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.
about AC signal calculations in Comsol
Posted 2012/07/26 6:58 GMT-4 Low-Frequency Electromagnetics 4 Replies
Please login with a confirmed email address before reporting spam
Hi all,
I am curious about AC signals in Comsol.
Why do we need to add an analytic function from the functions subtitle to work with an AC signal?
I added a waveform of square wave. If I do not place it in analytic function, the program couldn't calculate the jakobian, thus doesn't work.
Simply, why adding a waveform only does not work?
thanks a lot.
--
physics undergrad
loves photon.
I am curious about AC signals in Comsol.
Why do we need to add an analytic function from the functions subtitle to work with an AC signal?
I added a waveform of square wave. If I do not place it in analytic function, the program couldn't calculate the jakobian, thus doesn't work.
Simply, why adding a waveform only does not work?
thanks a lot.
--
physics undergrad
loves photon.
4 Replies Last Post 2012/08/02 13:55 GMT-4
Please login with a confirmed email address before reporting spam
Posted:
1 decade ago
Hi
there are a few things you should notice about time series and the continuity of your functions.
If you have a "square" BC function your derivative will be discontinuous and this will give trouble for the solver to estimate next step to take (Jacobian error), if you use an "smooth" function it would probalby work OK, and you have a series of functions awailable in COMSOL with "smoothed transition", specially for this.
Then the defalt "automatic" time stepper algorithm, asssumes a solution of a diffusion type, hence for periodic BC conditions, you should use an "intermediate" or strict time stepping (se solver sub-node settings) and give enough time steps to solve correctly your signal, else the solver will skip your steps easily (this was so for <= v4.2 I expect it's still so for v4.3, but COMSOL continuously updates their algoritm, so it might have become cleverer now ;)
--
Good luck
Ivar
there are a few things you should notice about time series and the continuity of your functions.
If you have a "square" BC function your derivative will be discontinuous and this will give trouble for the solver to estimate next step to take (Jacobian error), if you use an "smooth" function it would probalby work OK, and you have a series of functions awailable in COMSOL with "smoothed transition", specially for this.
Then the defalt "automatic" time stepper algorithm, asssumes a solution of a diffusion type, hence for periodic BC conditions, you should use an "intermediate" or strict time stepping (se solver sub-node settings) and give enough time steps to solve correctly your signal, else the solver will skip your steps easily (this was so for <= v4.2 I expect it's still so for v4.3, but COMSOL continuously updates their algoritm, so it might have become cleverer now ;)
--
Good luck
Ivar
Please login with a confirmed email address before reporting spam
Posted:
1 decade ago
thank you, Mr.Kjelberg, I see the reason now.
--
physics undergrad
loves photon.
--
physics undergrad
loves photon.
Please login with a confirmed email address before reporting spam
Posted:
1 decade ago
Hi again,
I have a question regarding the AC potential. I need to create a varying electric field in square wave form.
I chose one boundary as the varying potential, other as the ground, as thinking after one half period the varying potential boundary will be minus magnitude and the ground will be greater than that.
Is it correct? If not, what is the way of creating a varying potential between two boundaries?
thanks!
--
physics undergrad
loves photon.
I have a question regarding the AC potential. I need to create a varying electric field in square wave form.
I chose one boundary as the varying potential, other as the ground, as thinking after one half period the varying potential boundary will be minus magnitude and the ground will be greater than that.
Is it correct? If not, what is the way of creating a varying potential between two boundaries?
thanks!
--
physics undergrad
loves photon.
Please login with a confirmed email address before reporting spam
Posted:
1 decade ago
Hi
if you check the equations you well notice that a GND boundary is simply a electric potential with V=0, so indeed if you have a varying potential in +/- in this way
--
Good luck
Ivar
if you check the equations you well notice that a GND boundary is simply a electric potential with V=0, so indeed if you have a varying potential in +/- in this way
--
Good luck
Ivar