On Wed, 19 Apr 2006, Andreas von Rhein wrote:
Please have a look at the attached files.
square.dat and square2.dat are calculated with eigen.ctl as given.
Only difference is the M-Point. For square.dat I've chosen (vector3 0.5 0.5 0)
and for the second (vector3 0.5 -0.5 0)
So meep does calculate modes with k=0, in contrast to your explanaition. Sure
f=0 couldn't be found, therefor the bands jump.
I didn't say it doesn't calculate k=0 modes, I said it doesn't compute
f=0 modes. In your original simulation you had an empty cell and you were
exciting with a planewave source, and therefore you *in that case* should
only excite the f=0 mode (modulo numerical noise).
In M-dir some bands couldn't be found by meep (as in my own code)
but with changing to (0.5 -0.5 0) harminv detects them, any idea?
Don't excite with a planewave source. Excite with a point source at a
random location. The basic problem is that if you excite with a source
having too much symmetry, you will only get modes with that symmetry and
you will not excite other modes. (Similarly, run-k-points uses harminv at
a point given by your source location, and this should not be a
high-symmetry point.)
Steven
_______________________________________________
meep-discuss mailing list
[email protected]
http://ab-initio.mit.edu/cgi-bin/mailman/listinfo/meep-discuss
