Hi I have a problem with meep. I am creating a pretty big simulation of photonic crystals. The size of the volume is about 200x200x200 and it is running on MPI. There is a photonic crystal structure inside and a nanocavity which is supposed to have a Q-factor of about 10^7 so to calculate it I need also pretty many steps. The problem is that meep's simulation diverges after about 15000 steps. What it means is that suddenly in the place where the lightsource was ( and it already stopped working ) there appears something like a cone of tiny waves that is growing rapidly from step to step. The energy inside the calculation volume ( which i measure with fields::energy_in_box ) also starts to grow extremly fast although there are no more working lightsources inside. It reaches levels like 10^80. This is why I am sure it is not a physical solution and the problem must come from some numerical divergences. The code for that has already few thousand lines in C++ so I can not show it here but the main characteristics are:
- size in meep 10x10x10 with resolution 20 - pml layer with thickness 0.2 - photonic crystal with a nanocavity inside - symmetries, no symmetries, same result - divergence - using fields::use_real_fields (otherwise it freezes in MPI) - exciting with one gaussian_src_time with freq 0.7, width 3.0 for the first 10000 steps Does anyone have any idea what it could be ? Thomas Jefferson _______________________________________________ meep-discuss mailing list [email protected] http://ab-initio.mit.edu/cgi-bin/mailman/listinfo/meep-discuss
