http://db.audioasylum.com/cgi/m.mpl?forum=hug&n=51950&highlight=Hancock&r=&session=#postfp
I’ve read the material on Linkwitz’s site a while ago. This AA thread can’t be written to.
Anyone know what a Linkwitz-style **M baffle is?
Cheers
I’ve read the material on Linkwitz’s site a while ago. This AA thread can’t be written to.
Anyone know what a Linkwitz-style **M baffle is?
Cheers
Richard, I think he means W frame dipole. This is a standard dipole type which Linkwitz has made a particular version of which is perhaps slightly easier to build. A w dipole has the baffles holding the drivers in a W shape. It is apparently better than Linkwitz simplified version due to the absense of resonances caused by the parallel walls in his design.
A true W frame looks llike this:
or this:
If you haven't seen the site this image came from , then I think you will find it interetsing
A true W frame looks llike this:
An externally hosted image should be here but it was not working when we last tested it.
or this:
An externally hosted image should be here but it was not working when we last tested it.
If you haven't seen the site this image came from , then I think you will find it interetsing
rick57 said:
Hi Rick,
The m baffle I was refering to has all 45 degree angles, as opposed to the W baffle that has odd angles. The important difference is that the m baffle has one narrow slit in the midle of the back and two narrow slits at the sides of the front, which behave as one wide slit. You can see a picture of the dipoles in the following thread:
http://www.diyaudio.com/forums/showthread.php?s=&threadid=17642&highlight=
It is this narrow/wide arrangement that gets you a boost in directivity as you go up. However, you only get that boost when the width of the dipole is around 1/4 wavelength or more. If this dipole is for a subwoofer, then it's probably not important. I am crossing over my dipoles at 350Hz and it is important for my setup.
Hi John
Am I missing something: the thumbnail’s baffle has one slit in the back and two narrow slits at the front (the reverse of a Phoenix woofer), but no 45 degree angles?
I’m intending to cross my dipoles at (4th order) about 200 maybe up to 250 Hz, augmented with sealed subwoofers from about 50/ 60 Hz.
Do you see benefit over that frequency range in doing a M (per thumbnail) or 45 degree baffle?
(I’ll be getting the wood cut, I’m not a woodworker) Is it really that much harder to assemble a M over a U (NaO)/ 45 degree/ W?
And my 10 inch PHL mids & tweters are quite efficient (98 dB):
While it’ll be active, over 60 to 200-250 Hz, with Peerless XLS 12 (which I already have several of), which baffle shape is most efficient??
Thanks!
Am I missing something: the thumbnail’s baffle has one slit in the back and two narrow slits at the front (the reverse of a Phoenix woofer), but no 45 degree angles?
I’m intending to cross my dipoles at (4th order) about 200 maybe up to 250 Hz, augmented with sealed subwoofers from about 50/ 60 Hz.
Do you see benefit over that frequency range in doing a M (per thumbnail) or 45 degree baffle?
(I’ll be getting the wood cut, I’m not a woodworker) Is it really that much harder to assemble a M over a U (NaO)/ 45 degree/ W?
And my 10 inch PHL mids & tweters are quite efficient (98 dB):
While it’ll be active, over 60 to 200-250 Hz, with Peerless XLS 12 (which I already have several of), which baffle shape is most efficient??
Thanks!
rick57 said:
All the wood is joined at 90 degree angles, not 45--brainlock, there. Sigfried runs his dipoles below the point where you get the directivity effect I was refering to, so for the Phoenix it doesn't matter which direction you point the baffle. I run my dipoles up higher, so for me it does matter.
At 250Hz, a quarter wave is 331/250/4=33cm (13 inches). Keep in mind that this effect starts below the 1/4 wave point and that unless you have a very steep crossover, you will have meaningful output above the crossover point as well. So if the width of your dipole is around 10" or higher, then you will get a directivity effect from an m baffle--directivity will increase and be more biased toward the front.
The question, then is do you want a directivity effect. I am using a horn mid and tweeter, so yes I do because the this directivity boost matches almost perfectly the directivity of the horn in the crossover region. This is important because it keeps the power response of the speaker relatively smooth.
If you are using dipole mids, then you probably don't want this effect and would be better off with a baffle that has a more symmetrical radiation pattern, such as the w baffle.
As for the woodwrking, none of these baffles are too hard to do.
What really matters for efficiency is the external dimensions of the baffle, not the internal shape. An m baffle (an to a lesser extent the w baffle) is advantageous over an H baffle because the vibrations of the drivers cancel each other out. However, with the you m and w baffles, you have to worry more about the helmholz resonance. For m and w baffles, the helmholz frequency is 331/4/d Hz, where d is the depth of the baffle in meters. You get a resonance centered at the helmholtz frequency. For an H baffle, this resonance is centered at 331/4/(d/2) Hz--ie twice as high as the m and w baffles. At the Helmholz frequency, the m baffle has a sharper resonance than the w baffle.
Again, you can easily filter out this resonance with a notch filter, so you don't necessarily need to worry about it. If you are using an analog filter, the tradeoff is that you have some added distortion. I'm using a digital filter, so there is negligible impact from the filter.
hope that helps, John
Does the NaO avert theory with damping
Do you think that john K’s Nao model is likely to hold, without drawbacks - basically that damping the rear wave reduces the cancellation, and gives an efficiency gain? John K’s rear of the U frame is stuffed with 3 ½ inch fibreglass.
As he puts it:
Undamped there is more of a disadvantage to a U-frame than advantage compared to an H frame, since the on axis response is about the same. There is a little more radiated power to the sides, but the system resonance is also an octave lower, thus closer to the crossover point.
The story changes when damping is added. To envision what happens, consider that the response from the front and rear are composed of two components.
First there is the part due to the driver motion. Call that D. Then there is the part due to the resonance of the U-frame, call it R.
The front response is just D, and the rear response is -D +R, where the negative signs are used to note that the rear driver response is 180 out of phase with the front. Ideally, damping could be added such that R was driven to zero so the front a rear response would be identical.
In such a case the U-frame would generate a true cardioid polar response and would also be on axis SPL 6 dB greater than the H frame with similar damping over its useful range.
Unfortunately, R can not be completely damped without having other effects. At very low frequency, below the range of useful response, the damping has very little effect.
As the frequency rises towards the typical resonance (150 to 200 Hz, depending on U-frame length) the damping becomes effective and the 1/4 wave peak can be significantly controlled.
However, as the frequency continues to rise, the damping continues to be effective so that it is never possible to obtain front and rear radiation that is identical. The test is to find the optimum damping that will result in minimizing the rear radiation.
My experience is that the *rear radiation can be reduced by 10 to 15 dB over most of the U-frame’s useful range.
10 or 15 dB might not sound that great, but look at the implication on the effect of reflections from a rear wall. Consider a monopole, a dipole and a U-frame with 10 dB reduction in rear radiation.
With a monopole the reflection from the rear wall will radiate back forward and, neglecting differences in level, the reflected wave will sum with the direct front radiation either in phase or 180 degrees out of phase or somewhere in between. The limits of in and 180 out yield either a 6 dB peak or a null.
With a dipole, the result is the same except since the rear wave starts out 180 out of phase the frequency where the peaks and nulls occur will be inverted with respect to a monopole.
Consider the U-frame. Again the rear radiation will sum within the limits of in phase or 180 degrees out of phase. But since the level is 10 dB below the front radiation the reflected wave has less of an effect. Summed in phase the result is a maximum of a 2.4 dB peak. Out of phase the results is a worse case of a 3.3 dB dip. Obviously the in room case is much more complex than this, but there are clearly advantages with the front to back room resonances with a U-frame.
* So in summary, the correctly damped U-frame can yield a nominal 6 dB greater on axis sensitivity than an H frame dipole with the same "foot print" over its useful frequency range, which translates to 6 dB greater max SPL for the same driver excursion.
Total radiated power lies somewhere less than a monopole but greater than a dipole (a true cardioid has the same radiated power as a dipole). It potentially has advantages over both mono and dipole systems with respect to front to rear room resonances; and the single drawback is that the resonance peak is an octave lower that a dipole with the same footprint. This proves to be of little concern when used as a woofer system with appropriate crossover.
A dipole with the same sensitivity and max SPL limitations can be constructed - by making each leg of an H frame of length L so that the foot print of the dipole would be 2L, twice that of the U-frame. Of course, this dipole would also have the same 1/4 resonances as the length L U-frame. The idea driving me to use of the U-frame was the greater sensitivity in a smaller package.
Do you think that john K’s Nao model is likely to hold, without drawbacks - basically that damping the rear wave reduces the cancellation, and gives an efficiency gain? John K’s rear of the U frame is stuffed with 3 ½ inch fibreglass.
As he puts it:
Undamped there is more of a disadvantage to a U-frame than advantage compared to an H frame, since the on axis response is about the same. There is a little more radiated power to the sides, but the system resonance is also an octave lower, thus closer to the crossover point.
The story changes when damping is added. To envision what happens, consider that the response from the front and rear are composed of two components.
First there is the part due to the driver motion. Call that D. Then there is the part due to the resonance of the U-frame, call it R.
The front response is just D, and the rear response is -D +R, where the negative signs are used to note that the rear driver response is 180 out of phase with the front. Ideally, damping could be added such that R was driven to zero so the front a rear response would be identical.
In such a case the U-frame would generate a true cardioid polar response and would also be on axis SPL 6 dB greater than the H frame with similar damping over its useful range.
Unfortunately, R can not be completely damped without having other effects. At very low frequency, below the range of useful response, the damping has very little effect.
As the frequency rises towards the typical resonance (150 to 200 Hz, depending on U-frame length) the damping becomes effective and the 1/4 wave peak can be significantly controlled.
However, as the frequency continues to rise, the damping continues to be effective so that it is never possible to obtain front and rear radiation that is identical. The test is to find the optimum damping that will result in minimizing the rear radiation.
My experience is that the *rear radiation can be reduced by 10 to 15 dB over most of the U-frame’s useful range.
10 or 15 dB might not sound that great, but look at the implication on the effect of reflections from a rear wall. Consider a monopole, a dipole and a U-frame with 10 dB reduction in rear radiation.
With a monopole the reflection from the rear wall will radiate back forward and, neglecting differences in level, the reflected wave will sum with the direct front radiation either in phase or 180 degrees out of phase or somewhere in between. The limits of in and 180 out yield either a 6 dB peak or a null.
With a dipole, the result is the same except since the rear wave starts out 180 out of phase the frequency where the peaks and nulls occur will be inverted with respect to a monopole.
Consider the U-frame. Again the rear radiation will sum within the limits of in phase or 180 degrees out of phase. But since the level is 10 dB below the front radiation the reflected wave has less of an effect. Summed in phase the result is a maximum of a 2.4 dB peak. Out of phase the results is a worse case of a 3.3 dB dip. Obviously the in room case is much more complex than this, but there are clearly advantages with the front to back room resonances with a U-frame.
* So in summary, the correctly damped U-frame can yield a nominal 6 dB greater on axis sensitivity than an H frame dipole with the same "foot print" over its useful frequency range, which translates to 6 dB greater max SPL for the same driver excursion.
Total radiated power lies somewhere less than a monopole but greater than a dipole (a true cardioid has the same radiated power as a dipole). It potentially has advantages over both mono and dipole systems with respect to front to rear room resonances; and the single drawback is that the resonance peak is an octave lower that a dipole with the same footprint. This proves to be of little concern when used as a woofer system with appropriate crossover.
A dipole with the same sensitivity and max SPL limitations can be constructed - by making each leg of an H frame of length L so that the foot print of the dipole would be 2L, twice that of the U-frame. Of course, this dipole would also have the same 1/4 resonances as the length L U-frame. The idea driving me to use of the U-frame was the greater sensitivity in a smaller package.
passive tuning of a dipole* midrange
I’m also wondering about the benefit of damping the low end of a dipole* midrange, as an alternative to active equalisation (a la SL designs).
Denser and thicker fibreglass absorbs more at low frequencies (obvious really).
From http://www.bobgolds.com/AbsorptionCoefficients.htm I learnt there’s a *huge range of materials with different patterns of absorption coefficients over the frequency range, from around the world.
After massaging the data on 300+ materials in Excel, I compared the absorption at 250 Hz with absorption at 1000 Hz - the range of relative absorption coefficients of products listed there, is from 13% to 333%! Here’s a sample, with Selected Absorption coefficients (AC)
"'Columns": Product thickness density mounting 125hz, 250hz, 500hz, 1000hz, % AC @25 AC/ AC @1kHz, 4000hz
Owens Corning 700 series Rigid Fiberglass:
705, ASJ 25mm on wall 6.0 pcf (96 kg/m3) 0.2 0.64 0.33 0.56 114% 0.54 0.33
703, plain 1" (25mm) on wall 3.0 pcf (48 kg/m3) 0.11 0.28 0.68 0.9 31% 0.93 0.96
Owens Corning Sound Attenuation Fire Batt
Fire Batt 4" (102mm) 0.97 1.28 1.25 1.1 116% 1.1 1.09
Owens Corning Fiberglass Batts (fluffy pink), on the wall, and 16" from the wall
with Paper Out 3.5" R11 on wall 0.58 1.11 1.16 0.61 182% 0.4 0.21
Roxul Rigid Rockwool
AFB 1" (25mm) 3.0 pcf (48 kg/m3) 0.14 0.25 0.65 0.9 28% 1.01 1.01
from Jons Manville
Permacote® Linacoustic® R-300 1.5" (38mm) 3.0 pcf (48 kg/m3) 0.14 0.52 1.01 1.07 49% 1.03 1.97
IS 150 FSK 2" (51mm) 1.5pcf (24kg/m3) 0.25 1.09 1.11 0.58 188% 0.26 0.12
Knauf
Rigid Plenum Liner 1" (25mm) 3.0 pcf (48 kg/m3) 0.13 0.24 0.56 0.83 29% 0.92 0.98
CSR Bradford - Rockwool (Australia)
Glasswool Anticon Roofing Blanket - R2.0 Thermofoil 75mm kg/m3 0.6 1.21 0.9 0.41 295% 0.28 0.1
Glasswool Anticon Roofing Blanket - R2.5 Thermofoil 95mm kg/m3 0.72 1.43 0.82 0.43 333% 0.26 0.14
I hope that's readable - let me know if not!
Analysing the table on Bob’s web site gives much more confidence that a baffle can be built & measured, without detailed modelling, and later tuned to reasonable flatness with a degree of accuracy of finding the right material to give an appropriate degree of *passive equalisation.
If anyone wants the spreadsheet analysing the data in Bob's many tables, let me know (or tell me how to upload it!).
Cheers
I’m also wondering about the benefit of damping the low end of a dipole* midrange, as an alternative to active equalisation (a la SL designs).
Denser and thicker fibreglass absorbs more at low frequencies (obvious really).
From http://www.bobgolds.com/AbsorptionCoefficients.htm I learnt there’s a *huge range of materials with different patterns of absorption coefficients over the frequency range, from around the world.
After massaging the data on 300+ materials in Excel, I compared the absorption at 250 Hz with absorption at 1000 Hz - the range of relative absorption coefficients of products listed there, is from 13% to 333%! Here’s a sample, with Selected Absorption coefficients (AC)
"'Columns": Product thickness density mounting 125hz, 250hz, 500hz, 1000hz, % AC @25 AC/ AC @1kHz, 4000hz
Owens Corning 700 series Rigid Fiberglass:
705, ASJ 25mm on wall 6.0 pcf (96 kg/m3) 0.2 0.64 0.33 0.56 114% 0.54 0.33
703, plain 1" (25mm) on wall 3.0 pcf (48 kg/m3) 0.11 0.28 0.68 0.9 31% 0.93 0.96
Owens Corning Sound Attenuation Fire Batt
Fire Batt 4" (102mm) 0.97 1.28 1.25 1.1 116% 1.1 1.09
Owens Corning Fiberglass Batts (fluffy pink), on the wall, and 16" from the wall
with Paper Out 3.5" R11 on wall 0.58 1.11 1.16 0.61 182% 0.4 0.21
Roxul Rigid Rockwool
AFB 1" (25mm) 3.0 pcf (48 kg/m3) 0.14 0.25 0.65 0.9 28% 1.01 1.01
from Jons Manville
Permacote® Linacoustic® R-300 1.5" (38mm) 3.0 pcf (48 kg/m3) 0.14 0.52 1.01 1.07 49% 1.03 1.97
IS 150 FSK 2" (51mm) 1.5pcf (24kg/m3) 0.25 1.09 1.11 0.58 188% 0.26 0.12
Knauf
Rigid Plenum Liner 1" (25mm) 3.0 pcf (48 kg/m3) 0.13 0.24 0.56 0.83 29% 0.92 0.98
CSR Bradford - Rockwool (Australia)
Glasswool Anticon Roofing Blanket - R2.0 Thermofoil 75mm kg/m3 0.6 1.21 0.9 0.41 295% 0.28 0.1
Glasswool Anticon Roofing Blanket - R2.5 Thermofoil 95mm kg/m3 0.72 1.43 0.82 0.43 333% 0.26 0.14
I hope that's readable - let me know if not!
Analysing the table on Bob’s web site gives much more confidence that a baffle can be built & measured, without detailed modelling, and later tuned to reasonable flatness with a degree of accuracy of finding the right material to give an appropriate degree of *passive equalisation.
If anyone wants the spreadsheet analysing the data in Bob's many tables, let me know (or tell me how to upload it!).
Cheers
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.