Correct usage of waveguide ports

[mugundhan1]

Hi,

I have been trying to learn openEMS, and potentially use it in my research work too. I was trying to simulate a waveguide with a WR42 structure, similar to the one in the example, but used FreeCAD as a front-end to draw the structure.

The flow is as follows: Draw structure in FreeCAD, export components as stl files, import them in OpenEMS, and simulate.

When I do this, the structure gets imported properly, but when I run the simulation, i get the following warning:

Create FDTD operator (compressed SSE + multi-threading)
FDTD simulation size: 62x62x62 --> 238328 FDTD cells
FDTD timestep is: 9.47131e-13 s; Nyquist rate: 20 timesteps @2.63955e+10 Hz
Excitation signal length is: 505 timesteps (4.78301e-10s)
Max. number of timesteps: 1000000000 ( --> 1.9802e+06 * Excitation signal length)
Create FDTD engine (compressed SSE + multi-threading)
Warning: Unused primitive (type: Box) detected in property: port_excite_1!
Running FDTD engine... this may take a while... grab a cup of coffee?!?

I have uploaded my files here. I would be glad if someone helps me identify where the error is.

Steps I've taken:

1. I found that the structure of the port was originally larger than the waveguide 'opening', so I made it to the extent of 'a' and 'b' dimensions of the waveguide.
2. I have ensured that between the exciter and the it/ut and the port excitation is atleast spanned by 3 mesh-cells atleast.
3. I have tried placing the rectangular waveguide inside and outside the waveguide structure.
4. Tried refining the definitions of the dump box too.

I have added my stl files from FreeCAD and the matlab code (most of which was lifted and modified from the rectangular waveguide example.
I would be glad if someone can guide me here! :)

Sincerely,

Mugundhan
WR42.zip

[VolkerMuehlhaus]

@mugundhan1 One general comment: you have uploaded the *.m file and the STL, but it takes a bit of work to get this running on other machines. Not everone is willing to spend that amount of time to see what you did.

If you upload the *.XML model also (the file that is displayed in AppCSXCAD), it is more convenient for other users to quickly check what you did.

[mugundhan1]

Hi,

I'm sorry I missed this. I understand. I uploaded all the files just so if someone needs they have the original design files too. I attach the .xml file here.
WR42.xml.txt

[VolkerMuehlhaus]

Please upload the all *.STL also, unfortunately this data is not in the XML file.

[mugundhan1]

Hi,

Please find attached the stl files as zip.
WR42.zip

Mugundhan

[mugundhan1]

Hi, seemed like I was not meshing the boundaries properly. It works now, and I see some output for the Simulation. Will check the accuracy of it now.

[mugundhan1]

Update: I was able to get rid of the ''primitive not found'' error by placing mesh lines on all the metal edges and smoothing mesh-lines around it. But now I get a S11 and S21 value that is not very realistic. I'll work on finding out why it is that bad!

Mugundhan

[VolkerMuehlhaus]

Yes, that is what I also saw: you mesh was not aligned with the metals, which is why I don't like the STL geometries for my work.
I think your ports are also not aligned with the mesh.
And from what I remember, your ports where located somewhere inside the waveguide, at quite a distance to the flanges and at quite a distance to the boundary conditions. I think you should check again how this is done in the waveguide example, so that you don't have a wave propagating away from the port in two directions.

[mugundhan1]

Hi, I have tried adjusting the port positions now. I placed the start position of the port at the entrance of the waveguide and have shortened the walls of the mesh to be only the waveguide's start position - 8mesh_res, and end as stop + 8mesh_res.

%%
cd /home/mugundhan/Desktop/personal/freecad/WR42_Wg/
clear;
clc;
close all;
%%
physical_constants;
unit = 1e-3;

%%
f_start = 10e9;
f_stop  = 40e9;
f_0     = 24e9;%(f_start + f_stop)/2;
lambda0 = c0/f_0/unit;
lambda_fstop = c0/f_stop/unit;
%WR2_Dims
a = 10.668; % mm along x axis
b = 4.318; % mm along y axis
c = 6.35;
d = 12.7;
e = 22.41;
f = 22.41;
flange_thickness = 4.60; % in the z-axis
waveguide_length = 20; %mm along z axis.

%%
TE_mode = 'TE10';
%Specify the mesh resolution as 30 cells per wavelength in all directions:
mesh_res = lambda0./[30 30 30];

%%
FDTD = InitFDTD('EndCriteria',1e-4);
FDTD = SetGaussExcite(FDTD,0.5*(f_start+f_stop),0.5*(f_stop-f_start));

%%
BC = {'PML_8','PML_8','PML_8','PML_8','PML_8','PML_8'}; %pml in pos. and neg. z-direction
FDTD = SetBoundaryCond(FDTD,BC);

%% Define prelim mesh
mesh.x = [-15,-e/2,-d/2,-a/2,0,a/2,d/2,e/2,15];
mesh.y = [-15,-f/2,-c/2,-b/2,0,b/2,c/2,f/2,15];
mesh.z = [0-8*mesh_res(3),0, 4.6, 15.4,20, 20+8*mesh_res];
mesh.x = SmoothMeshLines(mesh.x,mesh_res(1),1.3);
mesh.y = SmoothMeshLines(mesh.y,mesh_res(2),1.3);
mesh.z = SmoothMeshLines(mesh.z,mesh_res(3),1.3);
CSX = InitCSX();
CSX = DefineRectGrid(CSX, unit,mesh);

%%
CSX = AddMetal(CSX,'waveguide');
CSX = ImportSTL(CSX,'waveguide',10,'/home/mugundhan/Desktop/personal/freecad/WR42_Wg/WR42-Waveguide.stl','Transform',{'Scale',1});

CSX = AddMetal(CSX,'Flange1');
CSX = ImportSTL(CSX,'Flange1',11,'/home/mugundhan/Desktop/personal/freecad/WR42_Wg/WR42-Flange1.stl','Transform',{'Scale',1});

CSX = AddMetal(CSX,'Flange2');
CSX = ImportSTL(CSX,'Flange2',12,'/home/mugundhan/Desktop/personal/freecad/WR42_Wg/WR42-Flange2.stl','Transform',{'Scale',1});

%% Define ports
start=[-a/2 -b/2 0];
stop =[a/2 b/2 0+4*mesh_res(3)];
[CSX, port{1}] = AddRectWaveGuidePort( CSX, 1, 1, start, stop, 'z', a*unit, b*unit, TE_mode, 1);

start=[-a/2 -b/2 20-4*mesh_res(3)];
stop =[a/2 b/2 20];
[CSX, port{2}] = AddRectWaveGuidePort( CSX, 0, 2, start, stop, 'z', a*unit, b*unit, TE_mode);

CSX = AddDump(CSX,'Et','FileType',1,'SubSampling','2,2,2');
start = [mesh.x(1) mesh.y(1) mesh.z(1)];
stop  = [mesh.x(end) mesh.y(end) mesh.z(end)];
CSX = AddBox(CSX,'Et',0 , start,stop);

%%
Sim_Path = 'tmp';
Sim_CSX = 'WR42.xml';
[status, message, messageid] = rmdir(Sim_Path,'s');
[status,message, messageid] = mkdir(Sim_Path);
WriteOpenEMS([Sim_Path '/' Sim_CSX],FDTD,CSX);
CSXGeomPlot([Sim_Path '/' Sim_CSX]);

%%
RunOpenEMS(Sim_Path, Sim_CSX)
%%
freq = linspace(f_start,f_stop,4001);
port = calcPort(port, Sim_Path, freq);

s11 = port{1}.uf.ref./ port{1}.uf.inc;
s21 = port{2}.uf.ref./ port{1}.uf.inc;
ZL = port{1}.uf.tot./port{1}.if.tot;
ZL_a = port{1}.ZL; % analytic waveguide impedance

%% plot s-parameter %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
figure
plot(freq*1e-6,20*log10(abs(s11)),'k-','Linewidth',2);
xlim([freq(1) freq(end)]*1e-6);
grid on;
hold on;
plot(freq*1e-6,20*log10(abs(s21)),'r--','Linewidth',2);
l = legend('S_{11}','S_{21}','Location','Best');
set(l,'FontSize',12);
ylabel('S-Parameter (dB)','FontSize',12);
xlabel('frequency (MHz) \rightarrow','FontSize',12);
ylim([-100,0])

%% compare analytic and numerical wave-impedance %%%%%%%%%%%%%%%%%%%%%%%%%%
figure
plot(freq*1e-6,real(ZL),'Linewidth',2);
hold on;
grid on;
plot(freq*1e-6,imag(ZL),'r--','Linewidth',2);
plot(freq*1e-6,ZL_a,'g-.','Linewidth',2);
ylabel('ZL (\Omega)','FontSize',12);
xlabel('frequency (MHz) \rightarrow','FontSize',12);
xlim([freq(1) freq(end)]*1e-6);
l = legend('\Re(Z_L)','\Im(Z_L)','Z_L analytic','Location','Best');
set(l,'FontSize',12);

%% Plot the field dumps %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
figure
dump_file = [Sim_Path '/Et.h5'];
PlotArgs.slice = {0 0 0};
PlotArgs.pauseTime=0.01;
PlotArgs.component=0;
PlotArgs.Limit = 'auto';
PlotHDF5FieldData(dump_file, PlotArgs)

The S21 I get doesn't have any reflections, but seems pretty attenuated.

[mugundhan1]

Hi @VolkerMuehlhaus

Sorry, I wasn't able to fix the error yet. I came back to read your suggestions:

"And from what I remember, your ports where located somewhere inside the waveguide, at quite a distance to the flanges and at quite a distance to the boundary conditions. I think you should check again how this is done in the waveguide example, so that you don't have a wave propagating away from the port in two directions."

I have now made the ports to start and end in the flanges. The structure I simulate is the same as in the waveguide example, but the lengths are different (20 mm) instead of 50 considered there. I use PML8 in all directions because I want to see the waveguide walls to be like the PEC they would be. Does this mean I should have a perfectly absorbing layer?

I'm sorry, I'm new to FDTD, so are there any rules of tumb to follow on boundary conditions, or port placement?