TAGGED: fdtd, photonics, ring-resonators
-
-
December 29, 2022 at 2:53 pm
jasper
SubscriberHi, in a FDTD simulation of a Ring Resonator Cavity I get the following result, which has the expected shape except for a weird power build up at the end of the simulation time:
What is going wrong here? Could this be caused by the PML or meshing? This is the simulation setup:
Where I use PML on the x axis boundaries and the source is on the left side.
-
December 29, 2022 at 8:06 pm
Guilin Sun
Ansys EmployeeThis indicates that the simulation suffers the so called "late time divergence". For single bus ring resonators, since only one bus for in-out, less light can be coupled out in early simulation, it needs longger simulation time, and thicker PML. Please increase the PML number and make sure the total thickness is roughly half-wavelength.
Â
In addition, make the bus waveguide long enough to extend outside of PML.
-
January 6, 2023 at 9:54 am
jasper
SubscriberThank you for the recommendations. I have increased the simulation time from 5*10^5fs to 10^6fs and changed the PML from the "standard" profile with 64 layers to the "stabilized" profile with 256 layers, which has removed the late time divergence. However, this has introduced new absorption lines in the T spectrum which were not present before, such as these:
I assume this is due to the PML change. Can you recommend other PML settings?
Â
-
January 6, 2023 at 4:55 pm
Guilin Sun
Ansys EmployeeDoes the simulation end at 100%? I guess so. please use smaller autoshutoff min such as 1e-6 or smaller, and use much longer simulation time, say 1e8, and make sure the simulation terminates at less than 95% (you can check it in the log file). If it still has data larger than 1 when the simulation terminates at less than 90%, then reduce the autoshutoff min further.
Do not use higher mesh accuracy for initial test, as it needs more number of PML. In your case, I think you can simply use the Standard PML, with the proper number of layers to make sure the PML thickness is about half wavelength. Please use mesh accuracy 2, and use an override region in the coupling region.
Make the port size larger, as long as the mode is correct.
You can search this forum for more suggestions from other posts.
-
- The topic ‘Power Buildup at End of Simulation Time’ is closed to new replies.
-
3977
-
1461
-
1272
-
1124
-
1021
© 2025 Copyright ANSYS, Inc. All rights reserved.