Error message : "RuntimeError: meep: Can only phase in similar sets of chunks: 1 vs 2"

[andrew-ueltd]

Am getting the error message "RuntimeError: meep: Can only phase in similar sets of chunks: 1 vs 2" when trying to use "sim.fields.phase_in_material". The structure comprises two layers whose conductivity is changing as a function of time. Is there a list of error messages anywhere? This particular message incurs no hits on a google search. What does this one mean? Any help appreciated.

[stevengj]

Basically you need to make sure that Meep discretizes the two structures in the same way. They should have the same materials, just with different values. If one of the conductivity values is 0.0, I would try changing it to something tiny like 1e-15 to make sure that Meep doesn't delete it.

The "phase-in" feature hasn't been worked on for a long time and doesn't have much in the way of tests of documentation; I'd really like to replace it with something better: #207

[andrew-ueltd]

There looks to be other problems with the phase-in option. I've tried modifying the "phase-in-mat3.py" example in #207, setting up for a normal-incident Gaussian pulse plane wave with PMLs with a fixed-position
dielectric cylinder. I've used the phase-in option between two identical cylinders with the same dielectric constant so the action of the phase-in should make no difference. However, when I use this the fields diverge. Two files (one with and one without phase-in) and their log files enclosed. It looks like the phase-in is doing something wrong, but I can't tell what ...
phase_in_problem.zip

[andrew-ueltd]

A correction to above remarks - it's probably not field divergence that I'm seeing, but a normalization problem with the field measures used as markers for divergence being extremely large - I haven't examined these numbers at early time steps to be certain.