Learning Forum

[NvN]

Hello!

I am running an FDTD simulation with a uniform meshing defined by the maximum mesh step (25nm) in all dimensions. I am setting a mesh override in one region at the center of the simulation with a maximum mesh step that is less coarse (ideally 1 nm), however this is not working! Not only does recalculating the simulation mesh not do anything but the simulation is clearly stuck on the coarser meshing at all points in the simulation. Changing to an auto non-uniform mesh globally doesn't change anything, neither does changing the mesh constraint. Any suggestions on how to continue?

Best,

Nika

[NvN]

Furthermore, the autoshutoff level looks curious to me as it switches so dramatically quickly to the threshold. I would expect it to gradually decay. What could this be a symptom of?

[Guilin Sun]

In the case of uniform mesh, override can create non-uniform mesh outside of its region (plus buffer), so it does not work.
However, in the auto non-uniform mesh, it should o=work. Somehow the software may memorize some settings for the uniform mesh, so please delete the FDTD region and add it again.
Please note that using uniform mesh is generally not recommended. You can use it only at some specific case and you know the consequence. The non-uniform mesh is optimal please use it.

The autoshutoff min may be related to this. After you readd the FDTD region and it still shows that same, please write a separate post so we can focus the issue one by one.

[NvN]

Thank you gsun for your response.
I have deleted the FDTD region and added it again with an auto-non uniform mesh with a minimum mesh step of 25nm. Unfortunately it is still clear from the plotting of Pabs that the mesh override around my structure does not work.
P.S. as the autoshutoff issue also persisted I will write a separate post for that.

[Guilin Sun]

This is a different issue. The unclear image is due to coarse mesh, and each pixel is the mesh size. YOu have two methods to make it smoother:
1: using interpolation interp - Script command please interpolate the data before using matlab to plot, if you use matlab
2: using finer mesh, such as 5nm or 1nm in this region only.
the 2nd method is the ultimate way with higher accuracy.

[NvN]

The region I am plotting in the figure is the region where I imposed the mesh constraint so that is where I am expecting the pixel size/mesh to be 1nm.

[NvN]

This is clearly untrue as the pixel size/mesh in the figure is 25nm. So the mesh constraint continues not to work.

[Guilin Sun]

please check what you said and measure the pixel: "with an auto-non uniform mesh with a minimum mesh step of 25nm."
and zoom in the image, measure it.

[NvN]

Yes, I used a auto non-uniform mesh with a min mesh step of 25nm. I ALSO used a mesh constraint around the structure with a maximum mesh step of 1nm. However in the image it is clear that the pixel size is 25nm. So the 1nm mesh constraint again didn't work, as was my original problem.

[Guilin Sun]

In this case, I suggest that you use one online example, do not use uniform mesh but only the default mesh, and with override mesh enabled/disabled and see what happens.
You may also need to uninstall and reinstall it again, or close all the applications, restart your computer. or do both.

[Guilin Sun]

Here is my test result: using uniform mesh of 0.1um:


Applying the override mesh:

then re-calculate the mesh:

It works as expected.