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.
I need help about AC/DC Electrostatic field
Posted 05.07.2012, 22:43 GMT-4 Low-Frequency Electromagnetics 7 Replies
Please login with a confirmed email address before reporting spam
Dear Ivar and everyone,
I am simulating sensor..
In my case, i use: copper for electrode and silicone for dielectric material.
i choosed Electrostatics field with Ground for 1 Rx Electrode and Terminal for Tx Electrode.
my mesh size is extremely line.
i choosed Staionary in study 1.
when i solved this error occur:
"
Failed to find a solution.
System matrix is zero.
Returned solution is not converged. "
this is my file.
Thanks
Best regard,
Dat
I am simulating sensor..
In my case, i use: copper for electrode and silicone for dielectric material.
i choosed Electrostatics field with Ground for 1 Rx Electrode and Terminal for Tx Electrode.
my mesh size is extremely line.
i choosed Staionary in study 1.
when i solved this error occur:
"
Failed to find a solution.
System matrix is zero.
Returned solution is not converged. "
this is my file.
Thanks
Best regard,
Dat
Attachments:
7 Replies Last Post 29.07.2012, 14:52 GMT-4
Please login with a confirmed email address before reporting spam
Posted:
1 decade ago
Hi
check that you mesh the full model, you have not a mesh for the air (nor is it defined in the materials)
--
Good luck
Ivar
check that you mesh the full model, you have not a mesh for the air (nor is it defined in the materials)
--
Good luck
Ivar
Please login with a confirmed email address before reporting spam
Posted:
1 decade ago
Hi Ivar,
Thank you so much,
I already solved that problem, i hope to meet u in another day.
Best regrad,
Dat
Thank you so much,
I already solved that problem, i hope to meet u in another day.
Best regrad,
Dat
Please login with a confirmed email address before reporting spam
Posted:
1 decade ago
hi ivar
I asked u in other thread about ac field in time dependent . i used ac/dc in electrostatic nd applied 10*sin(2*pi*f*t).
in terminal settings. t varied between 1-10 sec. but i dont know what i did wrong . potential came in negative and field distorted. in DC it looks fine.may be i am doing something basic wrong. I m new to comsol so plz look into this.
I asked u in other thread about ac field in time dependent . i used ac/dc in electrostatic nd applied 10*sin(2*pi*f*t).
in terminal settings. t varied between 1-10 sec. but i dont know what i did wrong . potential came in negative and field distorted. in DC it looks fine.may be i am doing something basic wrong. I m new to comsol so plz look into this.
Attachments:
Please login with a confirmed email address before reporting spam
Posted:
1 decade ago
Hi
your DC looks proably reasonable, but the AC is basically "0", no ?
if you use a time series, for 10 seconds or more, what kind of frequency do you have ? if its in the MHz domain, you need to sample (time step) in the sub microsecond level, henceit will calculate for ages perhaps 10E6 times ?
Are you sure its not better to use a harmonic solver (frequency domain) ? and solve for the amplitude, than use a transient solver in time series ?
If you use a time solver, be sure you use a strict or intermediate time stepping, with a time sequence short enough to resolve each frequency oscillation
--
Good luck
Ivar
your DC looks proably reasonable, but the AC is basically "0", no ?
if you use a time series, for 10 seconds or more, what kind of frequency do you have ? if its in the MHz domain, you need to sample (time step) in the sub microsecond level, henceit will calculate for ages perhaps 10E6 times ?
Are you sure its not better to use a harmonic solver (frequency domain) ? and solve for the amplitude, than use a transient solver in time series ?
If you use a time solver, be sure you use a strict or intermediate time stepping, with a time sequence short enough to resolve each frequency oscillation
--
Good luck
Ivar
Please login with a confirmed email address before reporting spam
Posted:
1 decade ago
hi ivar
its a great help ur doing here for beginners like me . It was the simplest and straight forward method for me in COMSOL to apply ac field in time dependent method. I would do it in frequency dependent if i could . i know where to apply frequency nd potential nd time. so should i vary frequency using sweep. OR is there any other way to do it properly.
sir , if u could share a simple model applying AC filed in frequency solver it would be a great help to new users like me.
thank you
vipin
its a great help ur doing here for beginners like me . It was the simplest and straight forward method for me in COMSOL to apply ac field in time dependent method. I would do it in frequency dependent if i could . i know where to apply frequency nd potential nd time. so should i vary frequency using sweep. OR is there any other way to do it properly.
sir , if u could share a simple model applying AC filed in frequency solver it would be a great help to new users like me.
thank you
vipin
Please login with a confirmed email address before reporting spam
Posted:
1 decade ago
Hi
I time stepping solver is good for transients or for some special cases of regular AC oscillations, but it requires quite some time to solve, as you need to go through each time steps, and with small steps sufficient to solve each wave oscillation.
The other way is to say you have a steady sate AC oscillation, you performa Laplace development (harmonic development) of your equations and you define the frequency, or rather omega = 2*pi*freq.
This is what happens when you choose a harmonic solver case (check the equations that COMSOL proposes, you will se the "t" change to "omega" ) in this mode you define amplitudes for the BCs and you do not give any "sin(2*pi*f*t)" equation at all. COMSOL then requests you to define the frequency, or a sweep over a series of frequency, in the harmonic solver node under "frequency" entry (use 100[Hz] or range(10,10,100) or whatever you want
A harmonic solver case is easier to solve and reuires less RAM, but assumes a steady state AC oscillation apart some special new special solver cases as the harmonic - transient case, see the doc, its worth to read it carefully
--
Good luck
Ivar
I time stepping solver is good for transients or for some special cases of regular AC oscillations, but it requires quite some time to solve, as you need to go through each time steps, and with small steps sufficient to solve each wave oscillation.
The other way is to say you have a steady sate AC oscillation, you performa Laplace development (harmonic development) of your equations and you define the frequency, or rather omega = 2*pi*freq.
This is what happens when you choose a harmonic solver case (check the equations that COMSOL proposes, you will se the "t" change to "omega" ) in this mode you define amplitudes for the BCs and you do not give any "sin(2*pi*f*t)" equation at all. COMSOL then requests you to define the frequency, or a sweep over a series of frequency, in the harmonic solver node under "frequency" entry (use 100[Hz] or range(10,10,100) or whatever you want
A harmonic solver case is easier to solve and reuires less RAM, but assumes a steady state AC oscillation apart some special new special solver cases as the harmonic - transient case, see the doc, its worth to read it carefully
--
Good luck
Ivar
Please login with a confirmed email address before reporting spam
Posted:
1 decade ago
hi ivar
thank you very much for this explanation . one more question sir how would i know or how could i calculate the accurate time step for proper result. i used 0 to 1e-9 with step 1e-10. it gave me result similar to DC.
And is there any example available in comsol model library using frequency domain.
thank you very much for this explanation . one more question sir how would i know or how could i calculate the accurate time step for proper result. i used 0 to 1e-9 with step 1e-10. it gave me result similar to DC.
And is there any example available in comsol model library using frequency domain.