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.
What causes a model to converge at one frequency, but not at another?
Posted May 1, 2016, 7:02 a.m. EDT 3 Replies
Please login with a confirmed email address before reporting spam
I have created a coil and am running it at different turn number and frequencies to verify that the model is outputting correct inductance.
Using "magnetic and electric fields" (mef) and a frequency domain study, I am running the following parametric sweeps:
N=5,10,15 turns
f = 10, 50, 100, 500 Hz
However, after running N=5 through all frequencies, N=10 does not converge at f=500 Hz.
Does anyone have any tips on how to tweek the solver?
I am currently running FGMRES and it "slows down" between a convergence of 2-5 while SSOR iterates to 1 000 000 and it gives up.
--
:)
Using "magnetic and electric fields" (mef) and a frequency domain study, I am running the following parametric sweeps:
N=5,10,15 turns
f = 10, 50, 100, 500 Hz
However, after running N=5 through all frequencies, N=10 does not converge at f=500 Hz.
Does anyone have any tips on how to tweek the solver?
I am currently running FGMRES and it "slows down" between a convergence of 2-5 while SSOR iterates to 1 000 000 and it gives up.
--
:)
3 Replies Last Post May 2, 2016, 2:57 p.m. EDT
Please login with a confirmed email address before reporting spam
Posted:
9 years ago
Hi
I believe you do not give us enough info to give back reasonable guesses.
One possibility, taken from the blue, with N=10 and f=500Hz you are hitting a LC resonance without any resistance to damp the signal hence a numerical singularity, but it could be fully other things, like a mesh density issue ...
--
Good luck
Ivar
I believe you do not give us enough info to give back reasonable guesses.
One possibility, taken from the blue, with N=10 and f=500Hz you are hitting a LC resonance without any resistance to damp the signal hence a numerical singularity, but it could be fully other things, like a mesh density issue ...
--
Good luck
Ivar
Please login with a confirmed email address before reporting spam
Posted:
9 years ago
Hi
I believe you do not give us enough info to give back reasonable guesses.
One possibility, taken from the blue, with N=10 and f=500Hz you are hitting a LC resonance without any resistance to damp the signal hence a numerical singularity, but it could be fully other things, like a mesh density issue ...
--
Good luck
Ivar
Greetings, Ivar. It is an honor to have you in my thread :)
My fault is almost bordering idiotic, but I forgot to disable the study "coil geometry analysis" after disabling the multiturn coil.
After that it worked out nicely, and I got good results:
Physical inductor: measured by impedance method at 128 uH
Modelled replikaka in comsol: 126.55 uH
I hope it's not pure dumb luck
--
:)
Attachments:
Please login with a confirmed email address before reporting spam
Posted:
9 years ago
Hi
It's never to late to learn ;)
and we learn more from our "errors" often, than from our successes, so continue like that :)
--
Have fun COMSOLing
Ivar
It's never to late to learn ;)
and we learn more from our "errors" often, than from our successes, so continue like that :)
--
Have fun COMSOLing
Ivar