Perhaps counter-intuitive but CLD should work better with a thinner layer.
(This increases the shear on the lossy material)
So the silicone buttons seems like extra complication with no real benefit, as far as I can see.
Also, silicone is not very lossy, so Bostik V60 seems less than ideal acoustically.
Of course there are always complications in the real world, like the extra mass of the thicker lossy layer and perhaps small alteration in the stiffness of the panel.
So I don't claim to be authoritative, because I don't have test data on my own panels yet - but I would recommend no spacers, keep the layer thin.
Also uses less damper material, to save your money.
Best wishes
David
Last edited:
Haha, true
(He used Bostic V60). Well, Bostic is not available here but i can source green glue and copy the method with silicone buttons to control the height. Is green glue a good candidate for this kind of job in terms of vibration dampening properties and its thickness/ease of application?
I am thinking about 2-3mm thickness.
Nooooooooooooooooooooooooooo
I built a CLD cabinet with green glue. Green glue never hardens and its consistency is similar to petroleum jelly (vasoline).
The result was the inner and outer MDF layers were decoupled and thus not very stiff. With the low stiffness and the added mass of the green glue, the panel resonant frequency was much lower than usual but not well damped by the knuckle test.
I used a fairly thick layer of green glue, parallel beads from a caulk gun, may have only had 50% coverage to let air escape, was years ago. You could try 1-2 mm layer, spreading uniformly will be a challenge.
I couldn't remove the stuff as it gets into the pores of MDF or ply, so I have two tower-speaker sized workbench stands in my basement.
Two possibilities
1) the EAR corporation makes blue viscoelastic sheets that seem promising. It's a very stretchy rubber - I couldn't figure what adhesive would reliably bond it to MDF or ply. You may have better ideas. The adhesive needs to be stiffer than the sheeting so the sheeting stretches - I worried this rules out 3M spray adhesives but I didn't test them.
E-A-R Aearo Technologies LLC - noise, vibration, shock, damping, sound, control
2) you could try bonding 12mm ply to 3 or 6mm aluminum using the Sonic Barrier brand loaded vinyl viscoelastic sheeting sold by Parts Express to damp the metal panels in car doors.
I used this sheeting between 12mm MDF and thin bathroom tile and it worked well. Used an acrylic loaded mortar to bond the tile to the vinyl layer (mortar sets even in an enclosed space). tile is on the cabinet inside, added after cabinet was glued together. There is a link to this work in post #76.
This sheeting had no benefit if both layers were MDF, ply, or a mix of the two. Seems designed for stiffer materials.
CLD is quite complicated, the optimum thickness is not easy to calculate, but I expect your layer was too thick.
The optimum depends on the stiffness of the other layers so it varies with their thickness and material.
It also depends on the stiffness properties of the viscoelastic material.
As I already stated, I believe that for typical speaker use a thinner layer is more effective, it's certainly stiffer.
I haven't done the analysis or experiments to confirm this yet but your experience makes sense.
Best wishes
David
Last edited:
Your second reply contradicts your first, but I generally agree with you that the optimal layer is hard to determine and that thicker is NOT better. But too thin would not be good either. The ideal is somewhere in between.
I didn't write as clearly as I should have.
My first post was to emphasise the point that, as you say but contrary to expectation "thicker is NOT better", and that a thinner layer actually can increase the shear and therefore the losses.
In practice the layer is probably too thick and so "thinner is better"
But that is a bit over simplified so I added the second to expand the point, I think we are accord.
Have you ever actually calculated the optimum thickness for typical speaker construction?
I haven't yet, but it should be easy to estimate the stiffness for moderate thickness plywood.
More difficult for the visco-elastic layer but typical values for polyurethane or rubber should be in the ballpark.
Or have you any experimental data for the same construction with different visco-elastic layer thickness?
Best wishes
David
As luck would have it I am/was in the process of doing this with the intention of writing an article for the web. It is a subject that comes up in discussion here quite regularly but without any trustworthy evidence based conclusions being reached. I also want/need the information for the tweeter/midrange cabinet I am supposed to be building for the drivers sitting on the table downstairs nagging away at me.
Unfortunately having found sets of open software on the web that claim to handle the sound radiation from a cabinet and the nonlinear vibration of a cabinet (high loss viscoelastic materials have properties that vary strongly with frequency) and spending a significant amount of time and effort learning how to use it and writing scripts to test it with reasonably representative cases I have come to the conclusion that some of the aspects we need for this type of problem (nonlinear vibration is fairly rare and not widely supported) are incompletely implemented to the extent they are too difficult to use for real work.
I am currently writing software for the sound radiation part which is going OK and should be tested and usable in a week or two. I had looked at improving the open source code but it looked too poorly written to be worth getting involved and extending.
The open FE software for the cabinet vibration is large and has been developed for decades to support engineering simulations. It is good code and well worth improving/fixing/extending the handling of nonlinear viscoelastic materials. Unfortunately it is in French which I don't read and the core Fortran routines are surrounded by layers of "helpful" computer scientist C++ code. Not sure whether to persevere or lash up some limited FE software for the task. Whatever it is delaying things but I am currently still plodding on.
One of the more significant papers that helped me understand damping, extensional and constrained layer, was the chapter "Applied Damping Treatments" by David I. G. Jones in the "Harris' Shock and Vibration Handbook". The chapter is a basic engineering text, so skim over what you don't need (or don't comprehend). However the following three graphs contained in the chapter are enlightening. You don't have to be an engineer to interpret the curves, just a little diligent:
1) Temp-Frequency Nomogram for Butyl Rubber Composites
Illustrates the effect of temperature on stiffness and damping factor on viscoelastic damping materials, butyl rubber in this case.
2) Graph of n/n2 vs h2/h1 For a Free Layer Treatment
Illustrates the impact of damping layer thickness and damping layer loss factor on a simple extensional damping system. Family of curves shown for different thickness ratios.
3) Typical Plots of n/n2 Versus Shear Parameter G
Illustrates the relationship between shear layer thickness and damping factor versus system constrained layer damping effectiveness. Family of curves shown for different thickness ratios.
Previous editions of this chapter are accessible on the internet.
1) Temp-Frequency Nomogram for Butyl Rubber Composites
Illustrates the effect of temperature on stiffness and damping factor on viscoelastic damping materials, butyl rubber in this case.
2) Graph of n/n2 vs h2/h1 For a Free Layer Treatment
Illustrates the impact of damping layer thickness and damping layer loss factor on a simple extensional damping system. Family of curves shown for different thickness ratios.
3) Typical Plots of n/n2 Versus Shear Parameter G
Illustrates the relationship between shear layer thickness and damping factor versus system constrained layer damping effectiveness. Family of curves shown for different thickness ratios.
Previous editions of this chapter are accessible on the internet.
Last edited:
I will be interested to see your results, I had in mind a rather different approach to try.
My idea was to take a typical panel say 600 mm span in 12mm ply laminated thru a constrained layer X mm thick to another 12 mm ply panel.
Pick realistic typical values for losses and stiffness in the constrained layer.
Differentiate the loss as a function of X, find the maximum.
First year calculus problem, basically.
Should provide a clue for "sensible" values.
Solve it with the skin stiffness as a parameter to learn more about how to optimise the problem.
I have seen this done in text books but never followed it up, their solution looked more complicated than I wanted.
I did a similar problem to find the optimum core thickness for a sandwich panel (a stiff skinned honeycomb core for yachts, or speakers).
Turns out there is a nice, simple solution for maximum stiffness at a specified mass.
The problems are not too dissimilar so I suspect there may be a simple answer for the maximum loss with specified skin stiffness.
Simplicity probably requires a few assumptions that are reasonable in the context of speakers but that the text books don't/can't make.
Best wishes
David
Thank you, I will track this down.
Just what I wanted, you posted as I wrote my previous reply.
Best wishes
David
Dan,
Loudspeaker enclosure damping falls under the 'structural damping' field due to the thickness of the walls (and associated levels of stiffness, mass, damping, energy, and transmission).
Here's a link to a constrained layer damping application, though a thin layer application. The link may take a little while to load.
http://www.acoustics.asn.au/conference_proceedings/AAS2008/papers/p69.pdf
Loudspeaker enclosure damping falls under the 'structural damping' field due to the thickness of the walls (and associated levels of stiffness, mass, damping, energy, and transmission).
Here's a link to a constrained layer damping application, though a thin layer application. The link may take a little while to load.
http://www.acoustics.asn.au/conference_proceedings/AAS2008/papers/p69.pdf
Having done a bunch of simulations for the woofer cabinet where damping is relatively unimportant (making a linear model for damping sufficient) little reads across from what you would deduce from a set of results for a panel to what is important in the full cabinet.
If you opt for a constraining layer the same thickness as the structural layer (optimum with same material for both) then most of the thickness is not carrying load. This is unlikely to be wise in a well balanced design.
How well does a box in a box perform? Should the baffle be maximum stiffness and little damping? If the constraining layer is internal how to go about bracing? Which cabinet modes actually need damping? Etc... These types of questions need a model of the full cabinet and radiated sound and not just some qualitative insights from considering a panel in isolation.
In truth one could answer most questions with a linear model plus a bit of fiddling but having started with the expectation of working with a full nonlinear forced frequency response over the range of the main energy containing modes of the cabinet I shall persevere for a while yet. What will force me to revert to the back up plan is getting too close winter and an overly cold shed.
If so then I really look forward to your results.
But I haven't seen an optimisation of even a simple panel, let alone a full cabinet, so I plan to start simple, try to understand the components before I combine them.
Not sure I understand your point here, will think it over.
In fact for my personal project I have an 18 mm ply cabinet and a 3 mm aluminium skin for the front panel and side transitions.
I haven't laminated the skin yet, wanted to do the maths first.*
But I decided a symmetrical example would be a nice, simple test case, and useful for builders of typical speakers.
*Also because it's not easy to bend a 3 mm sheet of ally 1220 wide.
Best wishes
David
Last edited:
Yes: appropriate damping material for application following stiffness and variation with frequency rather than peak loss, use of linear damping model (I think and wishing to use an even simpler unphysical model), measurements and predictions not plotted on the same graph,... but nonetheless still answering some of the questions the project sought.
This looks a better approach to me for a practical design leading to a wall thickness of less than 25 mm rather than 40 mm or so.
For loudspeaker no. But I did a lot of work on CLD back at Ford and so I have a feel for what works and what doesn't. Back then we tested the effects using cantilever beams and the response at the end from an excitation at the base. From this one can graph out the pertinent variable.
I doubt that the cabinet vibrations are nonlinear. Being frequency dependent is very different than being nonlinear.
Finally, I have found that internal damping structures (braces etc.) are at least as effect, if not more effective, than CLD. But no one seems to be talking about this.
