Tutorial examples/absorbed_power_density.py problems

[slagwad]

Hello, when I attempt examples/absorbed_power_density.py from https://github.com/NanoComp/meep/blob/master/python/examples/absorbed_power_density.py it gives me an error stating the yee_grid is not understood in dft_fields. When I comment out the yes_grid command it works and gives roughly equivalent results to those posted at https://meep.readthedocs.io/en/latest/Python_Tutorials/Basics/#absorbed-power-density-map-of-a-lossy-cylinder.

My second issue is that if I replace the material from SiO2 to cSi at 0.8um wavelength the results show:
flux:, -1.1632079183598192e-05 (dft_fields), 1.570299363693065 (dft_flux), 1.0000074075551788 (error)

where the dft_flux seems realistic but the dft_fields is clearly not right. I would appreciate your help resolving these two inconsistencies. Thank you

[oskooi]

The yee_grid parameter of add_dft_fields was added in #1095 and released with Meep v1.13. You need to upgrade to the latest version which produces the correct result for cSi at 0.8 μm:

flux:, 1.5648277747465522 (dft_fields), 1.5702993636930598 (dft_flux), 0.003484424099643892 (error)

[slagwad]

Thank you. I tried to update meep inside my conda wrapper by performing a "conda activate mp" followed by "conda update pymeep" and agreed to the updates but even after starting a new shell window and activating mp, it still gives me the yee_grid error. What is the right way to update my pymeep inside this conda container?

Also, when I run the above but with a continuous source "mp.ContinuousSource" instead of the Gaussian, it ends up with an effectively infinite run time. I tried shortening the source width by 2*pml assuming that it was an interaction between the source and the assumed metallic side walls but that did not help. Thanks again!

[slagwad]

I figured out the update problem. I used "conda update -c conda-forge pymeep" and it worked.
I would still like help with the "mp.ContinuousSource" question. Much appreciated

[oskooi]

That's because the run condition in that tutorial is sim.run(until_after_sources=100) which for a ContinuousSource is infinite and thus the simulation will never terminate. In general, you need to use a GaussianSource whenever you are computing DFT fields as explained in this FAQ.

(Note: GitHub issues are not a general help desk. They are for reporting bugs and feature requests. For questions related to usage, you should use the mailing list.)

[slagwad]

Hello Ardavan, I have searched for the mailing list specific to Meep but could not find it. Would you kindly provide the link? Is it the one for www.simpetus.com? Thx again