Learning Forum

[iqnlo]

Hi All

I just got an unbelievable result when I was doing the propagation sweep in EME solver. The S12^2 shows that the transmission is greater than 1 and I think this is totally wrong. Because the transmission cannot be greater than 1.

The S21^2 is 99% which looks fine. Can I send the project file to you somehow?

[iqnlo]

Sorry, maybe the previous post is not clear. I'm simulating the coupling between fibre and a taper. So the port 1 is the fundamental TE mode of the fibre. And the port 2 is the end of taper.

[Guilin Sun]

Since this happens at smaller gap, it might be related to mesh size. Please use mesh size that can well resolve the gap.
From the curves, they are not smooth. It may indicate that the simulation time is not long enough, autoshutoff min level is to high. Please increase simulation (setting), reduce the autoshut off level.
In addition, it might be also due to PML reflection. Make sure the PML has about half wavelength thick, usually by increasing the number of layers. If it has larger angle incidence of the em wave, you may use SteepAngle PML with enough thickness.
If the result does not improve, please post some screenshots directly so to help us diagnose.

[iqnlo]

Hi!
Thanks a lot for your reply!
May I ask how to do this setting? "Please increase simulation (setting), reduce the autoshut off level."
Also, what's the meaning of "PML has about half wavelength thick"? The thickness is from where to where?

[Guilin Sun]

simulation time setting:
Autoshutoff min
PML thickness is the orange colored enclosure of the simulation volume:

you van use the "Ruler" to roughly measure it, or calculate it if you know how to get the mesh size, as you know the number of layers:

It seems you are new to FDTD. Please watch some edu videos online:

https://support.lumerical.com/hc/en-us/sections/360007813194-FDTD100-List-of-videos

[sagar]

I am facing the same issue. I disabled the autoshutoff option and increased the simulation time 3000fs, but did not work. How should I know what simulation time and auto-shutoff min I need to put?

[sagar]

Even with 5000fs simulation time and auto shut off min of 1e^-15, and 0.8-micron of PML width for 600nm dipole, I am getting transmission more than 1.

[Guilin Sun]

Q:How should I know what simulation time and auto-shutoff min I need to put?
A: you can check the time monitor result.
Q:Even with 5000fs simulation time and auto shut off min of 1e^-15, and 0.8-micron of PML width for 600nm dipole, I am getting transmission more than 1.
A: if you are using dipole source, the transmission can be larger than one, since internally it is simply normalized to sourcepower, not the dipolepower. For dipole source, they are not the same:
sourcepower - Script command
dipolepower - Script command

[iqnlo]

Hi !
Thanks a lot for your reply. But the problem I encountered was in EME solver. Is there also this simulation time setting in EME solver? Thanks!

[Guilin Sun]

EME is a frequency-domain solver, which is steady-date algorithm. It needs proper number of cells, and modes, as well as others, Please refer those links
EME Convergence Testing - An Intuitive Approach
Convergence testing process for EME simulations

[sagar]

thanks for your reply. One doubt! how would we know what simulation time and auto-shut off value we need to put to get the best results?

[Guilin Sun]

It is hard to say in a few words. Simple method is, set sufficient long simulation time, then use the "check point", you can change the autoshutoff min from 1e-5,1e-6,1e-7 etc to resume the simulation and compare the results. If there is no change, the previous autoshutoff min is enough.

[sagar]

thanks for your reply. I think the transmission is going more than one because of Purcell factor enhancement, as you mentioned earlier also. To get renormalized transmission we need to divide transmission with the Purcell factor. How much significant it would be?

[Guilin Sun]

You are right. The Purcell factor is the ratio of dipolepower/sourcepower. So if the dipolepower is high your result will be much larger than 1.