For decoupling the driver, rubber (damping material) should also be between the retaining screw/bolt head and the driver frame. And possibly in the frame hole between the screw threads frame hole internal diameter. The driver frame should be isolated completely from the baffle and screws.
I do find that mounting the tweeter on a rubbery gasket helps lose some HF harshness. 
I also chiselled out the strengthening front panel pine battens the other night. The front panel now has a lower frequency on the tap test, and sounds cleaner at high level.
If you are going to design a generally floppy sort of cabinet, keep it floppy EVERYWHERE. then the vibration just flows around, and eventually gets absorbed by the rubbery lining.
So the rigid beech battens or fillets are minimal and light. It's the panels that are the main event. I don't much care what they are made of.
I also chiselled out the strengthening front panel pine battens the other night. The front panel now has a lower frequency on the tap test, and sounds cleaner at high level.
If you are going to design a generally floppy sort of cabinet, keep it floppy EVERYWHERE. then the vibration just flows around, and eventually gets absorbed by the rubbery lining.
So the rigid beech battens or fillets are minimal and light. It's the panels that are the main event. I don't much care what they are made of.
I do find that mounting the tweeter on a rubbery gasket helps lose some HF harshness.
Hi and thanks for the advice. So it works. Good !
I guess you are referring to this picture here below ?
if so another confirmation of the importance to get a stiff front baffle.
And it is what i would do as a start
But any mod that can reduce flection should be very beneficial.
As i said i would look at the really best open baffle speakers to get some hints about the construction of the front baffle.
I do not think that there are ANY open baffle speakers with "floppy" front baffles out there ... i could be wrong but ...
i would say that i would never do that ... i am getting too many confirmation about the combinatio of extremely rigid baffle with "normal" cabinet behind that. Of course a front baffle rigid and self-damping would be the optimum.
Like those composite super material ... but the cost is an issue here.
Nevertheless i would love to listen a speaker with a baffle made out of those (i.e. panzerholz, corian, etc.)
I will try but they are very very rare ...
Important is that they work and your very interesting experience confirm.
I am sure some kind of distortion measurements would show the beneficial effects of front baffle stiffening. I am sure.
Thanks again, gino
Hi and thanks for the advice.
I think that is important that the tweeter does not move back and forth as a consequence of the vibes generated by the woofer ... that its flange stays still.
I would place it even on a support detached from the bass box and listen for differences in sound. As a test ...
Thanks and regards, gino
gino,
Since you've mentioned the Magico Mini a few times now, I thought it was perhaps worth noting that a few of the design features have been missed.
1 - attaching the baffles with threaded rods puts the baffles under tension and therefore increases stiffness even further (an excellent strategy when unable to afford to stiffest materials like aluminum or Panzerholtz).
2 - they will also distribute some of the unwanted energy to the back baffle as well (also detached from the cabinet).
3 - curved panels are stiffer than flat panels (think of an egg).
4 - squeezing the front and back baffles together with the rods will also put the curved sides (and also the internal X bracing) under tension, thereby increasing the stiffness of the cabinet even further.
It's really quite an elegant design.
Personally, I think it's also important to include sound transmission when talking about cabinet design which hardly ever happens in these types of discussions. Put a driver inside a cabinet, completely close it up, crank the volume and you are going to be shocked at what you hear. This will be equivalent to the energy coming from the back of each driver when attached to the baffle which the cabinet is supposedly suppose to contain and attenuate. Typically with something like simple 3/4" MDF and a litttle insulation, you are looking at very, very little loss of dB's in the lowest frequencies progressing to maybe 25 to 30dB in the highest.
So here I think you are mistaken to ignore the SPL's coming from your side panels. I look at it this way: if one is looking to use drivers with 3rd order (and greater) harmonic distortion that is 50 to 60dB below the fundamental in their usable frequency range, then one should apply the same standard to the unwanted sound coming from the cabinet panels right across the whole frequency range. If the harmonic distortion of your drivers is only down about 30dB, then fine, you're not going to really notice if the cabinet is adding less than that, but when it's the other way around, I think it makes a difference. Equal noise floors, in other words, for the drivers and cabinet is what I'm suggesting.
In this regard, I believe 2 things are your friend (besides internal absorption). The 1st is mass - more mass equals more reflected energy so more of the SPL remains inside the cabinet. And 2nd is again CLD where there is greater loss of energy when sound needs to travel through materials of very different Q, mass and density. And wow, isn't that Von Schweikert strategy looking good once again.
Just my 2 cents anyways.......
Since you've mentioned the Magico Mini a few times now, I thought it was perhaps worth noting that a few of the design features have been missed.
1 - attaching the baffles with threaded rods puts the baffles under tension and therefore increases stiffness even further (an excellent strategy when unable to afford to stiffest materials like aluminum or Panzerholtz).
2 - they will also distribute some of the unwanted energy to the back baffle as well (also detached from the cabinet).
3 - curved panels are stiffer than flat panels (think of an egg).
4 - squeezing the front and back baffles together with the rods will also put the curved sides (and also the internal X bracing) under tension, thereby increasing the stiffness of the cabinet even further.
It's really quite an elegant design.
Personally, I think it's also important to include sound transmission when talking about cabinet design which hardly ever happens in these types of discussions. Put a driver inside a cabinet, completely close it up, crank the volume and you are going to be shocked at what you hear. This will be equivalent to the energy coming from the back of each driver when attached to the baffle which the cabinet is supposedly suppose to contain and attenuate. Typically with something like simple 3/4" MDF and a litttle insulation, you are looking at very, very little loss of dB's in the lowest frequencies progressing to maybe 25 to 30dB in the highest.
So here I think you are mistaken to ignore the SPL's coming from your side panels. I look at it this way: if one is looking to use drivers with 3rd order (and greater) harmonic distortion that is 50 to 60dB below the fundamental in their usable frequency range, then one should apply the same standard to the unwanted sound coming from the cabinet panels right across the whole frequency range. If the harmonic distortion of your drivers is only down about 30dB, then fine, you're not going to really notice if the cabinet is adding less than that, but when it's the other way around, I think it makes a difference. Equal noise floors, in other words, for the drivers and cabinet is what I'm suggesting.
In this regard, I believe 2 things are your friend (besides internal absorption). The 1st is mass - more mass equals more reflected energy so more of the SPL remains inside the cabinet. And 2nd is again CLD where there is greater loss of energy when sound needs to travel through materials of very different Q, mass and density. And wow, isn't that Von Schweikert strategy looking good once again.
Just my 2 cents anyways.......
Gino, remember that Steve (System 7) just says stuff to see what we'll react with. Removing a stiffener from the front baffle and declaring that to be an improvement on an improvement is nothing other than the usual psycho-waffle behavior that goes on around here. JMO of course, but ironic that he's the waffler this time around
I'm sorry if I come over as a waffler here.
I did say very early on that cabinet discussions rarely lead anywhere. But we do have two schools of thought that seem relevant. Light and rigid and damped like BBC, or heavy and damped like some of Troels' designs. I like the idea of low stored energy for good resolution.
One thing for sure, is that speakers sound like what they are made of. You even have to consider that you will hear the metal speaker basket or the cone material. Or even the ferrite in the bafflestep coil. Or maybe the box or baffle or room. One of the strangest experiences I ever had was listening to a sound system in an anechoic chamber. It sounded frankly weird.
The ear is a very good sensor, IMO. It hears EVERYTHING.
Hi and thanks a lot for the very valuable advice
My interest is driven by the certainty the cabinets matter.
I still do not know how much. Drivers and x-overs of course matter more.
But the contribution of cabinets should not be underestimated.
i am not sure to understand well what you intend for loss.
You mean that the lowest freqs will pass through the cabinet to the listening room ?
This of course is bad. But i would look at the subwoofers to get some idea.
I see a very interesting arrangement with two identical woofere mounted back to back in the best woofers.
The action on the cabinet would be then nulled by the two woofers cones moving in opposite direction
Very very interesting ...
An externally hosted image should be here but it was not working when we last tested it.
problem is that the speaker will be a dipole ... with two emissions, frontal and back.
But the cabinet will stay still ... compressed and pulled but not moving in a specific direction.
Very very clever.
Thanks again and i promise to read more carefully that paper
But again ... if i see rightly a panel that does not move does not emitt sound at all
I hear a sound of a panel only if the panel vibrates ... am i right
So if i get a very stiff side panel it will not vibrate at all and consequently will not generate any sound at all.
I see aluminum speakers with extremely stiff panels and no damping of any kind inside the cabinet.
The panels are so rigid that they cannot vibrate or if they do they do only at a minimal level. They stay still even under the strong basses.
The arrangement of the woofer back to back is really winning.
But it could be used only below some frequencies ... at which the bass becomes omnidirectional ?
I think that this is really the best arrangement.
When the cones move forward the cabinet is pulled
When the cones move backwards the cabinet is compressed
But the cabinet is not push to move in a specific direction
I will study the paper you mention before rambling again
Thanks a lot again. Gino
Nope, not a dipole- a monopole. The wavelengths are long enough that the baffle size is irrelevant, the two drivers radiate in-phase, omnidirectionally, at all subwoofer frequencies.
Yes, I'm convinced cabinets make a difference as well.
But like I said, fully enclose a driver into a cabinet, turn it up to your usual listening level and you will be shocked at what you hear.
Yes, a stiffer cabinet will decrease sound transmission but I don't think quite as much as you think. It still won't be infinitely stiff and so sound will still be transmitted through it.
The main thing you are doing by increasing panel stiffness is raising its natural resonant frequency. Since the panel is now stiffer, it will also take more energy to excite that resonance and since there is actually less and less energy the higher in frequency you go, it's a win-win situation. Except just because a panel has a natural resonant frequency doesn't mean that that frequency is the only one it will transmit. It will be the one that takes the least energy to excite and the one that rings the longest but the rest of the frequency range is going to get through too, albeit less so for those frequencies with less energy.
Make a box out of steel and ask yourself if it's going to be soundproof. It will be stiffer than aluminum but the answer is still going to be no. Unless you make it massive enough. Or unless you add constrained layers with materials of highly varying Q.
Now exactly how soundproof an aluminum box across the whole frequency range would be I'm not sure, but it will be better than an equivalent thickness of MDF or BB plywood. Densities are going to vary for all 3 materials according to different grades and/or types (and I think twinter has also provided the stiffness figures), but figures I have put the aluminum at about 3 to 4 times as heavy as the other 2 and that will definitely (in combination with aluminum's higher stiffness) make a difference in the transmission loss but in and of itself, that aluminum box still won't be soundproof. Will it be enough? What I was trying to say above is that I think that also depends on the quality level of your drivers - the lower the noise floor of your drivers, the lower the noise floor of the sound coming from the cabinet needs to be as well.
One of the only sources I've seen to bring some of this up and attempt to deal with it is Art Ludwig. Give that a read although I would personally probably still go to greater lengths to deal with the issues.
But if you have the materials handy, throw a quick box together, enclose a driver in it and listen to the results.
But like I said, fully enclose a driver into a cabinet, turn it up to your usual listening level and you will be shocked at what you hear.
Yes, a stiffer cabinet will decrease sound transmission but I don't think quite as much as you think. It still won't be infinitely stiff and so sound will still be transmitted through it.
The main thing you are doing by increasing panel stiffness is raising its natural resonant frequency. Since the panel is now stiffer, it will also take more energy to excite that resonance and since there is actually less and less energy the higher in frequency you go, it's a win-win situation. Except just because a panel has a natural resonant frequency doesn't mean that that frequency is the only one it will transmit. It will be the one that takes the least energy to excite and the one that rings the longest but the rest of the frequency range is going to get through too, albeit less so for those frequencies with less energy.
Make a box out of steel and ask yourself if it's going to be soundproof. It will be stiffer than aluminum but the answer is still going to be no. Unless you make it massive enough. Or unless you add constrained layers with materials of highly varying Q.
Now exactly how soundproof an aluminum box across the whole frequency range would be I'm not sure, but it will be better than an equivalent thickness of MDF or BB plywood. Densities are going to vary for all 3 materials according to different grades and/or types (and I think twinter has also provided the stiffness figures), but figures I have put the aluminum at about 3 to 4 times as heavy as the other 2 and that will definitely (in combination with aluminum's higher stiffness) make a difference in the transmission loss but in and of itself, that aluminum box still won't be soundproof. Will it be enough? What I was trying to say above is that I think that also depends on the quality level of your drivers - the lower the noise floor of your drivers, the lower the noise floor of the sound coming from the cabinet needs to be as well.
One of the only sources I've seen to bring some of this up and attempt to deal with it is Art Ludwig. Give that a read although I would personally probably still go to greater lengths to deal with the issues.
But if you have the materials handy, throw a quick box together, enclose a driver in it and listen to the results.
No.
Yes and no, because the stiffer the panel the higher its Q.
At resonance the panel is virtually transparent.
It's not just the Q. It's the mechanical impedance which must be matched. That's why the inner constrained layer is usually very thin, so it's more sheared than bended.
For sound transmission there's the "mass law": the higher the mass the lower the sound transmission. Unfortunately, in the real world, there's always some stiffness, and it forms a resonant mechanical circuit where the panel is, again, virtually transparent at its resonant frequency.
In short:
- below the first resonance, you need mass to reduce sound transmission
- above the first resonance, you need mass and damping; the first to keep the overall level of sound transmission low, the latter to reduce the effect of the resonances
- you do not need stiffness (just so much that your cabinet doesn't collapse)
Well, it's possible I am incorrect as this topic is something new that I have only looked into recently, however my studies suggest otherwise, so .... Yes.
If you look into the variables involved in a panel's natural resonant frequency, you learn that it is possible to have the same natural resonant frequency on panels that vary in stiffness (the panels will need to be different in size, thickness and/or materials).
The result is that it will take different amounts of energy to excite each panel to the same amplitude even though they all have the same resonant frequency. Obviously the stiffer panels will require greater amounts.
Perhaps I should have added "to the same amplitude" to my first statement above to make it completely accurate. <shrug>
I can post the plate theory formulas if you like.
I believe for basic panel vibration concepts:
Stiffness controlled (Low)
Fundamental resonance frequency region, damping impacted
Mass controlled (Mid)
Coincidence frequency region, damping impacted
Damping controlled (High)
Noise insulation case
Last edited:
If a light, stiff and curved panel is used we do not require high mass for sub frequencies. I have built many a sub and nothing more than a sonotube is required below 140Hz. Above this it is fairly acoustically transperent but considering it is only1/8" thick cardboard, not suprising. My current mltl is a 8" (8.25" od) sakrete tube that is subdivided lengthwise with a 8" wide 18mm mdf that runs the length that results in a 74" length folded to ~48". End driven by a peerless 6.5" at sub frequencies there is little panel resonance, but it does move along the length axis due to lack of force cancellation and low mass. Again, not a suprise
Actually, my "No" is a bit too strict.
If you want more stiffness in a flat (not curved) panel you have basically three possibilities:
1) other material (with higher modulus)
2) thicker panel (doubled thickness => 8x stiffness)
3) bracing
How the resonances look like with 1) depends on the other material properties like density and damping. Let's not discuss it here.
2) is a way which is very popular in the audiophile world, because thicker panels also means higher mass, and higher mass is always GOOD!!!11!!
What audiophiles (and the manufacturers) always miss is that doubled thickness means 8x the stiffness and 2x mass means 2x resonant frequency and 4x Q (Fc ~ sqrt(stiffness/mass), Q ~ sqrt(stiffness*mass)). This 4x Q means 4x amplitude if driven with the same excitation, and 2x amplitude with a real driver (F is constant, but velocity is halfed). Actually, many measurements show that the amplitude of the higher resonant frequency is equal to that if you'd use a thinner panel, I have to admit that I have no explanation for this behaviour.
Bracing 3) adds stiffness without adding (much) mass, so we will still have a higher Fc and a higher Q, but the Q will not rise as much as in example 2), so I guess that the amplitude will be a little lower.
Unfortunately, resonances higher in frequency are more likely to hear, so I don't think that bracing will reduce the vibrations and transmission that they will be unhearable.
Still, the best solution is to add mass and damping, but no stiffness (or so much damping that the stiffness doesn't hurt). Using thin wood panels with thick bitumen layers is one good idea (BBC style), CLD is another good solution. Other methods, like the magico ultra-braced aluminum cabinets may also work well (if they push the resonances high enough in frequency), but they're stupid from an engineers POV because it is a waste of time, material and energy (but the marketing guys will be really happy about it).
Hallo
i understand that by adding a lot of bracing the resonance frequency of the panel is pushed above the "harmful" range as in that high that the resonaces won't be excited anymore. And the amplitude will be less / lower.
Example: closed cabinet for driver with an xo to tweeter about 2.5kHz.
Afaiu having stiff walls (plywood) + lots of bracing pushes reso. freq. above excitation frequency.
What are your thoughts on that?
Regards
i understand that by adding a lot of bracing the resonance frequency of the panel is pushed above the "harmful" range as in that high that the resonaces won't be excited anymore. And the amplitude will be less / lower.
Example: closed cabinet for driver with an xo to tweeter about 2.5kHz.
Afaiu having stiff walls (plywood) + lots of bracing pushes reso. freq. above excitation frequency.
What are your thoughts on that?
Regards
If only someone in the DIY community would assembly the equipment to perform vibration testing on enclosure walls; with accelerometer, frequency spectrum analysis, cumulative spectrum analysis, etc.
I have been constructing a three way tower speaker, approximately 41" tall. After years of reading numerous comments about MDF and Baltic birch plywood, I built the cabinet walls of 1" MDF, with a 2" thick front baffle (2 layers). I used significant 3/4" Baltic birch plywood bracing. Applying the low tech knuckle rap test, the enclosure appeared to ring at numerous points. I borrowed an accelerometer and a computer with FFT frequency spectrum analysis software (no CSD waterfall software), and found sharp resonances from 330 to 800 hertz. The Baltic Birch 4" x 8" internal braces were even resonating at 400 Hz, and they were attached on three sides. This matches the resonance frequency ranges of what others have posted on the internet for MDF and plywood baffles using Cumulative Spectral Decay (CSD) plots.
When I added Baltic Birch braces in the middle of the previously 4" x 8" internal braces, their resonance frequency went up to 625 Hz and their ringing definitely decreased. I'm not sure why the ringing decreased so much. If stiffness is used as an objective in wood material enclosures, then maybe a resonance of 600 Hz or better may be beneficial. This is a preliminary observation, after reading more than several articles, and my limited experience. Very few enclosures made of Baltic Birch or MDF are sufficiently stiff to have the majority of their panels with resonances at 600 Hz or greater. This requires bracing every 4 to 6 inches with one inch thick wall panels.
I also applied epoxy resin with fiberglass and carbon fiber woven roving, epoxy resin with powdered limestone, and extensional damping material to the interior walls. These were efforts to further stiffen and damp the enclosure walls, with some improvement noted though not as much as desired.
HDF, MDF, and plywood (Baltic Birch) do not have significant internal damping, nor do natural stone materials. If these materials are used, effective damping methods need to be applied for high performance results. Laminated bamboo is stiffer and has greater internal damping than Baltic Birch or MDF. Chipboard (European), which I believe is call particle board in the USA, also has greater internal damping than Baltic Birch or MDF.
The 1970's "BBC style" of using thin wood panels with thick bitumen layers is a good application of extensional damping. The resonance is accepted in the playing range, but its magnitude is significantly reduce by the proper application of damping material. For extension damping to be effective, the damping layer should be at least one half the thickness of the panel, and better with the damping layer the same thickness as the panel. The thin wood panels helps facilitate this.
Now moving into the twenty first century, using a damping material as proposed on the Qualia website, a isophthalmic polyester resin and bentonite (cat litter) mixture (50/50% by weight), would allow thicknesses up to 1". The mixture has a high damping factor, 0.52 to 0.63. The isophthalmic polyester resin differs from the more common orthophthalic polyester resins, as it has less heat developed and less shrinkage during curing. So isophthalmic polyester resins can be used in thicker pours and results in less internal stresses. If applied in 1/2 to 1 inch thicknesses, the loss in internal volume would need to be taken into account. Extensional damping material relies on bending to function, so high internal damping and (surprisingly) stiffness similar to the panel is desired, not excessive elasticity. So applying the isophthalic polyester resin and bentonite mixture, you can potentially have stiff, damped enclosure walls. This is just one simple direct damping method. I'm also fully aware than constrained layer damping is more effective than extensional damping. I have the isophthalmic polyester resin, just have not used it yet.
A link to the Qualia website:
damping factor values : damping factor values
Last edited:
Been there; done that: http://www.diyaudio.com/forums/multi-way/218981-enclousure-strengthening-needed-2.html#post3151455. Numerous others DIY have made such measurements, too. Also a lot of researchers did those test. Look for Harwood from the BBC. After reading his report there should be no question about constructing a cabinet.
Unfortunately, the industry, the consumers, and the DIY community vastly ignores it.
sayrum said:
Look at the post/pictures I've linked above. Even tweeters excite panel resonances. But, if you can push the resonances above 2 kHz, than it should be safe. The available suggests that the audability of resonances is very low above. But are you sure you can do this? And it is worth the effort? Why not just use 9 mm plywood with 9 mm bitumen?
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.