Discussion Closed This discussion was created more than 6 months ago and has been closed. To start a new discussion with a link back to this one, click here.
Modelling loudspeaker surrounds; What is your approach?
Posted 2022/07/11 11:44 GMT-4 Structural & Acoustics, Materials, Modeling Workflow 2 Replies
Please login with a confirmed email address before reporting spam
Hi there,
I am curious to see what other people's approach is towards modelling their transducers surrounds in fluid-strucutre-interaction problems, more especially concerning damping, as that has been somewhat of an unknown to me so far (and has led to uncertainties).
Usually, I start with the data I have about the textile or rubber material used for the surround, which I use to impose the correct density for that domain. An equivalent surround's Young modulus is inferred by tuning the model until a static deflection test simulation matches real measurements. Both of these procedures work well to reproduce the measured resonance frequency. So far, so good.
However, frequency domain simulations, more often than not, present huge surround breakups and a very high Q at the resonance frequency, which differs substantially from measurements. At the moment I can only think this might be related to the damping model (the ones available at Solid Mechanics->Linear Elastic Material->Damping) in use. Usually damping is assumed as a constant (traditional) isotropic loss factor, but to no help. The textile materials we use, in particular, are often impregnated with resins that help them achieve good damping characteristics, which would explain the lack of breakups seen in reality, but which lead me to conclude that what they do is far from traditional damping.
Low values of isotropic loss factors seemingly don't change the simulated FR curves all. High values will merely offset the output curves down, not really addressing the breakups as they're expected to.
Exploring what could happen with the rayleigh model, the effects I could see were not really addressing breakups as much as expected, but the overall output of the modelled transducer was very sensitive to the parameters used, in the other hand. I must say I was not able to explore well enough the other available models just yet.
What would be good suggestions to try next in this sense? TIA.
Please login with a confirmed email address before reporting spam
Alexandre,
I think these materials are pretty nonlinear and thus it is diffcult to use the frequency domain. You are also looking at quite large deflection I guess, so linearization may not work. Materials like rubber exhibit hysteresis, this may actually be the dominant dissipation mechanism. It might be worthwhile to look into material models that include hysteresis, but you will necessarily end up in time domain with this. We are not working in the field of nonlinear mechanics, so I am not sure if mechanical hysteresis is available in the nonlinear structural mechanics module. Jiles-Atherton hysteresis is available in electromagnetics and it may be possible to adapt it to mechanics.
Cheers Edgar
-------------------Edgar J. Kaiser
emPhys Physical Technology
www.emphys.com
Please login with a confirmed email address before reporting spam
Alexandre,
I think these materials are pretty nonlinear and thus it is diffcult to use the frequency domain. You are also looking at quite large deflection I guess, so linearization may not work. Materials like rubber exhibit hysteresis, this may actually be the dominant dissipation mechanism. It might be worthwhile to look into material models that include hysteresis, but you will necessarily end up in time domain with this. We are not working in the field of nonlinear mechanics, so I am not sure if mechanical hysteresis is available in the nonlinear structural mechanics module. Jiles-Atherton hysteresis is available in electromagnetics and it may be possible to adapt it to mechanics.
Cheers Edgar
Thanks for the answer Edgar. Actually, the issues I've been seeing took place when modelling tweeters, meaning that the deflections are really small. Regardless of the material, I don't believe it's fair to assume it would be somehow in the non-linear regime, right? This is why I believe the answer might be somewhere in terms of damping rather than non-linearities like those. The good news is that the damping (or whatever exact characteristic this actually is) we see in reality does not have to be so accurately described for the most part. The models without damping already predict the outcome quite well except for the artificial break-ups I am seeing and a high "Q" at the resonance frequency of the driver.