Discussion Forum

Hi
I believe you should remake it fully fresh in latest V4 version, that would help.

Then note that Ex is -Vx=-dV/dx so your derivative d(es.Ex,y) = -Vxy.
This is already a second derivative of the dependent variable, you should increase the discretization to at least third order to get smoother derivatives. (and probably round off your sharp edges to minimise the edge singularities on the V gradients)

check the KB (knowledge base) about flux calculations

--
Good luck
Ivar

Thanks for your help.

I tried restating in 4.2 and rounded the edges. Unfortunately now I get an error:
Failed to evaluate variable.
- Variable: material.parameter.epsilonr22
- Geometry: 1
- Domain: 1

I can not find the cause. I attached the new file.

Hi,

Here is an updated model using 4.2 and rounded edges. I still dont know how to calculate the forces acting on each line. I think it should be an integral if dE/dy (for Fy) but I always get zero. am I using the correct module?

Thanks,
Matthias

Hi

in my 4.3.a I get non zero values (where you seem to get "0", strange), now you need to take the side parts too, ideally the full finger turn

But on the other side if you plot Vxx and Vyy along the entire loop around your "finger section" you get quite some spikes that will give you rather large errors see below.

I would propose to increase the discretization (turn on in the view mode) for V

One advise: use rather GND for the "0V" and a "Terminal" at 1V for the others BC's (instead of the Electric potential, as you have a scalar and not a field), then you get lumped parameters to compare already made up for you. But it should not change anything for your force calculation. And as you have only 1 domain, you cannot get the "ES Domain Force Calculation" defined

Do not forget these values are for a 1 meter loooong device (along Z)

--
Good luck
Ivar

Thanks.

I changed the potentials to Ground and Terminal. I am also installing Comsol 4.3. I dont know how to change the "discretization (turn on in the view mode) for V"?

Matthias

it all works out well in comsol 4.3.

Thanks for your help.

Matthias

Hi

that's good, by the way you were also missing a Force calculation on the domain, then you need to select the correct up/down (up for your case) tensor for the forces

--
Good luck
Ivar

thanks, were do I find the domain calculation for the tensor , I cant seem to find it. (I was determining the force by taking the derivative of the energy density ~es.E^2)

Hi

turn on the equation view (preferences) and dig into the equations, its somewhere there-below ;)

--
Good luck
Ivar

Hi ,

i would calculate the electrostatic force for 3D comb moving....i found the energy but i dont now how i can calculate the force..... Here I attached the model. Could anyone help me with it? Thanks

Hi

in v4.3a you have a Force domain node, you can add and COMSOL's calculates the force for you (with some restriction on BC set-up see the doc), but I cannot remember if that existed in v4.1
If not you need to try an energy derivation.

one thin about your model, you have a very long and thin air field, if you apply correct symmetry the thin layer approach can be used, indeed, but locally you have a very coarse mesh for such fine features. I would propose that you "box" locally your structure and use a far finer mesh in that region at least 2-3 mesh elements across the gap to better resolve the fields

--
Good luck
Ivar

thank you ivar...i found the force calculation node in version 4.1 just i need to define the domain
can you tell me please how i can do for the energy derivation method please ??? i want to compare the result

thanks