I think the undamped materials will show much more differences than the materials with a substantial amount of damping added (and that is what we want). As the damping is increased the varied resonances of different materials would be well damped, the differing inherent resonances and Q's of the materials should be dominated by the added damping.
An MDF violin and a plywood violin would sound pretty different, but put half an inch of tar on both and they should be fairly similar.
Note that MDF is a generic term and there are a lot of commercial products available with slightly different characteristics. MDF is well liked in the industry because it machines easily, is cost effective and is generally heavier than plywood. The only speaker market that frequently uses plywood would be the music/PA market, where its ability to withstand hard knocks is a critical attribute.
David S.
An MDF violin and a plywood violin would sound pretty different, but put half an inch of tar on both and they should be fairly similar.
Note that MDF is a generic term and there are a lot of commercial products available with slightly different characteristics. MDF is well liked in the industry because it machines easily, is cost effective and is generally heavier than plywood. The only speaker market that frequently uses plywood would be the music/PA market, where its ability to withstand hard knocks is a critical attribute.
David S.
Harbeth are hardly mass-produced (there are some videos on Youtube). MDF wasn't available in the early 70s when the original research paper was produced. It's likely Harbeth are using it because of acoustic benefits rather than cost (it's not like their speakers are at the cheaper end of the market).
Alan A. Shaw's view on the matter:
BBC-style thin-wall cabinets. Why so special?
BBC-style thin-wall cabinets. Why so special?
Last edited:
Very interesting. This would explain why my cheapies with thin cabinets sound better than the other 1" thick MDF "high-end" stuff I own.
Using this principle, is it still important to brace large areas of thin wall? Or is that counterproductive?
Does anyone have a spec for the bitumen sheet? Im not sure if that product is freely available over here.
Using this principle, is it still important to brace large areas of thin wall? Or is that counterproductive?
Does anyone have a spec for the bitumen sheet? Im not sure if that product is freely available over here.
Go to a fancy car stereo installer and see what they well sell you. Look for "deadsheet" or similar products. They all use it.
Very good question about bracing. I'm not sure what the answer is.
David S.
Interesting.
I've never had any particular problems with the sound of cabinets where both the front and back panels are removable and screwed into place despite dire warnings of "experts" that good cabinets should only ever use glue joints not screws.
Mind you I'm not talking about small wood screws, rather a generous number of 50mm machine thread cross dowel bolts which are screwing into a threaded insert of some sort, along with a cork sheeting gasket for the panel to clamp onto.
It had occurred to me before that not gluing the front and back panels but instead clamping them via a flexible gasket would break up the "whole box" resonance that would form in a fully glued box, but I'd never done any direct comparisons to see if it was audibly beneficial or not.
I wonder whether a sufficiently thick flexible gasket clamped between the removable panels and the rest of the cabinet counts as a degree of constrained layer damping ?
Last edited:
To brace a thin wall cabinet, thus moving panel resonances up in frequency, is largely counterproductive.
That said, some of the larger cabinets have some form of braces. Look at Harbeth 40, they have braces on the sides, but not actually touching the wall itself. Some others have braces that connect two or three walls, often only by sticks. I've yet to see any substantial stuff, like shelf braces.
That said, some of the larger cabinets have some form of braces. Look at Harbeth 40, they have braces on the sides, but not actually touching the wall itself. Some others have braces that connect two or three walls, often only by sticks. I've yet to see any substantial stuff, like shelf braces.
Last edited:
If you add a matrix style brace between a large front and rear panel that divides it in half over the full width, (essentially like a shelf but with holes cut to join the two cavities together) you increase (double?) the resonant frequency of each remaining section of panel.
However what happens if you simply place a strut between front and back panels in the middle ? A strut that is for example 20mm x 40mm. Somehow I don't think the result is the same as a full matrix brace in terms of distribution of resonant frequencies.
Sure, you've divided the height and the width of the panel in half - but only for a very narrow "cross" centred on the middle of the panel. For the 1/4 and 3/4 positions in each axis the panel is still free to vibrate.
Has anyone measured or simulated the modal vibration pattern of a panel braced at only one point ?
I doubt that it simply doubles the resonant frequency of the panel, rather that it spreads the resonance to more than one frequency, but with each individual resonance at a lower amplitude than either the un-braced or fully braced (divided in half) case.
Yes ? No ?
Last edited:
I've seen simulations. The lowest cabinet mode is where all sides alternately bulge in and out in phase. The complete surface of every panel is moving in the same direction (no upper modes).
A brace from one side to the other will strongly resist that mode.
Higher modes such as when the panels start to break into 2 or more regions will not be resisted by that brace, as long as they naturally have a node at that point.
A bit like driving a room with a subwoofer in the center of the floor. Modes with a node there are not driven.
David S.
A brace from one side to the other will strongly resist that mode.
Higher modes such as when the panels start to break into 2 or more regions will not be resisted by that brace, as long as they naturally have a node at that point.
A bit like driving a room with a subwoofer in the center of the floor. Modes with a node there are not driven.
David S.
Some thoughts on this:
1) If I understand it correctly, the 2nd harmonic mode of a panel should in theory cancel at the listening position (half the panel is moving in while the other half moves out) and the 3rd harmonic mode should partially cancel (1/3rd moving out while 2/3rd moves in) with 4th harmonic once again cancelling out.
Of the various panel modes, does this mean it is always the fundamental mode (all of the panel moving in and out together) which produces the greatest amplitude at the listening position ? And that the next worst offender is the 3rd harmonic ?
(I realise each axis of a panel will have a different fundamental resonance if they differ in length, so it becomes more complicated with a non square panel)
2) Does that mean bracing, depending on its position relative to potential modes, does not increase the resonance frequencies of a panel, but rather it only selectively suppresses the amplitude of certain resonance modes ?
So for example imagine a 600mm high panel of a certain thickness had a fundamental resonance of 150Hz when un-braced and only supported at the edges, thus giving it a 300Hz 2nd harmonic mode, 450Hz 3rd harmonic mode and so on.
Placing a brace at the half way point would strongly resist the 150Hz mode and at least partially resist the 450Hz 3rd harmonic mode, but would not resist the 2nd harmonic mode at all. However if the second harmonic mode doesn't contribute significant net radiation, (due to the two halves being out of phase) that may not matter much. (?)
If you increase stiffness of a panel by making it thicker, you still have the same set of resonances occurring, (1st, 2nd, 3rd harmonic and so on) and probably in the same proportions, but you just push them higher in frequency, and by increasing the mass you make it more difficult to damp them - which goes against the whole concept of keeping the resonances as low in frequency as possible and low in mass that is being discussed in this thread.
I see a comment in this thread and also many times elsewhere that placing a brace in the middle of a panel just doubles the resonant frequency of the panel, which would go against the approach of keeping the resonance frequency of the panel low. But does it really do that ?
If you look at the 2nd harmonic mode, a strut in the middle should not affect its frequency or amplitude at all, as its placed at a node. If the 2nd harmonic mode has not gone up in frequency then by definition the fundamental cant have shifted either ? After all the bracing has not added extra stiffness to the overall panel, nor mass, it is just supporting it in a previously unsupported point, thus suppressing the fundamental mode.
Would it be fair to say that in going from no brace to a brace at half way, the amplitude of fundamental and 3rd harmonic vibrations will have dropped significantly, but the 2nd harmonic will be no worse than before and still at the same frequency ?
We have not doubled the resonant frequency of the panel, we still have the same harmonic structure but the fundamental is almost completely suppressed and odd order harmonics are reduced.
Even though an accelerometer half way between the brace and one side will show the 2nd harmonic as the lowest resonance, as if the fundamental resonance had been doubled, the two halves of the panel will be out of phase and largely cancel (provable with a phase comparison of two accelerometers on either side of the strut) - thus showing that its actually a 2nd harmonic mode, not a fundamental mode that has doubled in frequency.
Or am I missing something ?
From this it seems that bracing at the half way point doesn't go against the concept of keeping panel mass and resonant frequencies low, and in fact can help by suppressing the fundamental resonance, reducing odd order harmonic resonances, without making even order resonances any worse or raising the frequencies of any individual resonance modes.
This seems to suggest that bracing is much more effective at controlling panel resonances than making the panels thicker, and that you're actually better off using thinner heavily damped panels with some bracing than very thick un-braced panels...
3) How does a full width "brace" differ from a strut style brace or does it not differ along the axis which it is dividing ?
For example imagine a front/rear panel that is 600mm high by 400mm wide, and you can either brace it by using a 400mm wide by 20mm high "shelf" style brace, (with a couple of holes cut in it so as not to partition the inside space) or you can use a 40mm wide by 20mm high strut placed in the middle.
In the vertical axis are the two any different ? The full width brace will clearly divide the panel vertically into two equal sections, and should suppress the fundamental mode of the panel in that axis very well. However will that narrow 40mm wide "strut" fully suppress this mode, or will the flexibility of the panel itself allow the horizontal sections that are not braced to still vibrate somewhat at the fundamental mode ?
Last edited:
Thanks Dave, there is a product called Bostik Sound Deadener available here if anyone is looking.
If we say that all materials resonate and we need to push the resonant frequency either very high or very low, we have 2 options:
1. Use thin wall material with no panel damping and extensive bracing matrix to raise the panel resonant frequency as high as possible.
2. Use thin wall material with heavy damping and no bracing, in order to lower the resonant frequency as much as possible.
Most cabinets built these days go the opposite way - Thick MDF walls that lower resonant frequency, ineffectual damping (due to high wall thickness) and extensive bracing that then raises the resonant frequency, counteracting the whole purpose of using thick MDF. Sounds silly no?
Hi,
No. Thick MDF walls raise resonant frequency, as stiffness relates to the
cube of the thickness, whilst mass only increases linearly. Your right,
you can't damp thick walls effectively. No. Extensive bracing further
stiffens thick wall cabinets and does not counteract using thick MDF.
The point here is I guess damping requires the walls to move, or it
won't work, and will be ineffective. High stiffness cabinets seek to
reduce inevitable movement (i.e. colouration) to a minimum.
The point of thin walled cabinets with thick damping layers is to drop
all resonances in frequency such that the midrange is not resonant,
only makes sense when using a bass / midrange driver. Bass
preciseness is given up for clear midrange, and it does work.
I had a pair of Spendor Preludes, not small, (~ 11"x12"x24") with
6mm thick chipboard side walls and about the same again thickness
bitumen damping layer. No bracing, Chipboard 12mm back panel,
18mm front panel, both with with no damping, but it worked well.
And still does after many years, a friend has them. Essentially
warmish bass but with very low "boxiness" in the midrange.
rgds, sreten.
Thanks Sreten,
If I understand it correctly, the material should move or flex so that the damping material moves along with it, dissipating the energy through heat and avoiding energy storage in the cabinet wall.
If so we should be using the more flexible types of timber (or rubber!). What would work best here?
With respect to the thick stiff wall approach, unless the resonances can be pushed well passed the crossover point (when crossing to a sealed driver / tweeter that isn't firing into the box), it seems to me like an inferior way of building a box. And less sustainable!
The Harbeth article shows a large peak centred on 500hz for 18mm MDF. How much bracing and additional thickness would need to be added just to get that peak up to say 1khz? And even then it is still in the critical band and well below most crossover points.
If I understand it correctly, the material should move or flex so that the damping material moves along with it, dissipating the energy through heat and avoiding energy storage in the cabinet wall.
If so we should be using the more flexible types of timber (or rubber!). What would work best here?
With respect to the thick stiff wall approach, unless the resonances can be pushed well passed the crossover point (when crossing to a sealed driver / tweeter that isn't firing into the box), it seems to me like an inferior way of building a box. And less sustainable!
The Harbeth article shows a large peak centred on 500hz for 18mm MDF. How much bracing and additional thickness would need to be added just to get that peak up to say 1khz? And even then it is still in the critical band and well below most crossover points.
No, the point is to increase the ratio of damping to mechanical impedance (stiffness or mass) and greatly reduce the Q of the resonances, not to alter the frequency of resonances at all.
That is explained in the paper, that midrange resonances are audible when output approaches some level relative to the woofer cone output. Reducing the level of the resonant peaks will get them below the threshold of audibility. Harwood makes no attempt to "push them up" or "push them down" in frequency. That is never a practical proposition.
David S.
I'm mostly relying on the Harwood paper:
http://diy-audio.narod.ru/litr/1977-03.pdf
Okay, I guess he does intend to "push them down" by using thinner walls.
His criterion is to get the level of resonances more than 30dB down for frequencies above 500 and progressively less below that. So, yes, if you could get a resonance above 500 Hz to move to well below 500 Hz then you may get real improvement. Thinner panels will move the resonances down as Harwood shows in his figures 13 and 14.
Still, his measurements show the resonances dropping in level and not appreciably changing in frequency, when damping is applied. In the one example of the LS3/6 cabinet, he shows that both thinner panels and damping were required.
David S.
You say bracing raises the panel resonant frequency, but why would it ? The stiffness of the panel material isn't changed by the bracing, you're just supporting it in an additional location altering the modal vibration patterns of some modes.
Putting a brace half way along a single panel is not equivalent to making the panel thicker, nor is it equivalent to cutting the panel in half and terminating each cut half on a separate frame.
In the case where you cut the panel in half the fundamental resonant frequency does double because the length is halved, and each panel can bend independently. (They could both bow outwards at the same time for example)
However if you put a brace at the half way point, which is a node for the 2nd harmonic, the two halves of the panel that are separated by the brace are still connected to each other - and can't bend in opposite directions at the junction, they must bend together around the brace point like a single cycle sine wave passing through the x-axis. Half the panel moves out while the other half moves in. This is not a doubling of the fundamental resonance, but the pre-existing 2nd harmonic mode.
The half way brace is suppressing the fundamental modal resonance, leaving the 2nd harmonic resonance as the lowest frequency resonance of significant amplitude, but it is not doubling the frequency of the fundamental resonance - that 2nd harmonic resonance was already there without the brace.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.