Discussion Forum

Hi

typically an external load, such as an external Boundary force for a structural case, you give an amplitude and the solver considers this amplitude as oscillating at the frequency "freq"

--
Good luck
Ivar

Thank you very much!!!!!!!!!!!

Hi,
Can I ask you another question? I am confused of resonance frequencies and the eigenfrequencies. I thought they were referred to the same thing. Are they? If they are the same thing then why for a two mass and spring system, it has two resonance frequency( one symmetric and one antisymmetric) , but if I simulate on comsol with the eigenfrequency analysis, it gives me a lot of eigenfrequencies? Thanks a lot??????

Hi

I agree in 1D two spring masses attached rigidly at one end should give two modes only, (twist modes excluded) but in 2D or 3D you should see a few more for the different directions and new DoFs

in V4.1 you can normalise the modes to give you their participation mass values, which allows you to see in which direction x,y,z the modes are giving their energy, unfortunately COMSOL does not solve for the three angles, that are also of interest

--
Good luck
Ivar

Hi Ivar,
Thank you very much!! But if I make a frequency response analysis, the value I get is different with the eigenfrequency analysis, so I guess the frequency response is the solution for 1D? Is that correct? Thanks a lot!

Hi

An eigenfrequency analysis is taking a model, stripping external forces/loads, linearising the model and searching for the eigenmodes of the equation sets

A frequency response is obtained by applying a sinus excitation, giving an amplitude, a phase and then scanning the frequency. If you load is applied such to excite the fundamental modes you should get a peak for your previously calculated eigenfrequencies (I always use an eigenfrequency analysis to define my frequency scan values). But if your load is oriented in another direction you might get couplings and excite other "higher order" modes

--
Good luck
Ivar

Hi,
Can I ask you another question? I am confused of resonance frequencies and the eigenfrequencies. I thought they were referred to the same thing. Are they? If they are the same thing then why for a two mass and spring system, it has two resonance frequency( one symmetric and one antisymmetric) , but if I simulate on comsol with the eigenfrequency analysis, it gives me a lot of eigenfrequencies? Thanks a lot??????

Hi,

Eigenfrequencies of a system are related to the undamped, free motion of the system. Resonance frequencies are related to the damped, and possibly forced vibration of the system. Resonance frequencies can be combinations of eigenfrequencies depending on your excitation.

On this note, I have been wondering whether one can actually get the frequency response with the sine sweep (Frequency Domain Study). Typically, frequency response of mechanical system, such as a musical instrument, is measured with an impulse excitation, not sinusoidal excitation. Although, in room acoustics, the sine sweep is frequently employed. Anyone else wondering about this?

Hi

I agree that there is much more to say here, but normally one consider eigenmodes as fundamental entities independently of their way of excitation. When you excite with a sinus you enter the energy in one ray so you excite ideally only 1 mode at the time (or some harmonics). If you enter with an impulse you should do the correlation of the impulse frequency spectrum with the eigenfrequency spectrum. With different damping you then have different decay times (which is another story).

So if you talk about a music instrument you should optimise its design to give you just a few excited notes, and damp any parasitic mode, the its the way you hit tit that will excite more certain rather than other modes

This is then related to a PSD (power spectral density) of an excitation what are all the true hypothesis underlaying that allows us to use a PSD as a representative way to run a frequency sweep to excite "truly" a system. Most engineers forget them quickly (and violate them easily) but you can find books full on signal processing around these issues.

As an engineer I still use and accept a PSD as a good entry for frequency sweeps and analysis of structures, remaining with the issue how to simulate the true damping of a system such to have representative peak heights and energy per mode distributions


--
Good luck
Ivar

Hi Ivar,
Thanks again for the replies. Again I have another question about the frequency response analysis. I make a plot of the phase of a certain point in my structure and use the postprocessing domain plot to plot the phase of that point. However the result is different with what I expected. It only shows the phase of 180 and 0 degree, what I expected is the phase goes through all the values between -180 to 180 degrees of my structure under a certain force. Can you give me some ideas? Also I want to make a plot a typical bode plot of my frequency response, if I use matlab, how can access my displacement and phase ( Out/in ) values so that I can make a bode plot in matlab?

Hi

what (which physical properties) is governing the phase "smoothness" ?

In my way of expression the phase "domain" is the loss "domain", if you add some losses i.e. some Rayleigh or general type viscous damping you will see the transitions come (as well as you will get imaginary modes frequencies

Tray it out, perhaps on a simple cantilever


--
Good luck
Ivar

Hi,
I am a little confused of what you said about the " phase domain" related to the loss domain. I thought the phase domain related to the passive element of my circuit like C, L in electrical domain and mass, spring in the mechanical point. Since my circuit is equivalent to the general mass spring system, I thought it would give me phase difference, am I wrong about that? Thanks!

Hi

sorry to go too quick (I'm stuck in a structural frequency sweep model just now so I'm biased) the smooth "round" phase transitions will appear if you add some structural damping in your MEMS module (if I understand you correctly you have some structural MEMS, or are you fully in ACDC ? in which case probably some electric damping, i.e. impedance loss, is needed)

--
Good luck
Ivar

Hi, resurrecting a dead thread. You mention applying an external boundary force to produce a frequency domain. How would this apply to a periodic structure? Would comsol replicate the force along the position vector and apply it to every unit cell?

Essentially, how would one go about producing dispersion curves for a periodic structure?

Thanks!

Matt

Hi ivar and others,

I have simulated a mems capacitive accelerometer. Now I want to find its cross axis sensitivity (how much it will displace in other direction when it is accelerated in one direction?). I dont know how to find it in COMSOL. Waiting for your kind response.

Thanks.