Discussion Forum

Hi

the nanometer nm 1e-12[m] is rather small, especialy if you have other parts in the few hundred um = 1e-6 range, or more, side by side. Internally, in the COMSOL geometry handler there is a variable for absolute and relative resolution, by default set to 1e-6 for the relative scale, probably not enough for your case, try to push it higher, or try to avoid scale differences above 1:1e-6 (1:1e-4 would be better).

There are in fact two variables, one in the "File - import - Cad data files - Options tab" the "absolute import tolerance" (at default 1e-5) which means that objects smaller than 10um will be suppressed !
not aways of interest for mems or nano-optics, so do not forget to correct this Absolute scale when importing CAD data (it is reset to its default next time you open Comsol fresh, so ALWAYS check when importing small geometries). Note that there are other variables for STL/VMRL files that are important too for small features.

The second is the "repair tolerance" you find in "Draw - Create Composite Object" (if I remember well it is also in the repair/defeaturing window for imported CAD geometry in 3.5a) try to push this one to 1E-9 or smaller in your case.

Best rule, keep your geoemtry in a 1:1E-4 scale ratio, or divide it up into 2 distinct "Geom" at different scales and link the physics (by hand). Comsol applies internal scaling factors to compensate such that the matrices are still resolvable without to big numerical errors, but there are limits to what is possible in the binary representation of numbers on PC's (in fact we are rather LIMITED !)

Hope this helps
Ivar

Dear Ivar,

Thank you very much for your information. I am trying your suggestions. Thanks Again.

Manjula

Dear Ivar,

As you told once I decreased the size of large parts, the model was working ok. But I do need a large boundary which I couldnot use because of this limitation. I thought of use a couple geometry as you instructed. But my question is If I use a small boundary in one gemetry and If I connected ( through Identity conditions) it to a larger geometry would I get my solution correct. If there is something I might read Could you please let me know that. I am greatly appreciating your help.


Thanks

manjula

Hi

if you have such small items, I assume you are modelling with the RF modules, then most of the time you can replace the "large" boundaries by the perfectly matched layers PML's. this reduces drastically the required sub-domain sizes. Read carefully the doc on PML (there are also a few discussions on the forum, try a search (above to the left for the [Go]

Have fun Comsoling
Ivar

Hi,

I am currently trying to create a model of a photonoic crystal fiber for my work at SLAC and I have several questions about the COMSOL.


BOUNDRY CONDITIONS:

In order to get used to the features of Comsol, I started with a simple case of a regular fiber (Silica Core 10µm/ Cladding of doped silica glass 120µm and acrylate coating 230µm).
Everything went fine during the geometry design but when I got to the physics I couldn't define the boundary conditions. The software won't let me choose the surfaces for the boundary conditions like in the tutorial model provided with the software.

GEOMETRY DESIGN:
My second problem concerns the revolving option. Indeed, for the photonic crystal fiber, I drew a circles in order to make a periodic structure only by rotations.

THERMAL EFFECTS:

My ultimate goal is to evaluate the thermal consequences of the electric field inside the PCF. I was wondering if it's possible to consider the electric field as an thermal internal source in the heat equation. Once I know the temperature distribution, I want to see the changes on the index and then simulate again the light propagation.


Thank you

Rachik