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.

MEMS Microphone

Sanjay Mukuntha Madhava

Please login with a confirmed email address before reporting spam

I have to do a Simulation of a Microphone to make the Static Analysis. I have the option to do it with either ANSYS or COMSOL. (Am a beginner in both)

Is it possible to develop the 3D model of a microphone Silicon Membrane with Aluminium on both sides?

Regards
Sanjay

10 Replies Last Post 1 mai 2011, 17:38 UTC−4
Ivar KJELBERG COMSOL Multiphysics(r) fan, retired, former "Senior Expert" at CSEM SA (CH)

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 27 avr. 2011, 17:36 UTC−4
Hi

that is a challenging task, in any of the two programmes. For me ANSYS is the engineering tool, while COMSOL is first of all the physicist tool, the one with the best potential to do complex things and to really understand what you are doing. No black boxes. But as both allows many physics, there are quite some aspects to learn.

For your case, the most important is to analyse your model on a peace of paper before you start with a programming tool. Describe the geometry, the interfaces, the physics needed: structural, sound/acoustic, electric ? ... Then define all your boundary conditions required, for each of the physics.

Then start to model your device one physics at a time, I would start with the structure, possibly load it with a pressure to check that it deflects, then add acoustics (or electric if needed) etc

With COMSOL there are many examples in the "model library, its worth to study a few related ones carefully to grab the essentials

--
Good luck
Ivar
Hi that is a challenging task, in any of the two programmes. For me ANSYS is the engineering tool, while COMSOL is first of all the physicist tool, the one with the best potential to do complex things and to really understand what you are doing. No black boxes. But as both allows many physics, there are quite some aspects to learn. For your case, the most important is to analyse your model on a peace of paper before you start with a programming tool. Describe the geometry, the interfaces, the physics needed: structural, sound/acoustic, electric ? ... Then define all your boundary conditions required, for each of the physics. Then start to model your device one physics at a time, I would start with the structure, possibly load it with a pressure to check that it deflects, then add acoustics (or electric if needed) etc With COMSOL there are many examples in the "model library, its worth to study a few related ones carefully to grab the essentials -- Good luck Ivar

Sanjay Mukuntha Madhava

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 27 avr. 2011, 18:24 UTC−4
Thanks for your reply but seriously i dont know where to start. I have to pretty much do it with COMSOL (apparently the university is having issues with licenses for ANSYS)
For my 1st Step, i have to develop a closed rectangular plate silicon membrane with side lengths 1mm x 1mm and Membrane thickness of 10 x10^ (-6) m

Could you please tell me where to start and where to look for examples.

Regards

Sanjay
Thanks for your reply but seriously i dont know where to start. I have to pretty much do it with COMSOL (apparently the university is having issues with licenses for ANSYS) For my 1st Step, i have to develop a closed rectangular plate silicon membrane with side lengths 1mm x 1mm and Membrane thickness of 10 x10^ (-6) m Could you please tell me where to start and where to look for examples. Regards Sanjay

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 27 avr. 2011, 23:11 UTC−4
Hi

Interestingly, I was working on the same subject of modeling MEMS acoustic sensor/microphone today, using the acoustic module from COMSOL.

Overall, it is easy to model the cavity inside the microphone, the challenge remains at modeling the sensor, where there is a very thin perforated backplate, a very thin diaphragm, and a very thin air gap. I think the issues here are: (1) how to model the thin structures of um-level? (2) what type of boundary condition to assign to the diaphragm? (3) how to generate mesh around these thin domains?

Any examples available?

Best regards,

Weiwen
Hi Interestingly, I was working on the same subject of modeling MEMS acoustic sensor/microphone today, using the acoustic module from COMSOL. Overall, it is easy to model the cavity inside the microphone, the challenge remains at modeling the sensor, where there is a very thin perforated backplate, a very thin diaphragm, and a very thin air gap. I think the issues here are: (1) how to model the thin structures of um-level? (2) what type of boundary condition to assign to the diaphragm? (3) how to generate mesh around these thin domains? Any examples available? Best regards, Weiwen

Ivar KJELBERG COMSOL Multiphysics(r) fan, retired, former "Senior Expert" at CSEM SA (CH)

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 28 avr. 2011, 01:28 UTC−4
Hi

with a plate ratio 1:100 you are about at the limit where 3D "shell" physics becomes interesting.

If you have a powerful PC with plenty of RAM (i.e. >>8Gb) then 3D "solid" structural with a fine mesh could do, but you need at least 3-5 elements in the thickness, and then the elements should ideally be about same shape ratio hence at least 300*300*3 = 270'000 elements, with 1.4*number of dependent variables (=3-6) of DoF per element => 1.2 MDof. Yes you need more than >16Gb RAM for that (double precision floating point).
Conclusion it's worth starting in 2D or with shell physics ;)

I agree no direct close example, 1) check the videos on the main COMSOL site 2) but to understand eigenfrequency analysis, try the Model Library - Multiphyiscs - Tuning fork, (for later Joules heating with structural deformation (3 physics electric, thermal + structural the "thermal_actutor _kh_distributed, then you could try to port this example to 3D shell elements instead of 3D solid structural

Thereafter try a few examples in the acoustic (if you have that option), there are two in the main Multiphysics chapter for all users

--
Good luck
Ivar
Hi with a plate ratio 1:100 you are about at the limit where 3D "shell" physics becomes interesting. If you have a powerful PC with plenty of RAM (i.e. >>8Gb) then 3D "solid" structural with a fine mesh could do, but you need at least 3-5 elements in the thickness, and then the elements should ideally be about same shape ratio hence at least 300*300*3 = 270'000 elements, with 1.4*number of dependent variables (=3-6) of DoF per element => 1.2 MDof. Yes you need more than >16Gb RAM for that (double precision floating point). Conclusion it's worth starting in 2D or with shell physics ;) I agree no direct close example, 1) check the videos on the main COMSOL site 2) but to understand eigenfrequency analysis, try the Model Library - Multiphyiscs - Tuning fork, (for later Joules heating with structural deformation (3 physics electric, thermal + structural the "thermal_actutor _kh_distributed, then you could try to port this example to 3D shell elements instead of 3D solid structural Thereafter try a few examples in the acoustic (if you have that option), there are two in the main Multiphysics chapter for all users -- Good luck Ivar

Sanjay Mukuntha Madhava

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 28 avr. 2011, 03:35 UTC−4
Hi,

I was also expecting answer from you for my question. Like i had mentioned earlier, am a beginner and the First task for my Project would be

To develop a closed rectangular plate silicon membrane with side lengths 1mm x 1mm and Membrane thickness of 10 x10^ (-6) m

Could you please tell me where to start and where to look for examples.

Regards

Sanjay
Hi, I was also expecting answer from you for my question. Like i had mentioned earlier, am a beginner and the First task for my Project would be To develop a closed rectangular plate silicon membrane with side lengths 1mm x 1mm and Membrane thickness of 10 x10^ (-6) m Could you please tell me where to start and where to look for examples. Regards Sanjay

Sanjay Mukuntha Madhava

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 29 avr. 2011, 04:01 UTC−4
Can someone help me here please? I just need to know if there are similira examples for this and where to start
Regards
Sanjay
Can someone help me here please? I just need to know if there are similira examples for this and where to start Regards Sanjay

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 29 avr. 2011, 08:31 UTC−4
Frankly speaking, it is a difficult simulation job, given the cavity dimensions ranging from mm^3 to um^3. I am also wondering if Acoustic module or MEMS module alone is sufficient.

The other issues are the boundary conditions assigned to the sensor, in particular, the diaphragm is a vibrating thin layer that needs to be properly modeled.

I am not sure there is anything directly similar from COMSOL. In other words, count on myself. This is fun!

Weiwen
Frankly speaking, it is a difficult simulation job, given the cavity dimensions ranging from mm^3 to um^3. I am also wondering if Acoustic module or MEMS module alone is sufficient. The other issues are the boundary conditions assigned to the sensor, in particular, the diaphragm is a vibrating thin layer that needs to be properly modeled. I am not sure there is anything directly similar from COMSOL. In other words, count on myself. This is fun! Weiwen

Sanjay Mukuntha Madhava

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 29 avr. 2011, 08:45 UTC−4
Wei am sorry but then i started this topic to get help..
Wei am sorry but then i started this topic to get help..

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 29 avr. 2011, 16:36 UTC−4
Sanjay,

I would recommend using 2D axi-symmetry, if possible in the model. This would simplify your analysis significantly. However, if you do need to model 3D, then you might have to define the mesh parameters rather than using Comsol's auto-mesh. The existing loudspeaker models are a good place to begin with. These will help you get familiar with the boundary conditions and settings required for acoustic-structural interaction problems.

Hope this helps.

Mihir
Sanjay, I would recommend using 2D axi-symmetry, if possible in the model. This would simplify your analysis significantly. However, if you do need to model 3D, then you might have to define the mesh parameters rather than using Comsol's auto-mesh. The existing loudspeaker models are a good place to begin with. These will help you get familiar with the boundary conditions and settings required for acoustic-structural interaction problems. Hope this helps. Mihir

Sanjay Mukuntha Madhava

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 1 mai 2011, 17:38 UTC−4
Hi,
Somehow i have managed to create a model using your examples. Can you please check if the approach is right and i have three parameters s,d, p in the global parameters. I have made a parametric sweep for d and is it possible to display all the results simultaneously

Thanks and Regards
Sanjay
Hi, Somehow i have managed to create a model using your examples. Can you please check if the approach is right and i have three parameters s,d, p in the global parameters. I have made a parametric sweep for d and is it possible to display all the results simultaneously Thanks and Regards Sanjay

Note that while COMSOL employees may participate in the discussion forum, COMSOL® software users who are on-subscription should submit their questions via the Support Center for a more comprehensive response from the Technical Support team.