I use MDF, and oak bracing. I like MDF because of it's relatively low Q (resonances aren't as extreme as harder woods/materials. Plus it's relatively high mass eats energy better than a light weight material/wood. The oak bracing (and cross bracing) improves cabinet strength substantially, and can reduce what resonance the MDF does have (or raise them to frequencies outside of the passband of the driver).
The shape of the cabinet and how well it's damped internally with various acoustic damping materials is every bit as important IMO. I glue thick felt like material (upholstery padding for ex.) tight to all inside surfaces. Heavier material works better at lower frequencies. I double it up in the corners since they are even more resonant. Then a layer of foam rubber glued to that (maybe 1/2 inch to 1 inch thick depending on the size of the enclosure). Then the soft fluffy stuff, also glued to that so it doesn't end up right against the diaphram of the driver, or blocking its magnet vent.
You want what might be called a "gaussian arrestor", that will dissipate the acoustic energy at any frequencies that could otherwise resonate in the box. And you want to focus your absorption effort right where a bounce could happen (right against the internal cabinet surface). Fluffy stuff out in the middle of the cavity doesn't do much at low and mid frequencies. Mid frequencies are usually the ones that have the troublesome wavelengths, that interact with the internal dimensions of the enclosure.
Parallel walls are highly resonant. Other shapes are likely to be substantially less resonant. According to the Audio Cyclopedia, the ideal enclosure is a sphere with the diaphram at the surface of the sphere (or as close as you can get it) The worse enclosure shape is a sphere with the diaphram in the center. That would ring like a bell. Next worse is a cube with all internal dimensions being the same. You would think that a tube shaped enclosure (driver at one end) would be highly resonant, which it is, but with good damping I've found a tube to work VERY well on my midrange driver (300HZ - 8kHZ). Measured with pink noise, it was ruler flat over that entire range (driver was Peerless/Tympany TG9FD1008 3 inch - tube was 4 inch inner diameter ABS pipe, about 8 inches long, stuffing only in bottom half).
Trying to improve on MDF may not make much difference by itself.
The shape of the cabinet and how well it's damped internally with various acoustic damping materials is every bit as important IMO. I glue thick felt like material (upholstery padding for ex.) tight to all inside surfaces. Heavier material works better at lower frequencies. I double it up in the corners since they are even more resonant. Then a layer of foam rubber glued to that (maybe 1/2 inch to 1 inch thick depending on the size of the enclosure). Then the soft fluffy stuff, also glued to that so it doesn't end up right against the diaphram of the driver, or blocking its magnet vent.
You want what might be called a "gaussian arrestor", that will dissipate the acoustic energy at any frequencies that could otherwise resonate in the box. And you want to focus your absorption effort right where a bounce could happen (right against the internal cabinet surface). Fluffy stuff out in the middle of the cavity doesn't do much at low and mid frequencies. Mid frequencies are usually the ones that have the troublesome wavelengths, that interact with the internal dimensions of the enclosure.
Parallel walls are highly resonant. Other shapes are likely to be substantially less resonant. According to the Audio Cyclopedia, the ideal enclosure is a sphere with the diaphram at the surface of the sphere (or as close as you can get it) The worse enclosure shape is a sphere with the diaphram in the center. That would ring like a bell. Next worse is a cube with all internal dimensions being the same. You would think that a tube shaped enclosure (driver at one end) would be highly resonant, which it is, but with good damping I've found a tube to work VERY well on my midrange driver (300HZ - 8kHZ). Measured with pink noise, it was ruler flat over that entire range (driver was Peerless/Tympany TG9FD1008 3 inch - tube was 4 inch inner diameter ABS pipe, about 8 inches long, stuffing only in bottom half).
Trying to improve on MDF may not make much difference by itself.
I have quite a bit of experience with acoustic measurements and I don't think that what you propose would work very well. And actually what we want to know is how the fundamental affects the cabinet, not its harmonics. What we are talking about measuring is a secondary source of sound (a low level one at that), not the primary one, like the cone. It's trivial to measure cone's effect on sound radiation, but not so with cabinet radiation.
This is a common misconception. For a given volume all shape have the same number of resonances in a given bandwidth. The lowest resonance may change slightly, but the average of the first dozen or so will be the same for all shapes.
My first post. Hi all (and sorry for my poor english).
A very nice debate, and a lot of informations.
A small contribution to this thread with this link to a comparison of different panels, raw panels, constrained layers, etc... inspired by the study of Hardwood and Matthews.
HAUPT
In german, but go to the bottom of the page. Graphs and files are OK, and, I believe, quite intersesting.
I'm bulding a pair of Nomex 64 of Troels Gravesen. I respect the original design, but put a close box behind the tweeter.
The panels are made of 18mm Black MDF (higher density than MDF, approximately 25%), 4mm bitumen and 4mm plywood.
Front panel is made of 18mm Black MDF, 8mm MDF, 4mm bitumen and 4mm plywood.
The layers are seal with a product which keeps a certain flexibility.
I hope it's a good choice, surely better than a single panel of 18mm of MDF. But I realise that I'm in the middle of the debate. So...
I'll try to "decouple" the speakers with 4mm bitumen, 3 mm high density foam and Well Nut. I'll do a comparison with a classic front panel, and then decide. I'll post some measurements (Clio...)
A very nice debate, and a lot of informations.
A small contribution to this thread with this link to a comparison of different panels, raw panels, constrained layers, etc... inspired by the study of Hardwood and Matthews.
HAUPT
In german, but go to the bottom of the page. Graphs and files are OK, and, I believe, quite intersesting.
I'm bulding a pair of Nomex 64 of Troels Gravesen. I respect the original design, but put a close box behind the tweeter.
The panels are made of 18mm Black MDF (higher density than MDF, approximately 25%), 4mm bitumen and 4mm plywood.
Front panel is made of 18mm Black MDF, 8mm MDF, 4mm bitumen and 4mm plywood.
The layers are seal with a product which keeps a certain flexibility.
I hope it's a good choice, surely better than a single panel of 18mm of MDF. But I realise that I'm in the middle of the debate. So...
I'll try to "decouple" the speakers with 4mm bitumen, 3 mm high density foam and Well Nut. I'll do a comparison with a classic front panel, and then decide. I'll post some measurements (Clio...)
Placing two out of phase speakers face to face as close as possible goes a long way. Most sound from the speakers will interfere to its death, and what is left are sounds produced by the enclosures.
Because of distance, mainly. Also, don´t forget cabinet panels oth are large radiating surfaces and thus beam.
Do they beam forward at the listener? If so, what happens to those beams when the speakers are placed face-to-face and out of phase?
You can detect sound coming from the panels with any device that registers sound waves, such as a microphone.
Nice to see comparative measurements but one perhaps needs to be cautious about extrapolating from a single panel, single point measurements to a full structure with a quite different mode shapes and sound radiation patterns.
Does that mean you have a flexible bond between the damping material and structural and constraining layer? I would have thought a tight bond was important so that the full motion (which we want to be small) of the structural and constraining layers strains the damping material.
I look forward to seeing them.
??? Having placed the speakers face-to-face and wired out of phase are you suggesting pushing a microphone into the gap to measure the beaming from the front baffle?
Simply use a Stethoscope onto Any of the speaker box surfaces to detect 'sounds' from them.
Cheap, Easy and unambiguous.
Cheap, Easy and unambiguous.
I think this was already discussed and satisfactorily refuted in this thread posts #200-212. cheers
Those posts do not seem to say anything about stethoscopes.
A mechanics stethoscope is particularily good for finding resonances. It does rely on the human ear hearing the difference between touch and not touch, but i guess someone could figure out how to capture the output.
dave
A mechanics stethoscope is particularily good for finding resonances. It does rely on the human ear hearing the difference between touch and not touch, but i guess someone could figure out how to capture the output.
dave
While the cancellation technique may give some improvement, it still has a calibration problem. The question remains after the test "Can one hear this in the presence of the main radiation?" Comparative studies of panels and enclosures gives a comparative result, but lacks the "significance" estimation.
I tried to measure the effect of cabinet vibrations on the polar pattern which would have shown if the enclosure radiation was "significant" in the presence of the main radiation. I could not detect any change.
I tried to measure the effect of cabinet vibrations on the polar pattern which would have shown if the enclosure radiation was "significant" in the presence of the main radiation. I could not detect any change.
Earl, in the absence of ways to make absolute measurements, comparative measurements are at least something and can be used in a iterative process to improve enclosures.
For absolute measurements, a wide band accelerometer such as the kit produced by George Nvantaras is essential. But that is of course not a acoustic measurement.
For absolute measurements, a wide band accelerometer such as the kit produced by George Nvantaras is essential. But that is of course not a acoustic measurement.
Hi and sorry to jump in.
I would just like to know if there is an agreement on this: that the contribution of side and back panels to the overall performance is negligible
This would be for me extremely important.
Better to put the biggest effort on the front baffle design and construction ?
Thanks and regards, gino
Last edited:
In typical designs it can be expected to be larger but not negligible.
The hole for the driver weakens the front panel and the heavy driver hanging off it drives it directly and so a balanced approach would suggest more material for the front panel relative to the other panels.
In the end it is only the acoustic measurement at the listening position that matters, right?