Its tmww lay-out looks like mine system, also 8", 4", and 1.2" tweeter. Will try to add link to my post.
Was there mention in the book about what kind of attenuation/absorption coefficients one should look for to be effective enough? Any mention in loudspeaker enclosure context or was it for room acoustics only?
For example if a material has absorption coefficient of 0,1 (~1dB) at 200Hz and there is a resonance in a speaker enclosure on that particular frequency, the sound wave pass the material 200 times per second ( as rough simplification ). This leads to some kind of decay time. More absorption is better I guess, but I mean I have no idea what is good enough absorption coefficient? How short the decay should be? Is there this kind of info in the Tooles book? We know glass wool is good though. Maybe just use glass wool always like people have for decades
Just check out by measurements if more or less thickness is needed and get on with it.
Last edited:
The context in the book is room absorption but the principle is the same. In an room absorber you want it to be broadband as much as possible as low as possible. Which is 6 to 8 inches of most fibreglass.
Studios use huge thicknesses of low resistivity material as that gives the most broadband effect.
Inside a speaker you cannot hope to get the same effect.
Too much stuffing is also bad and sealed and vented will be different too. 200Hz is getting hard to absorb so cabinet damping thickness isn't going to do it. Placement of the damping inside the cabinet if it is long will help and bracing in the right places. No absorbent in a cabinet can fix a bad structure.For example if a material has absorption coefficient of 0,1 (~1dB) at 200Hz and there is a resonance in a speaker enclosure on that particular frequency, the sound wave pass the material 200 times per second ( as rough simplification ). This leads to some kind of decay time. More absorption is better I guess, but I mean I have no idea what is good enough absorption coefficient? How short the decay should be? Is there this kind of info in the Tooles book? We know glass wool is good though. Maybe just use glass wool always like people have for decades
In a sealed cab glass wool and felt works great and is cheap. Denim is similar and there are polyester products that are similar to the glasswool like those from Bradford. They are quite different to the pillow stuffing that many use which I found to have almost no actual effect on cabinet resonances.
Troels approach of using thicker felt in vented cabs seems sensible and practical.
Unless you are building a kit it is worth to try a few reasonable options and check the impedance to see what effect they had.
Imho stuffing remains trial and error. Using sound absorption coefficients acquired from impedance tube or diffuse sound field measurements isn't particularly useful because of the nature of the acoustic problem in an enclosure, being standing waves mainly. The location of the stuff in the enclosure is of far more importance than the nature of the material.
With most materials and some measurements good results are within reach. I use close miking BD plot measurements to evaluate the damping, preferably with 30dB or more range. As long as ridges from internal resonances stay visible, damping isn't sufficient.
With most materials and some measurements good results are within reach. I use close miking BD plot measurements to evaluate the damping, preferably with 30dB or more range. As long as ridges from internal resonances stay visible, damping isn't sufficient.
Thanks for all the great discussion on stuffing. It is relevant to my project.
I have always preferred loose wool for midrange enclosures, and I will continue with that. In this project, the midrange sub-enclosure is going to be about 5 liters.
For bass cabinets, I need to do something different this time. My previous projects used separate bass sealed boxes, with a big bass driver operating up to ~ 200 Hz. I used just a minimum of stuffing in that case, and the largest box dimension had a 1/2 wavelength frequency outside of the woofer pass band.
This time, the woofers will operate up to 300 Hz, and the largest internal cabinet dimension is 45", which has a 1/2 wavelength frequency of 150 Hz. So I definitely need to tame the first mode acoustic resonance inside the box. I might use wool, or I might use something else... So thanks for all the ideas.
I have always preferred loose wool for midrange enclosures, and I will continue with that. In this project, the midrange sub-enclosure is going to be about 5 liters.
For bass cabinets, I need to do something different this time. My previous projects used separate bass sealed boxes, with a big bass driver operating up to ~ 200 Hz. I used just a minimum of stuffing in that case, and the largest box dimension had a 1/2 wavelength frequency outside of the woofer pass band.
This time, the woofers will operate up to 300 Hz, and the largest internal cabinet dimension is 45", which has a 1/2 wavelength frequency of 150 Hz. So I definitely need to tame the first mode acoustic resonance inside the box. I might use wool, or I might use something else... So thanks for all the ideas.
I use sealed box for my woofers and midrange. These are filled with very well cleansed wool, so just the naked fibre structure ,no grease or anything else, and rquite short fibers.
The wool came in a very well pressed block (~ 200kilos, when i produced loudspeakers (many many years ago), and what we did was to pull this densely pressed wool apart in flocks of varying sizes of max a walnut size. Result is a amount of flocks of randomely varying size, and because of the pulling virtually no cavities in between.
Then in a sealed box, increase the amount in steps and measure the resulting resonance frequency. That will drop in frequency, with those steps you can draw a graph, and up to some step the graph is a linear line and then will start to bend and ultimately the frequency will rise a bit again.
From experience the point where it stops to be linear is the Q around 0.7 and where the res-frequency is the lowest the Q is around 0.5
Wool fibre when looked at under a microscope has a structure similar to a pine cone (those petals that can open and close). And with varying pressure it will open or close a bit, thus consuming energy thus gives damping. With a random varying size of flocks filling the box cavity its resonances are then also damped considerably, both in energy and in possibility to occur.
For the volume of the sealed box, just follow the normal calculators to start with, and without any corrections for apparent volume.
The wool came in a very well pressed block (~ 200kilos, when i produced loudspeakers (many many years ago), and what we did was to pull this densely pressed wool apart in flocks of varying sizes of max a walnut size. Result is a amount of flocks of randomely varying size, and because of the pulling virtually no cavities in between.
Then in a sealed box, increase the amount in steps and measure the resulting resonance frequency. That will drop in frequency, with those steps you can draw a graph, and up to some step the graph is a linear line and then will start to bend and ultimately the frequency will rise a bit again.
From experience the point where it stops to be linear is the Q around 0.7 and where the res-frequency is the lowest the Q is around 0.5
Wool fibre when looked at under a microscope has a structure similar to a pine cone (those petals that can open and close). And with varying pressure it will open or close a bit, thus consuming energy thus gives damping. With a random varying size of flocks filling the box cavity its resonances are then also damped considerably, both in energy and in possibility to occur.
For the volume of the sealed box, just follow the normal calculators to start with, and without any corrections for apparent volume.
Last edited:
Troels G. just launched a cheap SB design SBA-861-PFCR
Could be inspirating... or not ! Anyway cheap and even cheaper if made active with the Wondom DSP of the first post.
Could be inspirating... or not ! Anyway cheap and even cheaper if made active with the Wondom DSP of the first post.
Try and simulate the box with reflections in hornresp.
If you get the drivers positioned so that their acoustic center is middle of the longest dimension of the box you'll prevent the lowest mode happening. I observed this with single driver hornresp MLTL sim at least.
Reasoning from this the last box I built is very shallow, to get the depth dimension out of band. Then put the driver in the middle of the baffle, middle of the other two dimensions, to get rid all the first modes. At least in theory
I would very much like to hear those other reasons - why are they not good enough to be used as a midrange?
I think it depends on your design goals and processes.
I think it makes a good midrange, particularly between 200Hz and 2KHz in a monopole.
The reason I would lean towards a driver that is NOT the SB15N(B)/CAC is-
Needing higher sensitivity
Wanting a LP crossover point beyond 2KHz (cone size limitation)
Not needing a HP crossover point ~120Hz or below (excursion limited SPL)
I think it makes a good midrange, particularly between 200Hz and 2KHz in a monopole.
The reason I would lean towards a driver that is NOT the SB15N(B)/CAC is-
Needing higher sensitivity
Wanting a LP crossover point beyond 2KHz (cone size limitation)
Not needing a HP crossover point ~120Hz or below (excursion limited SPL)
Last edited:
Twenty years ago I built a pair of 50” tall, heavy, rear ported cabinets, each with a pair of Vifa M21, 8” woofers since they were $20 each then. I used some JBL car speakers for the top end.
Unfortunately I didn’t know what I wanted for sound, and wound up abandoning the project before I could make a proper crossover, better drivers. They weren’t exactly popular with my new wife at the time either.
They did make the gutters on my house rattle a few times though, and always have wondered if I could hear them now, what would I think?
Should be a nice set of speakers with the planning and research you’re doing…
Unfortunately I didn’t know what I wanted for sound, and wound up abandoning the project before I could make a proper crossover, better drivers. They weren’t exactly popular with my new wife at the time either.
They did make the gutters on my house rattle a few times though, and always have wondered if I could hear them now, what would I think?
Should be a nice set of speakers with the planning and research you’re doing…
That's my point.... If in need of a midrange... then around 300hz to 3000hz should be absolutely fine. I only see problems when insisting on a 2 way smaller speaker.
I wish for more real midrange drivers, and less midrange woofers. But maybe this is purely semantics. It just seems to me that most drivers are build to try and produce bass - which is to try and make the overall compromise, so that we can have a bit of bass and midrange from the same driver - even though we know that it is a bit of a stretch to have one driver to cover such a big frequency range.
I just bought the SB MW13TX in the hope that I'm getting a good midrange... we'll see
I agree. I would like to see a driver with an Sd of ~ 100 cm^2, an Fs of 60 Hz, sensitivity of ~ 92 dB/2.83V, an Xmax of 3 mm 0-pk, and a cone material of either aluminum or composite, with a first mode resonance above 6 kHz.
Then I would like to see smaller versions with an Sd of ~ 70 cm^2 and 50 cm^2, with proportionately higher Fs.
j.
I would wonder if people would be happier with pro sound mids with some of those requirements? Sticking with SB Acoustics the Satori mid ranges get close, but have a lower Fs. The FaitalPRO 6RS140 seems to fit right in what you are looking for in terms of Fs and other parameters, just not the cone material. It has also been used in a few successful hifi designs like Troel's 3WC.
I don’t mind any cone. But I want on my axis response to mimic the on axis; in a smooth way; at least an octave, or two beyond the passband.
The problem with many rigid cones is that even after you implement a perfect on axis response with your filter, you get all kinds of wacky off axis responses in the octave above.
The problem with many rigid cones is that even after you implement a perfect on axis response with your filter, you get all kinds of wacky off axis responses in the octave above.
Last edited:
I just spent several hours reading this entire thread, 1100+ posts (!)
The methodology you used in testing the acoustical damping strategy was very insightful. I was planning on using near field tone burst or CSD, but I will add impedance sweeps to the process as well.
Your line array is really cool, What a great project !
j.
I hope you found it useful, it certainly got longer than originally expected
What is nice about the impedance sweep is that the room is not much of a factor making it easier to get a look at the problem.
Agreed. The impedance tells a lot about the damping and the curve is more easily interpreted than a CSD. I left the CSD plots a long time ago in favor of the BD plots btw. Assuming audibility of resonances doesn’t change that much in a critical band of -say- 200 to 2000Hz, the BD picture lets you compare those resonances quite well. The only caveat indeed is you always measure reflections outside the enclosure too.
One obvious advice: when measuring impedance for this application, use slow sweeps or stepped sines.
One obvious advice: when measuring impedance for this application, use slow sweeps or stepped sines.