Learning Forum

[Ray Chen]

Hello,

I would like to inquire about the setup of the  source. I am currently optimizing the extinction ratio of a polarization beam splitter (PBS) at a wavelength of 1550 nm using PSO. The optimized Figure of Merit (FOM) is 35dB for TM mode. However, I have some problem when trying to simulate the bandwidth. I obtained the bandwidth through a center wavelength sweep, but the results showed two peaks in the bandwidth, which seems somewhat unreasonable.

 

Subsequently, I directly extended the wavelength span (1450nm~1650nm) in the  source and obtained a result from the model, which showed a decrease in the extinction ratio to 22 dB. This is completely different from the result obtained through optimization.

I would like to inquire about any mistake in my setup and how to rectify them.

Thank you.

 

 

 

 

[Guilin Sun]

Usually FDTD is better used for broadband, even though you only need results at a single wavelegth. If you only set one single frequency you may not know if it has resonance and since the actual injected signal is broadband, meeting the autoshutoff min does not mean this frequency reaches the same level.

To allievate the issue you faced, please tick this option:

and specify the min and max wavelengths for simulation. In this way, the simulation settings make sure the material fitting model is the same whether it is at the single frequency or broadband, and the mesh will keep the same at the same time. You can also separately set them: the material fitting range (advanced in material explorer) and the mesh  generation "override simulation bandwifth for mesh generation" shown above.

[Ray Chen]

Thank you for your reply, It's works.

Thanks !