This sounds right to me. I don't really think of them as being different regions.
This reminds me - I should add a "modal overlap" plot feature into my software, to show approximately where the "localized" region begins. I'm worried, though, that I'm adding in too many features that are interesting to people like me, Eric, and Christian, but are going to be confusing and alienating for new people!
Hello Eric,
The math seem ok to me. Nice finding.
From your graph, PETTaLS overestimates the mode frequencies for about 10% for a ratio lambda/thickness = about 7.
For the plates you tested :
Which is in fact already in your first graph!
This error being on the modes and as they have a high density in those ranges, I wonder what is the effect of this error. Other errors in the model like the damping behavior are also possible with a high impact on the SPL.
For now, I am more in the opinion of a use of a simulator in the low to mid frequencies where there are design choices to make. In HF (above some kHz?), the material leads the SPL. There is then almost no more choice than an other material if needed.
Christian
Hello Twocents,
What are the dimensions of the canvas?
About the panel material, I see often in the DIY stores close to me polystyrene for such panels. It is what I tested in directivity. It is a bit lighter than acrylic.
Christian
I am first interested for measurements in 360° to get a view of the front and rear SPL according to the angle. In practice 360° gives double measurements which may show some measurement error (my turning table is a bit archaic). More interesting, it shown also some panels for reasons not well explained have not a symmetric radiation left to right. Reasons might be the spine, the exciter off axis.
So yes, in practice, even if it is longer, 360°. For a simulation, as what impacts the rear (spine, exciter mask, cavity noise...) is not included, 180° front might be enough. 180° and not 90° until we understand the possible asymmetry.
About the target for the best exciter position, I would more look at the directivity with the open back dipole effect included than "only" the on axis SPL. With limits, the FR can be EQed, the directivity comes from the design with no possible correction after.
As an example, I understood from B Zenker's paper and video that a 0.5/0.35 position is a possibility from the modes point of view. I made some tests based on that, small strange peaks occur in the IR probably coming from the symemetry... something more to investigate?
Christian
Yep - I'm most interested in full range - that way I can drill down to specific uses.
@Eucyblues99
Hello Eucy,
In your post #13765 in the main thread, you show the simulation of a 900x600x3mm plywood.
Is the plot an export of PETTaLS? The frequency range seems linear not log?
Which E, rho, Q for this simulation? In my experience, at least with poplar, the coincidence frequency is higher (13kHz across the grain and maybe 10 to 11k in the other direction).
This is curiosity. It doesn't change what you wanted to show.
Christian
Hello Eucy,
In your post #13765 in the main thread, you show the simulation of a 900x600x3mm plywood.
Is the plot an export of PETTaLS? The frequency range seems linear not log?
Which E, rho, Q for this simulation? In my experience, at least with poplar, the coincidence frequency is higher (13kHz across the grain and maybe 10 to 11k in the other direction).
This is curiosity. It doesn't change what you wanted to show.
Christian
Hi Christian
Yep..I need to check the properties...I thought the same thing but didn't go back to check it.
It was supposed to be rho 450 kg/m3, E of 6.6e9, (can't model anisotropy yet), and Q of 8 which is probably low but only affects the spikes anyway.
Yes... It's a direct save from the Sim.
Eucy
Yep..I need to check the properties...I thought the same thing but didn't go back to check it.
It was supposed to be rho 450 kg/m3, E of 6.6e9, (can't model anisotropy yet), and Q of 8 which is probably low but only affects the spikes anyway.
Yes... It's a direct save from the Sim.
Eucy
Hi again Christian
I can't explain or repeat how I managed to get a linear plot - a mystery only Dave can maybe solve.
I was using too low a value for E - should be about 10e9 or 1e10 for the longitudinal direction. However, the coincident freq plots at around 4.2k for the amended E value.
And yes - it doesn't alter the outcome of the fixed vs free SPL level comparison
Eucy
Eucy
I'm not sure why it reverted to a linear frequency scale. Has anyone else ever seen this?
The outside dimensions of the framed canvas are true to A4 size - 296mm x 210mm and the acrylic panel 230mm x 140mm, therefore there is a 30mm (approximately) canvas suspension border around the acrylic panel.
I assume the polystyrene you are referring to is known as PST/HIPS? I have not seen it here in a shop yet, but it is available at specialist suppliers.
I don't think so, but perhaps this is an issue with agreeing upon the correct material parameters for each type of wood? It seems like woods have a broad range of possible densities, Young's Moduli (moduluses?), etc. This is obviously true because the quality and density of wood varies depending on rainfall, temperature, old growth vs. farmed, etc.
Poplar seems like it's fairly stiff but not very dense - closer, actually, to something like gatorboard than to acrylic. I had to reduce the thickness of poplar to 1mm before I got a coincidence frequency above 10 kHz. Maybe a poplar plywood is more dense, or less stiff, and therefore has a higher coincidence frequency?
Not me.
Eucy,Hi again Christian
I can't explain or repeat how I managed to get a linear plot - a mystery only Dave can maybe solve.
I was using too low a value for E - should be about 10e9 or 1e10 for the longitudinal direction. However, the coincident freq plots at around 4.2k for the amended E value.
And yes - it doesn't alter the outcome of the fixed vs free SPL level comparison
Eucy
That is a pretty nice frequency and directivity plot, I think. But I'm inclined to believe that the real Q for a free plywood panel would be over 20. In an early post on this thread I showed an estimate of Q=20 for a free PMMA panel based on impedance measurements. Christian reanalyzed my data with probably a better method, and got 15. I also have measurements of plywood panels (see below) that show considerably less damping than PMMA. I have not yet tried to tease out the Q from those measurements, but inspection of the impedance curves suggests a significantly higher Q for the plywood than for the PMMA. I would guess FFFF plywood to be somewhere between about Q=20 to 50. The only caveat I would add is that for such a large panel maybe there would be some damping from the "acoustic loading".
Eric
Not directly PETTals related, I haven't been able to try it yet as I'm not sure it will work as I only have a Mac and have to use a USB windows stick to boot and use windows programs.
If you want an alternative to canvas for the surround that still allows forward back movement I have had good results with a wide (40/3mm) foam tape surround. Leaving a 10mm gap between panel and frame, similar to a traditional speaker I guess. Doesn't allow the same tension adjustment as canvas though. The bass response is the best I've found this is with 600/400/15mm hd EPS and Xt32 exciter.
With my canvas experiments I had trouble with the edges (panel to canvas and canvas to frame) buzzing at louder volumes.
@JoskaNZ, I don't want to clutter this thread with my canvas experiment either but thank you for mentioning an alternative to canvas surrounds. I have also tried to lightly clamp the panel between soft latex rubber tubes, but it was not that successful. I will share my canvas build on the main DML thread if it proves a worthwhile option.
I just got a macbook so that I can compile the code for iOS - should be coming within a few days!
The sample I have is transparent. It was a glass coming with a frame. Ikea frames mention also polystyrene for their glass.
It is sold here as a substitute to real glass in different thickness.
I heard about PST/HIPS as high impact polystyrene. I will try to see if it is referenced like that in the DIY store.
Christian
HIPS is PS copolymerized with rubber. Probably a little less stiff and with a little more damping than pure PS. Not sure how much though.
Eric