New feature 'FDTD-Ports' - Frequency analysis tab

[evachat]

Hello,

i would like to clarify some points of the new feature ‘Ports’ of FDTD. I want to calculate S-parameters in a broad frequency range. In the ‘Modal Properties’ tab i select 10 frequency points an in the frequency analysis tab i run a Frequency sweep (is it necessary?) which results in artificial peaks like here https://kb.lumerical.com/en/pic_passive_dispersion_analysis.html. However, in Material Explorer there isn’t the ‘fit with multi-coefficient model’ button to solve the above problem. Is there another way to do this?
The peaks are also observed in S-parameters (abs^2) when i select 10 frequency points whereas the s-parameter curve is smooth when i choose 1 frequency point.

Thanks in advance

[bkhanaliloo]

Hi @evachat

It took me a while to replicate the problem but I guess I understood the problem. The Warning box that appears when you click on Frequency analysis causes confusion and I requested to modify it.

Since FDTD is time domain simulations, it automatically use a fit to sampled data. The discontinuity problem that you are facing in frequency sweep, I think, is because the mode of interest is not tracked properly. As a result the effective index experiences sudden and significant changes. Increasing the mesh size, number of test mode in the Frequency analysis, and lowering the bandwidth might be some tricks that you can use to avoid the problem.

i run a Frequency sweep (is it necessary?)

No, this is not necessary but it is a good way to visualize what modes are injected (although software uses Chebyshev grids to inject the light).

I hope this solves the confusion as well as the second part of the problem.

Please let me know if you have further questions.

[evachat]

Thanks, for the answer. I will check your reccomendations.

[bkhanaliloo]

Hi @evachat

This problem is solved with 2017a R4 (v8.17.1102) version, which is available for download from software download center.

Thanks

[lih076]

Hi, I’m wondering how to check the injected mode field by using multifrequency mode calculation in 3D FDTD. The link(https://kb.lumerical.com/en/ref_sim_obj_mode_source_-_broadband.html) tells me that the frequency points here do not represent the fields that are actually injected into the simulation… Thank you.

[bkhanaliloo]

Dear @lih076

You can use a frequency domain field and power monitor in front of the source to visualize the injected mode. In the 2017a-R1 version (under result view window for the source), you could see the actual injected mode profile (the interpolated and actually injected mode), but this option is not available in the recent versions.

If you want to learn more about Chebyshev coefficients, you can take a look at this example:

https://kb.lumerical.com/en/index.html?ref_scripts_chebin.html

Thanks