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.

Q factor and Fourier Spectrum

Please login with a confirmed email address before reporting spam

Hello,

I request some guidance on the subject of viewing a Fourier spectrum based on a parametric sweep of a membrane. I have analyzed my model and have gotten eight Eigen frequencies. I want to find the Q factor for all of the Eigen frequencies, but first I would like to see a fourier spectrum of the eight modes' response to a nanoNewton applied point load. I was thinking that just doing a parametric sweep using the frequency analysis solver would be the job to do this. However I have set my parametric values ranging from my first eigen frequency to my last eigen frequency. A probe plot however reveals only one spike at the fundamental frequency and no response from the remaining seven frequencies. Can anyone or Ivar maybe point in the proper direction.

I would really appreciate the help and thank you for your time,
Daniel

4 Replies Last Post 08.12.2010, 16:33 GMT-5

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 07.12.2010, 22:34 GMT-5
I have Comsol 3.5a and 4.0 at my disposal. I forgot to mention
I have Comsol 3.5a and 4.0 at my disposal. I forgot to mention

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 08.12.2010, 02:24 GMT-5
Hi

if you are t use COMSOL for the future, I can only suggest that you get hands on the latest update and use v4.1 (or newer) since the explanations are rather different in v3.5.

I suspect what you are getting is that your frequency sweep solver is "jumping over" the narrow resonances. because the default settings of the iterative solver is to expect a diffusion type exp or log type behaviour reacing a saturation / assymptote and it is by default not set to catch small glitches in an otherwise smooth (close to zero) line.

Therefore try:
1) do an eigenfrequnecy analysis and note down the frequencies (or get them into a table for later use)
2) make a frequency sweep and manually define ranges with som 10 points spaced around each of your frequency peaks and a few points in-between this means many range(,,) range(,,) ... defined one after the other on the same line
3) plot your amplitude in a log scale
4) to get the wrapped phase use the atan2(imag(...),real(...)) where ... us typicall "v" for a 2D horizontal canteliver tip amplitude vibrating vertically (along Y) (there might be a sign correction in there too ( I do nothave COMSOl here now so I cannot try and check, but passing a 2nd order resonance you should get a -180° phase drop at the resonance frequency

You will also note that your physe shift are rather "square" because an eigenfrequency analysis does not take into account any damping (or just a litle articifial mathematical/numerical damping to help the solvers to converge, physics dependent)

Now if you want a "Q factor out you need to add some damping. This is again another story how to best do it, check in the documentations there are a few discussions about correct damping settings. Damping is the most difficult, as its a material and assembly dependent value of rather experimental type, and material damping are mostly negligible compared to assembly induiced damping at least in most structures for ST

--
Good luck
Ivar
Hi if you are t use COMSOL for the future, I can only suggest that you get hands on the latest update and use v4.1 (or newer) since the explanations are rather different in v3.5. I suspect what you are getting is that your frequency sweep solver is "jumping over" the narrow resonances. because the default settings of the iterative solver is to expect a diffusion type exp or log type behaviour reacing a saturation / assymptote and it is by default not set to catch small glitches in an otherwise smooth (close to zero) line. Therefore try: 1) do an eigenfrequnecy analysis and note down the frequencies (or get them into a table for later use) 2) make a frequency sweep and manually define ranges with som 10 points spaced around each of your frequency peaks and a few points in-between this means many range(,,) range(,,) ... defined one after the other on the same line 3) plot your amplitude in a log scale 4) to get the wrapped phase use the atan2(imag(...),real(...)) where ... us typicall "v" for a 2D horizontal canteliver tip amplitude vibrating vertically (along Y) (there might be a sign correction in there too ( I do nothave COMSOl here now so I cannot try and check, but passing a 2nd order resonance you should get a -180° phase drop at the resonance frequency You will also note that your physe shift are rather "square" because an eigenfrequency analysis does not take into account any damping (or just a litle articifial mathematical/numerical damping to help the solvers to converge, physics dependent) Now if you want a "Q factor out you need to add some damping. This is again another story how to best do it, check in the documentations there are a few discussions about correct damping settings. Damping is the most difficult, as its a material and assembly dependent value of rather experimental type, and material damping are mostly negligible compared to assembly induiced damping at least in most structures for ST -- Good luck Ivar

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 08.12.2010, 15:38 GMT-5
Ivar,

Thank you very much, I appreciate your assistance. I have one more question that you may be able to guide me towards the answer's direction.

With regards to eigen frequency damping, I have looked at the thermoelastic damping tutorial in the MEMS model library and I tried to implement this method into my design however the solution takes forever to converge. The process suggests that I set a heat generation equation to couple to my eigen frequency analysis, it however says to set my boundary conditions with regards to conduction module as 'thermal insulation'. Is this because I am setting a heat generation for the subdomain and therefore this subdomain setting overrides my boundary settings? It seems to me that it would make more sense to set the boundary conditions of interest to heat flux and just set a convective coefficient and a reference temperature. Is this a relatively good process to move forward?

Thanks again for answering my previous question!

Daniel
Ivar, Thank you very much, I appreciate your assistance. I have one more question that you may be able to guide me towards the answer's direction. With regards to eigen frequency damping, I have looked at the thermoelastic damping tutorial in the MEMS model library and I tried to implement this method into my design however the solution takes forever to converge. The process suggests that I set a heat generation equation to couple to my eigen frequency analysis, it however says to set my boundary conditions with regards to conduction module as 'thermal insulation'. Is this because I am setting a heat generation for the subdomain and therefore this subdomain setting overrides my boundary settings? It seems to me that it would make more sense to set the boundary conditions of interest to heat flux and just set a convective coefficient and a reference temperature. Is this a relatively good process to move forward? Thanks again for answering my previous question! Daniel

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 08.12.2010, 16:33 GMT-5
Hi

yes there you are touching a delicate task, I havent tried it out in V4, and I do not (yet) feel very comfortable with the thermoeleastic dissipation analysis,

I prefer to add some structural damping either via the linear material subnode "Damping" check the doc for the different options to select, or explicitely as a damping factor*lambda equation. The first one is the easiest as everything is readily set-up
--
Good luck
Ivar
Hi yes there you are touching a delicate task, I havent tried it out in V4, and I do not (yet) feel very comfortable with the thermoeleastic dissipation analysis, I prefer to add some structural damping either via the linear material subnode "Damping" check the doc for the different options to select, or explicitely as a damping factor*lambda equation. The first one is the easiest as everything is readily set-up -- Good luck Ivar

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.