On Wed, 11 Apr 2007, matt wrote:
What I tried checking was if any dimensions were too small; PML thickness, cell width, and distances from flux planes to sources and scatterer. I also played with df, trying to make the source have a wider df than the flux regions. Finally, I increased the grid resolution, which also had no effect. I think gridding problems would manifest themselves at shorter wavelengths, so that makes sense.
Try putting a *narrower* bandwidth source around the frequencies where you have having problems. Sometimes a very wide-bandwidth source (short pulse) will lead to accuracy problems, in part because the stop-when-decayed is no longer accurate --- almost all of the energy of the short pulse will propagate through in a small time, and so the simulation stops, but there is still energy from low frequencies that hasn't gotten through yet because its timescale is longer.
Steven _______________________________________________ meep-discuss mailing list [email protected] http://ab-initio.mit.edu/cgi-bin/mailman/listinfo/meep-discuss
