On Sep 28, 2010, at 2:49 AM, Geethaka Devendra wrote:
I am totally confused about the harminv function in Meep. I am trying to figure out the modes supported in a normal waveguide by using harminv.
(Note that MPB is probably easier for this sort of thing.)
When I change the location of the point where the harminv is run, output of harminv changes, which I can not understand. why would this be different for different locations in the waveguide?
In general, realize that harminv is solving a fairly tricky signal- processing problem. It is not bulletproof. Sometimes it will miss modes, for example. This is especially true if you give it a broad bandwidth containing lots of modes. Even when it finds all of the modes, there is some error bar in the values that it computes, e.g. the frequency and Q.
At different points in the waveguide, different modes will have different relative amplitudes, and this will change the harminv results at least slightly. If one of the mode amplitudes changes drastially (e.g. if you pick a point near where that mode has a node, e.g. in a symmetry plane) then harminv may miss it entirely.
Also the other problem is, say if harminv outputs 3 frequencies, and I select the first for fcen of the gaussian, it again outputs more than one, shouldn't this be then one, corresponding to the fundamental mode then?
If you want to find just one mode, you have to change more than just the fcen, you also have to change the bandwidth to excite just that one mode.
Steven _______________________________________________ meep-discuss mailing list [email protected] http://ab-initio.mit.edu/cgi-bin/mailman/listinfo/meep-discuss
