gedlee said:
I'm not interested in equalizing structural modes. I don't even see how that would be possible unless you incorporated active damping elements into the enclosures (possible, but probably excessive and very expensive). I'm interested only in equalizing the frequency response (impulse response) so that the modal resonances affect on cone motion is reduced to zero.
If an enclosure resonance is imparting a force on the cone, wouldn't it be possible to feed the driver coil an electrical signal equivalent to the resonance in amplitude but of opposite phase. The forces cancel and the effects of the resonance on the frequency response are gone.
Hi Thadman,
I think this paper is relevant to the topic.
http://www.fesb.hr/~mateljan/arta/papers/im-aaaa2007.pdf
As far as standing waves are concerned, I think a 1.5”- 2” layer of open cell foam covering the internal surfaces of the enclosure will reduce the magnitude/Q sufficiently for them to be considered inaudible. On the other hand, untreated panel resonances in my opinion can be more of a problem, because the surface area of an enclosure is many times greater than the area of the driver’s cone, so panels only need a very small excursion for their acoustic output to equal that from the driver itself. For this reason, I always treat large enclosure panels with bracing.
To treat these resonances with some form of processing could help, but personally, I would opt for the solutions above before I resorted to processing techniques, simply because they are so inexpensive to implement.
There may be a greater opportunity to limit resonances due to standing waves by processing, because in its simplest form, standing waves exit through the loudspeakers cone. So in spatial terms, they are in a single dimension, however, panel resonances are in three-dimensional space, so processing cannot effectively treat them in my opinion.
Regards
Peter
I think this paper is relevant to the topic.
http://www.fesb.hr/~mateljan/arta/papers/im-aaaa2007.pdf
As far as standing waves are concerned, I think a 1.5”- 2” layer of open cell foam covering the internal surfaces of the enclosure will reduce the magnitude/Q sufficiently for them to be considered inaudible. On the other hand, untreated panel resonances in my opinion can be more of a problem, because the surface area of an enclosure is many times greater than the area of the driver’s cone, so panels only need a very small excursion for their acoustic output to equal that from the driver itself. For this reason, I always treat large enclosure panels with bracing.
To treat these resonances with some form of processing could help, but personally, I would opt for the solutions above before I resorted to processing techniques, simply because they are so inexpensive to implement.
There may be a greater opportunity to limit resonances due to standing waves by processing, because in its simplest form, standing waves exit through the loudspeakers cone. So in spatial terms, they are in a single dimension, however, panel resonances are in three-dimensional space, so processing cannot effectively treat them in my opinion.
Regards
Peter
FWIW...
I have A/B listened to pairs of speakers where one had an external vibration damper attached. The differences WERE audible AND significant.
Despite some of what I've read here, if an enclosure vibrates past some threshold it is audible and undesirable. Vibrating surfaces can make sound - if that surface is the speaker cabinet and not the transducer, artifacts are being added to what reaches the listener's ears that were not in the original recording.
At the least I would prefer to overbuild the cabinets and err on the side of caution.
I have A/B listened to pairs of speakers where one had an external vibration damper attached. The differences WERE audible AND significant.
Despite some of what I've read here, if an enclosure vibrates past some threshold it is audible and undesirable. Vibrating surfaces can make sound - if that surface is the speaker cabinet and not the transducer, artifacts are being added to what reaches the listener's ears that were not in the original recording.
At the least I would prefer to overbuild the cabinets and err on the side of caution.
Re: FWIW...
Of course this was a blind test!
Past some threshold it will, agreed. But I claim that threshold is a lot higher than what typically gets thrown about.
First, the "radiation efficiency" of the cabinet is going to be very low (if you don't understand that term, it means the level of radiated sound to the level of the vibration). Loudspeakers are designed to radiate sound and so they tend to have high levels of radiation efficiency. Second, the actual level of the vibration of the cabinet to that of the speaker is miniscule. The net efect is a very low level of sound radiated from the enclosure itself.
Doing nothing to reduce enclosure vibrations is going to be a problem, but a few key techniques will reduce the sound levels radiated to below the threshold of annoyance (if not audibility).
I have tried to measure the amount of sound radiated by the cabinet and found it to be extremely difficult. That makes it, at best, a very small effect.
sdclc126 said:
Of course this was a blind test!
sdclc126 said:
Past some threshold it will, agreed. But I claim that threshold is a lot higher than what typically gets thrown about.
sdclc126 said:
First, the "radiation efficiency" of the cabinet is going to be very low (if you don't understand that term, it means the level of radiated sound to the level of the vibration). Loudspeakers are designed to radiate sound and so they tend to have high levels of radiation efficiency. Second, the actual level of the vibration of the cabinet to that of the speaker is miniscule. The net efect is a very low level of sound radiated from the enclosure itself.
Doing nothing to reduce enclosure vibrations is going to be a problem, but a few key techniques will reduce the sound levels radiated to below the threshold of annoyance (if not audibility).
I have tried to measure the amount of sound radiated by the cabinet and found it to be extremely difficult. That makes it, at best, a very small effect.
Re: Re: FWIW...
Good evening Earl,
I am very much interested in your founding, since they seems in contrast with what Barlow and Stevens found out in theirs 1975 AES preprints: if I remember well, Stevens measured a 10dB difference between the back pannel radiation and the drive output, using an undamped 50 liters box.
How much did you measured, and which setup did you used?
gedlee said:
Good evening Earl,
I am very much interested in your founding, since they seems in contrast with what Barlow and Stevens found out in theirs 1975 AES preprints: if I remember well, Stevens measured a 10dB difference between the back pannel radiation and the drive output, using an undamped 50 liters box.
How much did you measured, and which setup did you used?
Interesting thread...
If I had to pick a single item which has caused me some of the most difficult-to-solve problems in DIY speaker building, it is cabinet resonances. I find I am incredibly sensitive to them. Maybe it's because I owned Magneplanars for many years and I am used to a dipole sound. Despite this, I am not a total torch carrier for the dipole crowd. I also like speakers in cabinets for reasons too numerous to go into here... Both have their strong points. BTW, dipoles DO have undesirable structure resonances. Even Sigfried Linkwitz found this out.
Let's talk about the issue of cabinet resonances. It has been proposed that if a cabinet contributed no nonlinearities but had resonances, that with proper filtering, you could avoid exciting these resonances and therefore avoid their audible effects. If a cabinet has a resonance at 300 Hz which augments the output from the speaker at that frequency, one must simply reduce the magnitude of the signal to the speaker by the proper proportion at that frequency and end up with a "flat" frequency response despite the cabinet resonance. Let me start by saying that I have actually used this scheme with some significant measure of success, but it is by no means perfect. Here's why... With the non-linearities of speakers, let's say you put in 2 signals; one at 700 Hz, and one at 1000 Hz. Just one of the most likely of distortion products to come from the speaker is the difference of 1000 - 700 = 300 Hz. This signal is an unplanned-for signal and we did not put it in. There are innumerable combinations of frequencies which, when you put them into our real world speaker driver, give distortion products at 300 Hz. This is why, although you may not be putting IN any frequencies at 300 Hz, intermodulation distortion products will exist at or around 300 Hz which will be augmented by the 300 Hz resonance in our cabinet. This will produce the "resonant signature" which is so immediately identifiable. And remember, the more complex the music is, the more likely it is that the distortion byproducts will coincide with that resonance. That is why using music sources with many voices and/or instruments will so mercilessly reveal cabinet coloration, even when you filter the input signal.
If I had to pick a single item which has caused me some of the most difficult-to-solve problems in DIY speaker building, it is cabinet resonances. I find I am incredibly sensitive to them. Maybe it's because I owned Magneplanars for many years and I am used to a dipole sound. Despite this, I am not a total torch carrier for the dipole crowd. I also like speakers in cabinets for reasons too numerous to go into here... Both have their strong points. BTW, dipoles DO have undesirable structure resonances. Even Sigfried Linkwitz found this out.
Let's talk about the issue of cabinet resonances. It has been proposed that if a cabinet contributed no nonlinearities but had resonances, that with proper filtering, you could avoid exciting these resonances and therefore avoid their audible effects. If a cabinet has a resonance at 300 Hz which augments the output from the speaker at that frequency, one must simply reduce the magnitude of the signal to the speaker by the proper proportion at that frequency and end up with a "flat" frequency response despite the cabinet resonance. Let me start by saying that I have actually used this scheme with some significant measure of success, but it is by no means perfect. Here's why... With the non-linearities of speakers, let's say you put in 2 signals; one at 700 Hz, and one at 1000 Hz. Just one of the most likely of distortion products to come from the speaker is the difference of 1000 - 700 = 300 Hz. This signal is an unplanned-for signal and we did not put it in. There are innumerable combinations of frequencies which, when you put them into our real world speaker driver, give distortion products at 300 Hz. This is why, although you may not be putting IN any frequencies at 300 Hz, intermodulation distortion products will exist at or around 300 Hz which will be augmented by the 300 Hz resonance in our cabinet. This will produce the "resonant signature" which is so immediately identifiable. And remember, the more complex the music is, the more likely it is that the distortion byproducts will coincide with that resonance. That is why using music sources with many voices and/or instruments will so mercilessly reveal cabinet coloration, even when you filter the input signal.
Re: Re: FWIW...
The listening test was not, and should have been, at least single blinded. However, the other part of the test was to play only the speaker with the damper, and to remove and replace it several times during play (it attaches to a self-stick magnetic plate placed on the cabinet). The difference in a mid-bass hump was glaringly apparent.
That aside, I understand perfectly the rest of what you're saying, and it makes perfect sense. A thick, heavy slab of wood is obviously going to present a challenge for a transducer to set in motion in any appreciable way. This should be good news to speaker builders, for who "proper" cabinet design is an obsession and the subject of endless threads here. Most of us have been led to believe that any vibration is audible and must be eliminated; that a cabinet must be 100% vibration-free.
I for one find it refreshing that this is not the case - another speaker building myth busted - as it will make future speaker building much easier. I have "overbuilt" my current cabs and I'm absolutely positive the dampers will offer no audible benefit.
If you're interested I'd be happy to ship the dampers to you for testing - perhaps you could show a percentage drop in vibration on a given speaker cabinet? Probably won't show anything earth shattering though.
The listening test was not, and should have been, at least single blinded. However, the other part of the test was to play only the speaker with the damper, and to remove and replace it several times during play (it attaches to a self-stick magnetic plate placed on the cabinet). The difference in a mid-bass hump was glaringly apparent.
That aside, I understand perfectly the rest of what you're saying, and it makes perfect sense. A thick, heavy slab of wood is obviously going to present a challenge for a transducer to set in motion in any appreciable way. This should be good news to speaker builders, for who "proper" cabinet design is an obsession and the subject of endless threads here. Most of us have been led to believe that any vibration is audible and must be eliminated; that a cabinet must be 100% vibration-free.
I for one find it refreshing that this is not the case - another speaker building myth busted - as it will make future speaker building much easier. I have "overbuilt" my current cabs and I'm absolutely positive the dampers will offer no audible benefit.
If you're interested I'd be happy to ship the dampers to you for testing - perhaps you could show a percentage drop in vibration on a given speaker cabinet? Probably won't show anything earth shattering though.
Re: Re: Re: FWIW...
Yes I've often wondered if there could be some kind of secondary effect - though the vibrations of the cabinet could be very low in audibility in and of themselves, could they possibly have a cumulative effect with the transducer - the whole being greater than the sum of its parts?
planet10 said:
Yes I've often wondered if there could be some kind of secondary effect - though the vibrations of the cabinet could be very low in audibility in and of themselves, could they possibly have a cumulative effect with the transducer - the whole being greater than the sum of its parts?
Comparing open baffle vs boxed speakers...
are you comparing the loudspeaker enclosure to one without or are you comparing differences in room sound/radiation pattern?
Unless you isolate one parameter at a time you are only playing a guessing game where beliefs and opinions cloud the real problems.
A room can easily add peaks and dips in excess of +/-30dB at the listening position...does an enclosure do that much damage?
are you comparing the loudspeaker enclosure to one without or are you comparing differences in room sound/radiation pattern?
Unless you isolate one parameter at a time you are only playing a guessing game where beliefs and opinions cloud the real problems.
A room can easily add peaks and dips in excess of +/-30dB at the listening position...does an enclosure do that much damage?
gedlee said:
Let me get this straight: You do a complex and pricey composite cab.
And you HADN'T validated your reason for doing so beforehand? And now you're saying that you're right and every other speaker builder is wrong to bother with serious cabinet building?
Those of us who've built many speakers usually know that better results are found when good, solid cabinetry is used. Most of us have underbuilt cabs and gone towards better and better constructions over time.
I don't need numerical validation when it's as obvious as it is. Just as I know duct tape is stickier than scotch tape.
Re: Re: Re: FWIW...
I guess that might be the flaw in my argument as the enclosure in question was very well damped and one could certainly NOT describe it as a poor construction. But I felt that it was over-kill and so I tried to measure the difference as I reduced the damping. I was unable to measure any difference or detect it audibly. That's not to say that going weak and weaker would not have found the threshold, I sure that I would have. My point is simply that I have found the extremes of construction unecessary - been there, done that. It's not worth the time or the money. That said, I would say that what I build now is right at the sweet spot. Nothing outrageous, but certainly very effective.
Its all about being "effective", not being "extreme".
claudio said:
I guess that might be the flaw in my argument as the enclosure in question was very well damped and one could certainly NOT describe it as a poor construction. But I felt that it was over-kill and so I tried to measure the difference as I reduced the damping. I was unable to measure any difference or detect it audibly. That's not to say that going weak and weaker would not have found the threshold, I sure that I would have. My point is simply that I have found the extremes of construction unecessary - been there, done that. It's not worth the time or the money. That said, I would say that what I build now is right at the sweet spot. Nothing outrageous, but certainly very effective.
Its all about being "effective", not being "extreme".
Its all about being "effective", not being "extreme"
That can't be overemphasized!
YES, resonances can be filtered out. Resonances can be dampend.
Panel resonances are not non linear distortion. They shift the ampitude a bit and consequently alter the group delay. It's all a linear process. What one gets is a little hump here, a dip there.
To measure it: build a flimsy box to try out. Measure the amplitude response thoroughly. Pave the box with tiles or something else heavy. But thin as not to alter the boxes outer dimensions to much. Measure again. Difference? Build an other box less flimsy, try again to make a difference with tiles.
Why not go into that very cost effective investigation and find out? Do it at least once as a test to get a feel for what is going on there.
Folks has overdone the building of cabinets (but in DIY only, and to tremendous levels) since decades. I'm convinced when lets say 10 people would publish their findings on that the topic will be through. What's the use in a debate without some valid data? You have to dig for papers that are more than 30 Years old, and it is only one! Nowadays mearement gear is such cheap. Forget about outdated papers and Do It Yourself.
so long
gedlee said:
That can't be overemphasized!
YES, resonances can be filtered out. Resonances can be dampend.
Panel resonances are not non linear distortion. They shift the ampitude a bit and consequently alter the group delay. It's all a linear process. What one gets is a little hump here, a dip there.
To measure it: build a flimsy box to try out. Measure the amplitude response thoroughly. Pave the box with tiles or something else heavy. But thin as not to alter the boxes outer dimensions to much. Measure again. Difference? Build an other box less flimsy, try again to make a difference with tiles.
Why not go into that very cost effective investigation and find out? Do it at least once as a test to get a feel for what is going on there.
Folks has overdone the building of cabinets (but in DIY only, and to tremendous levels) since decades. I'm convinced when lets say 10 people would publish their findings on that the topic will be through. What's the use in a debate without some valid data? You have to dig for papers that are more than 30 Years old, and it is only one! Nowadays mearement gear is such cheap. Forget about outdated papers and Do It Yourself.
so long
I want to explain a few things that may not be clear to people about cabinet effects.
There are two paths of unwanted sound from the speaker, through the air inside the box and through the structure - they act completely differently.
The air effect is seen as, and is, the most dominate. But because it is air and couples very poorly to any structure, this effect will be primarily LF only. Basically above the point where the box is a wavelength or more in size the air effect of the box back on the cone goes to nothing in any box with some absorption material. It starts to fall off much sooner than this, so we are talking about a few hundred Hz tops. If the box has no damping them the back wave can often be seen as an effect on the cones motion and hence on its output. But this is almost always brought to very low levels with absorption inside the box, except at LFs were the wavelengths are longer than box dimensions. At these frequencies, the absorption does not do much, and the back wave has a significant effect on the cone. This effect is simply what we see as the shift in resonance when the driver is put in the box. It is well know and accounted for in any good design. It is NOT a problem.
Hence we can see that the boxs effect on the air load behind the driver is not significant IF a good amount of damping material is used inside the box.
What about this air pressure effect on the box as a structure. This CAN BE significant, but a few simple techniques are very effective. First consider that once again, the wavelengths of sound in air are going to be much greater than the box dimensions. Thus it is reasonable to consider the box as having a fairly uniform pressure force at all surfaces - all in-phase. Now a box will have radiation modes, starting with a monopole (all panels moving in phase) a dipole (1/2 in-phase and 1/2 not), a quadrapole, etc. on up higher and higher. The important point is that as the mode goes higher and higher it radiation efficiency falls rapidly. Thus the worst case is the monopole mode (and its the most strongly driven), next the dipole, etc. Basically kill the first two and the problems are not significant
A simple cross brace, from front to back and side to side, attached in the center does wonders for reducing the lower order modes - the worst ones. If this brace is rigid, like the oak that I use, then the brace will not compress (oak is VERY strong in compression along its grain) and it will thus connect the two opposite sides together fixing them to be a fixed distance appart. This will basically kill the monopole mode since the sides cannot all move outward or inward together as the braces won't allow this. What about the dipole mode? Well this one is killed too if the braces cross in the center and are attached together. Thats because in this configuration the center point is fixed in space and connot move and this elliminates the dipole mode. ( In a monopole the CM stays put but all surfaces move in phase, but in a dipole the CM has to move. Only quadrapole and above have a stationary CM.) Now the quadrapole modes are not affected by the braces being limited only by the stiffness of the corner joints. BUT the quadrapole modes are way way down in radiation efficiency and don't radiate much sound at all. Thus, this simple technique of cross bracing is extremely efective at reducing box radiation from internal pressures.
Finally we come to the structure to structure vibrations, which are actually most effective at higher frequencies. These can be a real problem in an enclousre where something is not done to damped them. But a simple contrained layer damping (CLD) front baffle and rear baffle (which are rigidly attached together) basically kills any structural vibration from the speaker getting into the box. Anything that does get in can be damped with CLD on the other panels. There are some proprietary tricks to optimizing this technique, but its basically quite simple.
I have found that these techniques reduce enclosure problems to the point of being negligable and going any further than this is a waste.
There are two paths of unwanted sound from the speaker, through the air inside the box and through the structure - they act completely differently.
The air effect is seen as, and is, the most dominate. But because it is air and couples very poorly to any structure, this effect will be primarily LF only. Basically above the point where the box is a wavelength or more in size the air effect of the box back on the cone goes to nothing in any box with some absorption material. It starts to fall off much sooner than this, so we are talking about a few hundred Hz tops. If the box has no damping them the back wave can often be seen as an effect on the cones motion and hence on its output. But this is almost always brought to very low levels with absorption inside the box, except at LFs were the wavelengths are longer than box dimensions. At these frequencies, the absorption does not do much, and the back wave has a significant effect on the cone. This effect is simply what we see as the shift in resonance when the driver is put in the box. It is well know and accounted for in any good design. It is NOT a problem.
Hence we can see that the boxs effect on the air load behind the driver is not significant IF a good amount of damping material is used inside the box.
What about this air pressure effect on the box as a structure. This CAN BE significant, but a few simple techniques are very effective. First consider that once again, the wavelengths of sound in air are going to be much greater than the box dimensions. Thus it is reasonable to consider the box as having a fairly uniform pressure force at all surfaces - all in-phase. Now a box will have radiation modes, starting with a monopole (all panels moving in phase) a dipole (1/2 in-phase and 1/2 not), a quadrapole, etc. on up higher and higher. The important point is that as the mode goes higher and higher it radiation efficiency falls rapidly. Thus the worst case is the monopole mode (and its the most strongly driven), next the dipole, etc. Basically kill the first two and the problems are not significant
A simple cross brace, from front to back and side to side, attached in the center does wonders for reducing the lower order modes - the worst ones. If this brace is rigid, like the oak that I use, then the brace will not compress (oak is VERY strong in compression along its grain) and it will thus connect the two opposite sides together fixing them to be a fixed distance appart. This will basically kill the monopole mode since the sides cannot all move outward or inward together as the braces won't allow this. What about the dipole mode? Well this one is killed too if the braces cross in the center and are attached together. Thats because in this configuration the center point is fixed in space and connot move and this elliminates the dipole mode. ( In a monopole the CM stays put but all surfaces move in phase, but in a dipole the CM has to move. Only quadrapole and above have a stationary CM.) Now the quadrapole modes are not affected by the braces being limited only by the stiffness of the corner joints. BUT the quadrapole modes are way way down in radiation efficiency and don't radiate much sound at all. Thus, this simple technique of cross bracing is extremely efective at reducing box radiation from internal pressures.
Finally we come to the structure to structure vibrations, which are actually most effective at higher frequencies. These can be a real problem in an enclousre where something is not done to damped them. But a simple contrained layer damping (CLD) front baffle and rear baffle (which are rigidly attached together) basically kills any structural vibration from the speaker getting into the box. Anything that does get in can be damped with CLD on the other panels. There are some proprietary tricks to optimizing this technique, but its basically quite simple.
I have found that these techniques reduce enclosure problems to the point of being negligable and going any further than this is a waste.
Let's get REAL simple.
Put your hand on a panel with a speaker playing on the high end of normal listening levels.
Feel it vibrating?
Now, set the speaker to where the output right on the woofer 'feels' about the same.
Can you hear the driver playing?
Now consider that there's likely more cabinet surface singing along than that woofer, and it's pretty obvious that these are audible effects.
Naturally it's a cost/benefit system, especially for a commercial product, but it's a real issue, and if you're doing a design that's as complete as you can, you may as well address it. These are easily dealt with, you can use something as simple as a couple ribs and a rubber cement and 1/4" plywood shell overlay to significantly improve matters.
And most of us DIYers don't want to make something with a known problem just because it's easier and cheaper.
Put your hand on a panel with a speaker playing on the high end of normal listening levels.
Feel it vibrating?
Now, set the speaker to where the output right on the woofer 'feels' about the same.
Can you hear the driver playing?
Now consider that there's likely more cabinet surface singing along than that woofer, and it's pretty obvious that these are audible effects.
Naturally it's a cost/benefit system, especially for a commercial product, but it's a real issue, and if you're doing a design that's as complete as you can, you may as well address it. These are easily dealt with, you can use something as simple as a couple ribs and a rubber cement and 1/4" plywood shell overlay to significantly improve matters.
And most of us DIYers don't want to make something with a known problem just because it's easier and cheaper.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.