides more output.
The sound was, actually, very good. That room has some problems at 40Hz which these dipole speakers effectively bypassed.
However, if you take the afore-mentioned sealed sub and apply some EQ to compensate for the room problems, you also get very good bass, with the addition of the heaps of headroom that using 15" drivers in a domestic setting will get you.
One very positive thing with dipoles is that they are small. They need moving into the room a little, but 25Hz from a pair of flat boards is still quite impressive.
Chris
I tried a pair of nice 15" drivers (Beyma 15P1200Nd) on open baffles, and had to apply 30dB of EQ to get flat to 25Hz. Without that EQ, they sounded completely gutless. And I mean completely. Sorry to contradict you, Ben, but dipole losses are very real and need serious attention, power, and driver displacement to get useful low end.
To give you an idea, at fairly moderate listening levels at 4m distance, the drivers were getting out to 30mm p/p travel, and I was hitting each one with around 300w peaks. One driver in a tiny sealed box easily provides more output.
The sound was, actually, very good. That room has some problems at 40Hz which these dipole speakers effectively bypassed.
However, if you take the afore-mentioned sealed sub and apply some EQ to compensate for the room problems, you also get very good bass, with the addition of the heaps of headroom that using 15" drivers in a domestic setting will get you.
One very positive thing with dipoles is that they are small. They need moving into the room a little, but 25Hz from a pair of flat boards is still quite impressive.
Chris
If you use a wrong driver og very small baffle, then you will need heavy eq. If you use big baffle and proper driver you don't need eq (or just very little).
Less room interaction! - you can obtain that as well with several small subwoofers, rather than one or two. Dipole radioation is just one method.
I still like open baffle. It's simple to build and I benefit from less room interaction for low frequencies. I don't like to have four or more smal subwoofers, but it would probably be a good solution also.
On a small baffle i don't like the dipole bass very much. No matter how much eq you apply, something is missing. A large baffle or folded baffle is better in my opinion.
Less room interaction! - you can obtain that as well with several small subwoofers, rather than one or two. Dipole radioation is just one method.
I still like open baffle. It's simple to build and I benefit from less room interaction for low frequencies. I don't like to have four or more smal subwoofers, but it would probably be a good solution also.
On a small baffle i don't like the dipole bass very much. No matter how much eq you apply, something is missing. A large baffle or folded baffle is better in my opinion.
How big was the baffle? Did it have "modest couple of feet of driver baffling"?
For sure, nobody is contesting the notion of out of phase wave cancellation or the validity of your experience. I'm just saying that in a complex room environment, the cancellation is hardly like the textbook model would have you fear.
Just for fun (or laziness), I once just hung a 12-inch driver on a cord from the ceiling. I taped a large toroid coil of few mH to the back wired in series. Pleasant sound.
B.
From my own test with an OB sub a few years ago, I concluded that the alleged reduction in room interaction when OB subs were used was basically a myth. Measurements trump suppositions every time. Ever wonder why there are very few published measurements for OB subwoofers?
The Subwoofer DIY Page v1.1 - Projects : An INF10 Dipole Subwoofer
The Subwoofer DIY Page v1.1 - Projects : An INF10 Dipole Subwoofer
With OB you eliminate some resonances that are inherent with a driver in a box. Closed box speakers have a resonance corresponding to the internal dimensions of the box. Vented box speakers add potential port resonances, and have worse transient response. Both closed and vented boxes can re-radiate the driver's backwave back out through the cone after it reflects off of inner box surfaces.
OB systems have none of these resonance issues, and in general will sound less colored. This an advantage, and for me the many benefits of OB systems outweigh the disadvantages.
I recently built 4 subs. Two are monopole, with an extremely stiff enclosure (3/4" BB + 1" high density urethane foam) and two are dipole (each cabinet has dual opposed drivers coupled magnet-to-magnet in a folded open baffle, also 3/4" BB with the foam) All use the same 12" paper cone drivers and servo amps from Rythmik. Each pair has been EQ'd in the room with DSP to about the same curve. The monoploes need a little phase adjustment because they are in the corners behind the mains, while the dipoles sit out in the room at the same listening distance. BTW, the monopole and dipole cabinets are the same size (16" x 16" x 25") although they need to be placed differently in the room of course. Both the mono and dipole setups are crossed at 80Hz LR4 to dipole mains. The room is long and narrow, 15 x 40 with a 10 ft. vaulted ceiling. This is all to say I'm really comparing apples to apples here.
Both setups sound great: tight, musical, articulate, blah blah blah. They both get down to the 20's easily, of course the monopoles can go lower with less effort.
However the dipoles are something special. Deep bass at high SPL with no windows and dinner plates rattling in my rickety old house. You don't feel the bass hitting your body like a monopole, it's more like it's just there, enveloping you. The absence of "pressure" in the room is wonderful and even at dance-party levels you don't feel battered by the sound.
Around the crossover frequency, the dipoles again are the winners, integrating effortlessly with the mains. The monopoles, while really good, have just a touch of that "boxy" quality, I guess attributable to reflections coming out of the box through the cone, but also have that punch goes well with certain kinds of music. As far as "room interaction" goes, both setups have their peaks and dips and need EQ, preferably DSP, to sound their best.
Both setups sound great: tight, musical, articulate, blah blah blah. They both get down to the 20's easily, of course the monopoles can go lower with less effort.
However the dipoles are something special. Deep bass at high SPL with no windows and dinner plates rattling in my rickety old house. You don't feel the bass hitting your body like a monopole, it's more like it's just there, enveloping you. The absence of "pressure" in the room is wonderful and even at dance-party levels you don't feel battered by the sound.
Around the crossover frequency, the dipoles again are the winners, integrating effortlessly with the mains. The monopoles, while really good, have just a touch of that "boxy" quality, I guess attributable to reflections coming out of the box through the cone, but also have that punch goes well with certain kinds of music. As far as "room interaction" goes, both setups have their peaks and dips and need EQ, preferably DSP, to sound their best.
Don't know if I've ever read as coherent a home listening test as that one.
Like with horns, pretty much all the folks who use dipole sound love 'em, with shortcomings ignored*.
I know I am boring to repeat this, but it makes you wonder what it is that the sim advocates are missing? Do they see the "trees" but their precise-looking calculations fail entirely to understand what the "forest" is all about?
Ben
*I've been using large dipole electrostatic panels for the last 40 years, with various subs below 130 Hz. Funny thing, Mine are almost touching because the stereo ping-pong effect is too disturbing when separated more (an observation that surely befuddles those who think a monopole speaker must be essential for good stereo)
Like with horns, pretty much all the folks who use dipole sound love 'em, with shortcomings ignored*.
I know I am boring to repeat this, but it makes you wonder what it is that the sim advocates are missing? Do they see the "trees" but their precise-looking calculations fail entirely to understand what the "forest" is all about?
Ben
*I've been using large dipole electrostatic panels for the last 40 years, with various subs below 130 Hz. Funny thing, Mine are almost touching because the stereo ping-pong effect is too disturbing when separated more (an observation that surely befuddles those who think a monopole speaker must be essential for good stereo)
Last edited:
This is a strong function of how large the open baffle or dipole subwoofer is made.
I always like to harp on the fact that, on-axis, the SPL of a dipole is, at the "dipole peak" +6dB compared to the same driver in a closed box (of any size, e.g. the "passband" SPL level). Below the dipole response peak, the SPL will fall off at 6dB/oct from the dipole front-back interactions. When you approach and pass the driver resonance frequency the response will turn down another 12dB/oct, but this is not exactly true in practice (see below).
In order to make the most of a driver for a dipole subwoofer you MUST make the front-to-back distance large. How large? Well the ideal size is such that the dipole peak is just about at the crossover point. This translates into some pretty large sized "baffles" if you are using a planar baffle, e.g. 4 feet (1.25m) wide and tall for a 100Hz crossover point. But you can "fold" the baffle into an H-frame for example so that it is visually smaller.
With a large H-frame, you now have +6dB response (on-axis, compared to the passband level of closed box) at the peak. The driver's resonance point is about the same as its free air value. In between you have a down-sloping response. But with some simple response shaping you can flatten it out and such that the curve is flat to 30-35Hz before turning down. Just like with a CB sub with a LT, you will need some boost to do this, e.g. 12dB.
Drivers for dipole subs can be of medium to high Qts (e.g. 0.4-1.5). In a long H-frame there will be some increase in the Q value because of increased air loading of the driver. With lower Qts drivers the dipole slope will more slowly turn down from 6dB/oct (due to dipole cancellation) to an effective 12dB/oct (above resonance) and may get steeper but I would not attempt to use the driver below its resonance because the power requirements go up significantly. There are some "OB specific" subwoofer drivers with higher Qts (e.g. 0.6-0.8) that will help to reduce the power needed near resonance. There are very few high Qts (Qts>1.0) drivers that make good "subwoofers" due to limited Xmax, or other reasons.
As several posters mentioned (including myself) back in 2016 when this thread was started, the big advantage of a dipole sub is the lack of cabinet resonances and the sound is often thought of as more "open" and "uncolored". But like anything this depends on the specific design and implementation and not every OB/dipole will have much of any deep bass unless it is designed with that in mind.
One interesting approach, if you have the space for it, is the SLOB (slot loaded open baffle). This is a large (e.g. 2m high, 2-3m wide) planar baffle with a plenum running top to bottom, like a vertical "trough". An even multiple of drivers is mounted on each side of the "trough" in opposition. The driver cones move either towards or away from each other and this tends to cancel most vibration. The large baffle provides large D, and by using multiple subwoofers the total displacement can be respectable. The one issue with this design is that the trough forms a cavity that resonates and the deeper and narrower the cavity the stronger the resonance (the higher its Q). Otherwise, while being huge but still needing to be "away" from a wall, this type of dipole sub can provide copious bass.
There is no real reason why the SLOB can't be made with a very "wide" slot. In fact, just make the slot about 1/3rd of the total width. This would make the Q of the resonance of the trough very low such that it would no longer be a issue. For all those big planar OB fans, you could stick one of those fancy and expensive Voaxtive drivers on each outer panel and you will have a single-panel-stereo FAST/WOW system with some REAL bass! I would love to see (and hear) something along those lines! It would be very easy to build, too. Some decoupling of the fullrange drivers from the actual baffle would help to prevent the subs from interacting with the fullrangers. But you would need a place to "Deploy" such a system that would need something that is effectively a "room privacy screen". Still, that would be a fun system!
By keeping the cross section of the H-frame larger than the driver diameter and no more than 0.25*D (D=total length) you will keep the line resonance Q low.
Can you point to journal articles, links (probably Linkwitz), etc. discussing the nature of the dipole peak as well as how to size an OB for a specific crossover point?
Linkwitz's page on Dipole Models gives good info:
Electro-acoustic models
Here is a figure from that page that shown the regions of the frequency response of a dipole that I mentioned in my post above:
This model is for a very small D, so it is not of a subwoofer, but the response features remain the same. It's just a model of two dipoles. This is not accurate as you get above 500Hz-1kHz but is a good predictor of response for low frequencies where sources behave like monopoles.
In section A-3 of that web page SL mentions some more accurate info about higher frequency responses that deviate from the two-monopole model and these are similar to what I have measured for some real-world systems around 1kHz and above. They tend to look more like this:
A more realistic looking response that you might get for a 5" midwoofer when used "nude".
But even this more realistic model does not capture things like resonances behind the frame and the blocking of the rear output by the magnet, and these tend to muck up the response above 3kHz.
I can't point to articles on crossover point choice. But for a subwoofer, I do not see a compelling reason why one should not make the crossover point equal to the dipole peak. Above the dipole peak there will be (for a sub) one or more notches on axis, so you cannot really operate that high. Below the dipole peak the response drops away at 6dB/octave, so using the sub in a lower frequency region well below the dipole peak throws away this higher SPL region, and you get low sensitivity. By making it as large as possible for a given crossover point you maximize the efficiency of the system. You do that by building the dipole large enough such that the dipole peak is at the crossover point.
The image above assumes the sources have infinite low frequency extension. You need to add on top of this the LF roloff of the driver itself, which will be to a first order approximation like the free-air response that you can predict from the TS parameters. This causes additional LF droop in the response. When the Qts of the drive is relatively low, e.g. 0.3-0.5, this additional rolloff of the driver's own response makes the effective LF rolloff more like 12dB/oct.
Last edited:
I was interested in dipole woofer, made some experiments (me and my friend built H-frames, M-frames, flat open baffles, U-frames, sealed and vented woofers). Tried drivers with different Q from ~0.25 to 1+ (I don't think low Q is a big problem for OB). Until recent time I didn't realized that with each 6dB loss you need 4 times (not 2 times) more power in order to compensate for that. And therefore dipole's 6dB/oct drop is something really important, you cannot just ignore it. So regarding power demands and cone excursions I can summarize everything with a picture below (frequencies given for typical average H-frame and sealed woofer sizes like 50-60cm, they will differ from size to size, but to change the chart dramatically you'll need to change the woofer size dramatically too).
Of course with dipole you can experiment with placement and get different results, it doesn't sound boxy etc but with monopole it is much easier to produce loud bass sound, it takes less power, less cone excursion. In other words you only turn volume knob slightly for a vented woofer and you get your desired loudness, with an H-frame you'll turn the knob to the max and still sometimes this won't be enough.
Of course we tried dipole woofer equalization but equalization cannot save from power demands and cone movement limitations, right? Not to mention mechanical forces and extra vibrations due to powerful cone movement.
If we'll look at Linkwitz best system, he has FOUR woofer drivers with great cone excursion abilities (you can tell that even just by their appearance) and that's how he got "enough" low frequency loudness (in my opinion).
So my advice to anyone making dipole woofers, if you don't want to spend money for 4x more drivers, go for sealed (or better vented) box. If you are in a quiet environment and in a small room, only then dipole woofer with single driver may be ok.
Nevertheless I personally prefer H-frames and I like the idea of stacking them (you can make one, or pair, then later another one or pair and so on). But again, that's not the cheapest way to get loud bass since drivers cost money and you'll need more amplifiers too.
Of course with dipole you can experiment with placement and get different results, it doesn't sound boxy etc but with monopole it is much easier to produce loud bass sound, it takes less power, less cone excursion. In other words you only turn volume knob slightly for a vented woofer and you get your desired loudness, with an H-frame you'll turn the knob to the max and still sometimes this won't be enough.
Of course we tried dipole woofer equalization but equalization cannot save from power demands and cone movement limitations, right? Not to mention mechanical forces and extra vibrations due to powerful cone movement.
If we'll look at Linkwitz best system, he has FOUR woofer drivers with great cone excursion abilities (you can tell that even just by their appearance) and that's how he got "enough" low frequency loudness (in my opinion).
So my advice to anyone making dipole woofers, if you don't want to spend money for 4x more drivers, go for sealed (or better vented) box. If you are in a quiet environment and in a small room, only then dipole woofer with single driver may be ok.
Nevertheless I personally prefer H-frames and I like the idea of stacking them (you can make one, or pair, then later another one or pair and so on). But again, that's not the cheapest way to get loud bass since drivers cost money and you'll need more amplifiers too.