Hi, Devin, as far as I understand the simulation, with the gaussian source you may excite plenty of cavity modes with different longitudinal and transversal (even) numbers. Your source is narrowband, selecting (almost) exclusively one mode which I would denote as "14".
I have ran your simulation at 99% the frequency and at 90% the frequency, which shows that with such a narrowband source you may excite any mode you wish. (See attachment.) The "10" mode you may be looking for seems to have much lower frequency than you are trying now (I guess around 80% thereof). I expect MEEP to be precise, may there be some mistake in the setup? >From my experience, I would recommend to use a very short pulse (df > fcen) and let harminv separate all the modes. At least it saves time. Additionally, I would reduce the resolution at least twice (so that it finishes 16 times faster), it should not influence the resonance frequency. You may also record the field and look at the peaks in transmission. They might look like this: http://www.optique-ingenieur.org/en/courses/OPI_ang_M01_C03/co/Contenu_02.html I would try it, but as a python-meep user I have no clue how to compute and plot Fourier transform in scheme. Filip On Wed, 3 Oct 2012 12:48:53 +0100 Devin Dunseith <[email protected]> wrote: > Hi Filip, > > See attached, > > I also show the electric field pattern of the dominant mode (the > e-field h5 file is too big to attach, so i just attached a few image > files). If I could find the resonant mode I wanted, I would excite > with a narrower band source so we wouldn't see this (I think 2,3 > Laguerre-Gaussian) mode. > > Cheers, > > Devin > > On 3 October 2012 12:11, Filip Dominec <[email protected]> wrote: > > Hi, could you send me the outputs, please? > > Maybe I will not tell you anything you did not know, but I am interested in > > it. > > F. > > > > 2012/10/3, Devin Dunseith <[email protected]>: > >> Hi Filip, > >> > >> Thanks! A quick update though...I've been working on it a lot, and I > >> gather that it might not be a matter of underestimating the > >> frequencies - when I look at the frequencies near where my resonance > >> should be, the field pattern is completely wrong (it should be a > >> fundamental longitudinal mode, with three half wavelengths in the > >> cavity). However, I still can't understand why it's not finding my > >> resonance though. > >> > >> Thanks again, > >> > >> Devin > >> > >> On 3 October 2012 11:28, Filip Dominec <[email protected]> wrote: > >>> Hi, Devin, > >>> I ran your simulation several times to experiment a bit. I will take a > >>> look at it again and write my suggestions later. > >>> F. > >>> > >>> 2012/10/2, Devin Dunseith <[email protected]>: > >>>> Hi, > >>>> > >>>> I am trying to simulate a microwave near-confocal Fabry-Perot cavity > >>>> with MEEP. I made the spherical cavity mirrors by first making > >>>> rectangular blocks of metal, and then iteratively replacing the metal > >>>> with cylinders of air, varying the size of each cylinder so as to > >>>> ultimately make a spherical concave mirror. I bring the fields into the > >>>> cavity via a rectangular waveguide and sub-wavelength aperture. > >>>> > >>>> After the sources (and some extra time), I run harminv, and have tried > >>>> running harminv both in the centre of the cavity and at then end of an > >>>> output waveguide. I have everything set up so that a=1cm, and Gaussian > >>>> optics theory and experiments tell me that the resonance i'm looking for > >>>> (the TEM 3,0 mode) should be at 14.22 GHz, which is 0.474 in my MEEP > >>>> units. > >>>> > >>>> However, no matter what resolution I use, I have not managed to get the > >>>> resonance in Harminv above 0.46 (about 13.8 GHz), which is too far from > >>>> what it should be to ignore. > >>>> > >>>> I have set up my geometry in 3d, but usually do calculations in 2d to > >>>> save time. I'm trying a simulation in 3d with low resolution to see if > >>>> that helps, but am otherwise at my wits' end. Has anyone experienced > >>>> anything similar? > >>>> > >>>> My code is below, and png from (output-epsilon) attached. > >>>> > >>>> Thanks, > >>>> > >>>> Devin > >>>> > >>>> (define-param D 9) ;mirror diameter > >>>> (define-param hole-size 2) ;input coupler hole size > >>>> (define R 7.5) ;mirror radius of curvature > >>>> (define-param fcen 0.475) ;centre frequency > >>>> (define-param df 0.05) ;frequency width > >>>> (define-param res 30) ;resolution > >>>> (define cav-length 3.5) ;cavity length > >>>> > >>>> (define (H x) (* (/ 1 2) (+ (* -1 (sqrt (+ (* -1 (expt D 2)) (* 4 (expt > >>>> R 2))) )) (sqrt (* -1 (+ D (* 2 R) (* -2 x)) (+ D (* -2 (+ R x)))) > >>>> )))) ; a function for use in my iterative creation of spherical mirrors > >>>> > >>>> (define topbot -1) ;parameter for controlling if I am making the top or > >>>> bottom mirror > >>>> (define (sphermirror top x x-max dx) ;a function to create mirror > >>>> concavity iteratively > >>>> (if (<= x x-max) > >>>> (begin > >>>> (if (= top 1) (set! topbot 1) (set! topbot -1) ) > >>>> (append > >>>> (list > >>>> (make cylinder (center 0 (* topbot (- (H x-max) (/ (H x) 2) (/ > >>>> cav-length 2))) 0) (material vacuum) > >>>> (radius (- (/ D 2) x) ) (height (H x)) (axis 0 1 0) > >>>> )) > >>>> (sphermirror top (+ x dx) x-max dx))) > >>>> '() > >>>> ) > >>>> ) > >>>> (set! geometry-lattice (make lattice (size 16 26 no-size))) > >>>> (set! geometry > >>>> (append > >>>> (list ;make the mirror foundation blocks > >>>> (make block (center 0 (- (/ (H (/ D 2)) 2) (+ (/ cav-length 2) > >>>> 0.1)) > >>>> 0) (size D (H (/ D 2)) D) > >>>> (material metal)) > >>>> (make block (center 0 (- (+ (/ cav-length 2) 0.1) (/ (H (/ D 2)) > >>>> 2)) > >>>> 0) (size D (H (/ D 2)) D ) > >>>> (material metal)) > >>>> ) > >>>> (sphermirror 0 0.005 (/ D 2) 0.005) ;make the mirrors concave > >>>> (sphermirror 1 0.005 (/ D 2) 0.005) > >>>> (list ;make the waveguides and input/output coupling holes > >>>> (make block (center 0 (- (/ (- 13 (/ cav-length 2) 0.1) 2) 13) 0) > >>>> (size > >>>> 2 (- 13 (/ cav-length 2) 0.1) 1.2) > >>>> (material metal)) > >>>> (make block (center 0 (- (/ (- 13 (/ cav-length 2) 0.1) 2) 13) 0) > >>>> (size > >>>> 1.6 (- 13 (/ cav-length 2) 0.1) 0.8) > >>>> (material vacuum)) > >>>> (make cylinder (center 0 (* -1 (+ (/ cav-length 2) 0.05)) 0) > >>>> (material > >>>> vacuum) (radius (* hole-size 0.05)) (height 1) > >>>> (axis 0 1 0)) > >>>> (make block (center 0 (- 13 (/ (- 13 (/ cav-length 2) 0.1) 2)) 0) > >>>> (size > >>>> 2 (- 13 (/ cav-length 2) 0.1) 1.2) > >>>> (material metal)) > >>>> (make block (center 0 (- 13 (/ (- 13 (/ cav-length 2) 0.1) 2)) 0) > >>>> (size > >>>> 1.6 (- 13 (/ cav-length 2) 0.1) 0.8) > >>>> (material vacuum)) > >>>> (make cylinder (center 0 (+ (/ cav-length 2) 0.05) 0) (material > >>>> vacuum) > >>>> (radius 0.025 ) (height 1) > >>>> (axis 0 1 0)) > >>>> ) > >>>> ) > >>>> ) > >>>> > >>>> (set! pml-layers (list (make pml (thickness 1.0)))) > >>>> (set! sources > >>>> (list > >>>> (make source ;excite the cavity with a TE10 mode gaussian source > >>>> in the input waveguide > >>>> (src (make gaussian-src > >>>> (frequency fcen ) (fwidth df) )) > >>>> (component Hx) > >>>> (center 0 -6.5 0) > >>>> )))) > >>>> (set! resolution res) > >>>> > >>>> (run-sources+ 10000 ;run the sources before performing haminv of the > >>>> correct e-field component at cavity centre and end of output waveguide > >>>> (at-beginning output-epsilon) > >>>> (after-sources (harminv Ez (vector3 0 0 0) fcen df)) > >>>> (after-sources (harminv Ez (vector3 0 6.5 0) fcen df)) > >>>> ) > >>>> > >>>> (run-until (/ 1 fcen) ;visualize the field over one period > >>>> (at-beginning output-epsilon) > >>>> (to-appended "ez" (at-every (/ 1 fcen 20) output-efield-z)) > >>>> ) > >>>> > >>
<<attachment: 90freq_Dunseith-ez.png>>
<<attachment: 99freq_Dunseith-ez.png>>
_______________________________________________ meep-discuss mailing list [email protected] http://ab-initio.mit.edu/cgi-bin/mailman/listinfo/meep-discuss
