Dave,
Was the acoustic radiation map you shared for the PIP case above coincidence? I guessing it was, is that right? It may be a purely academic question, but what if it was not above coincidence? It's entirely possible (at least in theory) that for a composite structure with a core of relatively low shear modulus, that the coincidence is never reached at any frequency. So for such a structure, I'm curious what the PIP radiation pattern would look like then. As always, no pressure if it is not easily tested.
Thanks,
Eric
@EarthTonesElectronics
Hi Dave
I just did a comparison run on the example acrylic panel you used in your video Part 3 - 0.2x0.3x.006 CCCC using the two exciters available (not the 32dia you used) and I noticed something odd in the critical freqs. Running the 19mm to 20kHz shows no exciter effect. Using the 25mm gives a critical freq of 3848 Hz, ie. lower than the coincident freq of 6098, and yet your sim using the 32mm exciter gives a critical freq of 8322 Hz - above the coincident. This doesn't seem to make sense (unless I mucked up something of course)
Also - I'm wondering what practical value lies in showing the average horizon level FR, (where some balanced modes radiate strongly but show nil in this plot)
I've also been trying to get some info on the coincident frequency analysis of anisotropic plates but keep running into significant paywalls.
Why am I doing this? Because I'm having difficulty conceptualising the effects of differing horiz and vert stiffnesses creating 2 separate coincident frequencies - surely the anisotropy creates amore complex radiation situation and hence trying to match PeTTaLS iso output (in our version at least) with Christian's analysis of plywood is fallacious??
Eucy
Hi Dave
I just did a comparison run on the example acrylic panel you used in your video Part 3 - 0.2x0.3x.006 CCCC using the two exciters available (not the 32dia you used) and I noticed something odd in the critical freqs. Running the 19mm to 20kHz shows no exciter effect. Using the 25mm gives a critical freq of 3848 Hz, ie. lower than the coincident freq of 6098, and yet your sim using the 32mm exciter gives a critical freq of 8322 Hz - above the coincident. This doesn't seem to make sense (unless I mucked up something of course)
Also - I'm wondering what practical value lies in showing the average horizon level FR, (where some balanced modes radiate strongly but show nil in this plot)
I've also been trying to get some info on the coincident frequency analysis of anisotropic plates but keep running into significant paywalls.
Why am I doing this? Because I'm having difficulty conceptualising the effects of differing horiz and vert stiffnesses creating 2 separate coincident frequencies - surely the anisotropy creates amore complex radiation situation and hence trying to match PeTTaLS iso output (in our version at least) with Christian's analysis of plywood is fallacious??
Eucy
Eucy,
I know you asked Dave, but I will share my two cents. I think you are correct that we simply won't be able to match Christian's analysis with an isotropic model. As you know, plywood is anisotropic, so just from that there should be two different values for coincidence frequency depending on the orientation of the panel. But there is even more to it than that. The coincidence frequency depends al lot on the shear stiffness of each of the core layers, their thickness, and their location in the the layup. And the the shear stiffness of each of the core layers depends very strongly on their orientation. So predicting the coincidence frequency of plywood (or frankly any multilayer composite) is pretty hard without doing some pretty thorough measurement (or modeling) of all of the relevant elastic constants.
Eric
@EarthTonesElectronics
Dave,
One thing I have been meaning to ask is this: How much harder is it to model a panel that is only "self baffled" than one that is infinitely baffled? I think most of us are making panels that are not embedded in a wall or ceiling, but rather panels that are standing substantially free without any additional baffle.
Eric
Dave,
One thing I have been meaning to ask is this: How much harder is it to model a panel that is only "self baffled" than one that is infinitely baffled? I think most of us are making panels that are not embedded in a wall or ceiling, but rather panels that are standing substantially free without any additional baffle.
Eric
Eric... That's what I can't rationalise... Radiation is polar and continuous, so I can't see edge radiation on one axis becoming quiescent then suddenly just popping around to the other axis at the second frequency. It must surely be a more complex situation than that, particularly when edge supported.
Eucy
Last edited:
I thought that the model was effectively open baffled?? Or does it have infinitely long side walls??
Last edited:
Of course, but once completed, the panel has overall characteristics which can be measured and modelled
Eucy
I must say that I don't know for sure. But I think you are likely correct that it can't jump suddenly from one frequency to another. I would suspect rather that it changes gradually changes between one and the other as the angle changes. Imagine that the you are standing at 80 to 90 degrees off axis, while someone spins the panel in its own plane from vertical to horizontal or vice-versa. I suspect that the coincidence frequency would vary as a relatively smooth function as the panel is rotated from vertical to horizontal.
Eric
I think yes, as long as the model takes into consideration all the necessary characteristics. Which may include as many as 6 elastic constants (2 elastic moduli, 3 shear moduli, and one poisson's ratio) for an orthotropic material. I think there may be one or two of these are are less important, but at least the 2 elastic moduli and shear moduli are critical, I suspect.
Eric
I don't love this aspect of the model, but I think it simply boosts the simulated SPL at low frequencies. As long as you understand that, you can kind of take it into account. But I'd be interested to hear Dave's perspective on this.
Eric
Nup - It appears to do the not much at all but opposite - Blue 50mm box - purple 500m
Eucy
So while chopping a tree today I mused that a relationship for the coincident frequency for plywood should be able to be simply developed by preparing 5 test strips.
One to be cut longitudinally, one transverse, and 3 biased at 30,45 and 60 degrees. These to be loaded and measured to calc E for the varying layer directions.
Applying the std formula should provide sufficiently accurate data points to allow curve fitting and generate an equation for a good estimation of fc at any angle for that material.
Then what is the question... Will knowing that be of any real use??
Eucy
One to be cut longitudinally, one transverse, and 3 biased at 30,45 and 60 degrees. These to be loaded and measured to calc E for the varying layer directions.
Applying the std formula should provide sufficiently accurate data points to allow curve fitting and generate an equation for a good estimation of fc at any angle for that material.
Then what is the question... Will knowing that be of any real use??
Eucy
It's a little more complicated because plywood is multilayer, so you need a means (model) to see how they behave in a stack. And it's not just E, but also G (shear modulus). But that said, yes, it's possible with a good enough model to estimate fc at any angle.
And the value of that would be that it would let you predict at what frequency you would have high off-axis SPL, like what is circled in red. Note, this simulation is in not intended to represent specifically plywood. Rather, it's just a panel with properties that illustrates clearly the effect of coincidence frequency on the radiation pattern.
I'm kind of on the fence myself about how important it is in a practical sense with most typical panels. But I think it is very interesting problem to understand and solve from a technical/academic perspective.
Eric
Eric..I don't follow that... Provided the min wavelength is about 6 times the panel thickness, the fc eqn relies on E, density and thickness. I believe this in the eqn used in PeTTaLS.
Why is stack behaviour involved? I'm not disassembling the plywood, it's a given, completed entity.
Eucy
Yes, it was... here's a test using 0.3mm aluminum, which has a coincidence frequency of around 35kHz.
