Just to add my input. The cone excursion required for dipole bass is substantial, even with multiple 15" drivers driven into their X limit, which is not a preferred method as they are no longer linear and distortion rises. I checked a range of different 15's in WinISD and power handling was severely compromised. For some they could only manage 20 or so watts even when used with a 40Hz HP. Running them from 80-100Hz on a 4th order, the power handling increased to well beyond their power rating, many in excess of 1000watts to meet their Xmax. For those interested, and although not perfect, you can simulate open back by assuming a very large cabinet of say, 1000 litres!! But you also need to add a shelf filter that replicates the 6dB roll off cancellation as a result of baffle dimensions.
This natural roll off is of course critical and to preserve the actual crossover point, you need to EQ the dipole with a +6dB shelf. As a rule of thumb, if the dipole is sitting on the ground, the frequency equivalent to twice the diagonal of the baffle is a rough starting point.
With a baffle holding 2 x15's (3ft x 1.5ft ish), it starts rolling between 180-225Hz and this is where you need to start your shelf EQ. Clearly then you will be around +12dB at 50Hz. No wonder you have no power handling!!
To maintain an accurate 100Hz X/over without EQ correction you are looking at 6 x 15's in a baffle around 5ft x 3ft approximately. Yes, that is 12 x 15'S in total for stereo!
Another aspect which may have some relevance, and comment welcomed, the dipole bass increases it's GD as the frequency drops. A Transform box, or small cabinet monopole, even with the required EQ can remain in very low GD figures down to the lowest required frequency.
And as for combining monopoles and dipoles, I had a classic system in the 70's, stacked Quads, 2 x 15 monopole bass bin and Kelly ribbon tweeters stuck between the Quads. A TEAC 12dB/oct active x/over divided them, a sensible room some 30ft x 14ft and the sound was just wonderful with no "join" between the Quads and the bass bins which were some 12 ft behind them against a wall....
This natural roll off is of course critical and to preserve the actual crossover point, you need to EQ the dipole with a +6dB shelf. As a rule of thumb, if the dipole is sitting on the ground, the frequency equivalent to twice the diagonal of the baffle is a rough starting point.
With a baffle holding 2 x15's (3ft x 1.5ft ish), it starts rolling between 180-225Hz and this is where you need to start your shelf EQ. Clearly then you will be around +12dB at 50Hz. No wonder you have no power handling!!
To maintain an accurate 100Hz X/over without EQ correction you are looking at 6 x 15's in a baffle around 5ft x 3ft approximately. Yes, that is 12 x 15'S in total for stereo!
Another aspect which may have some relevance, and comment welcomed, the dipole bass increases it's GD as the frequency drops. A Transform box, or small cabinet monopole, even with the required EQ can remain in very low GD figures down to the lowest required frequency.
And as for combining monopoles and dipoles, I had a classic system in the 70's, stacked Quads, 2 x 15 monopole bass bin and Kelly ribbon tweeters stuck between the Quads. A TEAC 12dB/oct active x/over divided them, a sensible room some 30ft x 14ft and the sound was just wonderful with no "join" between the Quads and the bass bins which were some 12 ft behind them against a wall....
Last edited:
I couldnt state it better myself: "going boxed below the lowest room node" Post nr #7
I have a big room 10meter long. 15-27Hz is handeled with 4 custom made 15" JBL elements moving air between the listening room and outside "nature". I so to say have an Infinite baffel.
27-200 (will be changed to 27-150 Hz) is managed by 4 dayton IB15" woofers in 2*15" configuration subs, that is acting as stands for QUAD Esl63 speakers and obviusly they go from 200Hz to infinity...
well not really, but without metal grille and no dustcovers, bypassed electrolytic cap, + some other mods, they sound good.
Point is. Below the lowest room mode there is no point going dipole.
I have a big room 10meter long. 15-27Hz is handeled with 4 custom made 15" JBL elements moving air between the listening room and outside "nature". I so to say have an Infinite baffel.
27-200 (will be changed to 27-150 Hz) is managed by 4 dayton IB15" woofers in 2*15" configuration subs, that is acting as stands for QUAD Esl63 speakers and obviusly they go from 200Hz to infinity...
well not really, but without metal grille and no dustcovers, bypassed electrolytic cap, + some other mods, they sound good.Point is. Below the lowest room mode there is no point going dipole.
Dipoles are not very well modelled in software, especially at the low end. Experiments give true answers. I use two H frame boxes each with two 10 inch good drivers (SEAS) and amps (Icepower). 6-8 dB EQ is sufficient for most situations. 30Hz is -3dB and covers all of the music I listen to. Surprisingly kick-drum tests sound strong at 20Hz.
Group Delays acceptance thresholds at low frequencies are much bigger than the mid-high
2-3 ms. Research is sparse but Rod Elliot quotes 20-30 ms and even 50ms for 20Hz.
Group Delays acceptance thresholds at low frequencies are much bigger than the mid-high
2-3 ms. Research is sparse but Rod Elliot quotes 20-30 ms and even 50ms for 20Hz.
Not true.
At low frequencies you can model a dipole as a combination of two monopole point sources. The wavelengths involved are long enough for this to be accurate. For frequencies below about 200Hz (but even higher is still relatively accurate) you only need one nearfield measurement taken at one side of the dipole. The second "source" (e.g. the rear output) is modeled by inverting the polarity. Then you add a delay corresponding to the relative distance between front and rear sources and the listening position. When you are on axis in front of the dipole, this is the familiar "D" or separation distance. At other positions you need to take into account the relative distances using geometry. I created a modeling program that generates polar plots for dipoles in this fashion only using Excel.
Interesting how we can find and resurrect old threads
My last build used 2x 12" SB34NRX75-16 chassis on a very narrow baffle, with a calculated dipole peak, IIRC, at around 500Hz, and a theoretical needed dipole cancelation correction for 300Hz and below (+6dB at 150Hz, +12dB at 75Hz, +18dB at 37,5Hz). I applied this shelving low pass filter in the miniDSP and the output was an overwhelming bass. I eventually took out the whole shelving low pass and got "good bass" at my listening position.
Some considerations: we have quite a "livid" room: lots of glass without curtains, concrete building. Also the woofers have a highish Qts of 0.7.
SB Acoustics :: 12” SB34NRX75-16
In any case I still do not quite understand why so little (or no) equalisation was needed in my case, but this is certainly not explained by most of the current models. The only explanation I found was this Dipole on axis
I am actually quite happy with this result and built a smaller "nude" design using one of the above woofers per channel and still happy with the overall SPL output.
Siegfried Linkwitz ((RIP) pioneered dipole acoustic modeling. But when it came to construction he experimented. See his LX521. His website is a great source of knowledge. I find the rolloff problems with dipole bass less severe than software models would suggest. The models are accurate very close to the speakers, in the listening position the reality is different. That applies to dipole peak too. With a wide baffle you can get away with very little or no EQ of the peak.Interesting how we can find and resurrect old threads
My last build used 2x 12" SB34NRX75-16 chassis on a very narrow baffle, with a calculated dipole peak, IIRC, at around 500Hz, and a theoretical needed dipole cancelation correction for 300Hz and below (+6dB at 150Hz, +12dB at 75Hz, +18dB at 37,5Hz). I applied this shelving low pass filter in the miniDSP and the output was an overwhelming bass. I eventually took out the whole shelving low pass and got "good bass" at my listening position.
Some considerations: we have quite a "livid" room: lots of glass without curtains, concrete building. Also the woofers have a highish Qts of 0.7.
SB Acoustics :: 12” SB34NRX75-16
In any case I still do not quite understand why so little (or no) equalisation was needed in my case, but this is certainly not explained by most of the current models. The only explanation I found was this Dipole on axis
I am actually quite happy with this result and built a smaller "nude" design using one of the above woofers per channel and still happy with the overall SPL output.
“Experiment, make it your motto day and night”.
Phase instead of delay really helps some times.
I believe modelling is limited and really liked "Experiments are true answers". I get that.
As others have said, it is about time to stop using that textbook model of the rear waves coming around and annihilating the front waves, at least in real rooms with sound bouncing about.
Reproducing music at home with dipoles, like with motional feedback, almost everybody who tries it says it is much better music (and i suspect esp for "classical" or acoustic music).
But matters are quite different in the bass where we are sensitive to almost nothing but loudness - not direction and not distortion on music.
Along with tossing out the old dipole model, we also should toss out lab tests of bass directionality. On music, no localization south of maybe 140 Hz, based on my long experience and posts by others. Easy to make great speakers that go down to a 140 Hz crossover, even ESLs.
So my point is not to fuss about matching subs to mids in sound quality or finding a model to dictate the choice of crossover region. And measure, measure, measure.
B.
Last edited:
Well, I think there is not a lot of people saying/explaining that.
In the end I do not really care what others do, but I find it a pity that a lot of people apparently get dismotivated for smallish dipole speakers because they will read that one needs that 6db/octave correction and, which can easily amount to about 20dB SPL loss. But in real life that loss is not that dramatic, and nice results can be obtained from a pair of 10" or 12" woofers and small footprint.
In the end I do not really care what others do, but I find it a pity that a lot of people apparently get dismotivated for smallish dipole speakers because they will read that one needs that 6db/octave correction and, which can easily amount to about 20dB SPL loss. But in real life that loss is not that dramatic, and nice results can be obtained from a pair of 10" or 12" woofers and small footprint.
Agree with the above two posts. I've built some dipoles and my experience differs from the existing models. As for bass, one can start with with two 10-12inch drivers per frame (H, W, V). With a good sub amp it is easy to match the bass with the mains. Try Dayton Audio SA1000 where for reasonable money you get all you need: variable gain and LPF, switchable bass boost and infrasound filters, EQ and more.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.