top of page

exercise 9.10

Question

ZIP with MATLAB scripts and note:

Small tag OK.jpg

exercise 9.10 notes:

Small tag OK.jpg
pozar_09_exercise_10_question.jpg
Background

Rectangular waveguide filled with gyrotropic material: the material spins polarization on the direction of the travelling wave.

 

Ferrite permeability mu is a tensor.

A gyrotropic material has natural electromagnetic resonance.

The delta H also called 'linewidth' is the Q factor or 3dB bandwidth Linewidth of  xe

  • mu=mu0*mur

  • mur=1+xe

  • xe : magnetic susceptibility

  • xe=xe1+1j*xe11

001.jpg

The following graph is the imaginary part of the magnetic susceptibility of an unspecified ferrite material.
 

Here gamma to calculate w0 is not the lossy material propagation constant alpha+1j*beta

 

but

gamma_gr=abs(qe)/me 

 

with 

qe : electron charge

me : electron mass

002.jpg
Circuit
003.jpg
006.jpg
Constants
c0 = 299792458                           % [m/s] light velocity
eps0  = 8.854187817e-12               % [F/m]=[C^2/N*m^2] permittivity, empty space  
mu0  = 4*pi*1e-7                              % [H/m]=[T*m/A] permeability, empty space  
m_electron=9.109e-31                    % [kg] electron mass
q_electron=-1.602176634e-19        % [C] electron charge
k1 factor [Hz/Oe]: needed to calculate precession frequencies:

From Quantum mechanics, the magnetic properties of materials like ferrites
arise from electron spin.

So electrons have angular momentum caused by orbital motion around nuclei
and magnetic moment caused by spin.

Lange factor g:
     g=1 electron moment is only caused by orbital motion
     g=2 electron moment is only caused by its spin, this is the case 
gyromagnetic ratio gamma_gr : ratio between angular and spin moment

gamma_gr=abs(q_electron/m_electron)    % [C/kg] gyromagnetic ratio

the gyromagnetic ratio is at the same time :
* electron mass to charge ratio
* and electron angular momentum to magnetic moment ratio

Magnetic field : H
magnetic field strength [A/m] in MKS units system
                                        [Oe] in CGS

Oe2Am=4*pi*1e-3           % conversion factor [Oe] to [A/m]

magnetic flux : B
[Gauss] : H through a given surface
[Wb]
1 [Mx] = 1e-8 [Wb]
1 [Tesla] = 1 [Wb/m^2]

G2Wb_ov_m2=1e-4

B in [G] = mur * H in [Oe]
conversion factor [Gauss] to [Weber/m^2]


A1=4*pi  % area of 1m radius sphere

k1= Oe2Am * G2Wb_ov_m2 * A1 * gamma_gr   % [Oe/Hz]



[POZAR] approximates k1 with 2.8e6
Question Inputs

a=.01                  % [m] rectangular waveguide horizontal size

f=1e10                 % [Hz]
lambda0=c0/f;
k0=2*pi/lambda0;

h0max=1500        
% [Oe]
delta_h=1700       % [G] same as 4*pi*Ms , Ms magnetization

er=13                  % relative permittivity
mur=1                 % no specific mur  defined so assume 1

N1=1e3                % amount steps in H0 range
dh0=h0max/N1     % H0 range step

H0=[dh0:dh0:h0max];


waveguide geometry fill-up parameters
because of surface and vpasolve attempts didn't work
I decided to fill the waveguide progressively.

N3=[1.01 2 10 1e2 1e3];

t=a/2*(1-1./N3)  
% ferrite(s) thickness
c=.25*(a-2*t)      % c: ferrite left hand side air gap span 
d=c                     % d: ferrite right hand side gap span 

Range for sought propagation constant

N5=1e3    % beta resolution, when not symbolic
Beta_range=k0*linspace(.5,4,N5);


f0 : precession (a.k.a. Larmor) frequency, 1st limit
material and H0 dependent

F0=k1*H0;  

fm : precession higher frequency, 2nd limit 
material dependent

fm=k1*delta_h  

k2 : mu permeability deviation
frequency, material and H0 dependent

K2=mu0*mur*fm*f./(F0.^2-f^2);


 
005.jpg
this exercise is about solving the following equation:
004.jpg
I tried with a surface but with tangent and cotangents turning the equation into transcendental, not even MATLAB standard functions can get this in a single shot. [POZAR] suggests Newton-Raphson.

I found a good approximation matching expected results with Douglas Schwarz support function intersections.m

%%  mue : Effective Permeability

MUe = mu0*mur*(1+fm*f./(F0.^2-f^2));    % (9.25a)

%% B contains all betas for each t/a

B=zeros(2,numel(H0)+1,numel(N3));


for s2=1:1:numel(N3)


% n4 low means thin ferrite, n4 large means ferrite completely
% filling up  waveguide cross-section.


   t1=t(s2);
   c1=c(s2);
   d1=d(s2);
  % c=d imposed, from question header but leave d in case future use

   beta=Beta_range;  % it was syms but didn't work  

   Beta=[t1;c1];

   for s1=1:1:numel(H0)

      Kfsq=4*pi*f^2*MUe(s1)*eps0*er-beta.^2;
      Kasq=4*pi^2/c0^2*f^2-beta.^2;


% building equation
      Eq1=Kfsq./(MUe(s1)).^2;                                                          % [1]

      Eq2=K2(s1).^2.*beta.^2./((mu0*mur)^2*MUe(s1).^2);             % [2]

      KEq3=abs(Kasq.^.5).*cot(abs(Kasq.^.5)*c1)./(mu0*MUe(s1));    % [3]
      Eq31=KEq3.*abs(Kfsq.^.5).*cot(abs(Kfsq.^.5)*t1);
      Eq32=KEq3.*K2(s1).*beta/mur;

      Eq4=Kasq/mu0^2;                                                                
   % [4]

      Eq5=cot(abs(Kasq.^.5)*c1).*tan(abs(Kasq.^.5)*d1);                  % [5]

      KEq6=abs(Kasq.^.5).*tan(abs(Kasq.^.5)*d1)./(mu0*MUe(s1));   % [6]
      Eq61=KEq6.*abs(Kfsq.^.5).*cot(abs(Kfsq.^.5)*t1);
      Eq62=-KEq6.*K2(s1).*beta./mur;


% assembling equation

      S1vor=Eq1+Eq2+Eq31+Eq32+Eq4+Eq5+Eq61+Eq62;

%% solving equation
% When plotting the following
% figure;plot(beta,S1vor);grid on;xlabel('\beta');title('S1vor')
% the asymptote is the sought point [beta(n) H0(n)].
% To capture this point support function intersections.m is particularly handy
% directly using the Y points generated by plot function 


      hf1=figure;
      hf1.Visible='off';
      hp1=plot(beta,S1vor);grid on;xlabel('\beta');title('S1vor');


% capturing asymptote
      [x0,y0]=intersections([beta(1) beta(end)],[0 0],beta,hp1.YData,'robust');

      if numel(x0)<2
          x0=[x0;0];
     end

      if numel(x0)>2
          x0([3:end])=[];
      end


%% logging single point result
      Beta=[Beta x0];

      close all;

% Not closing so many figures causes forced shut down that in turn
% shows up as the following low level error on launching MATLAB again:
%
% Warning: MATLAB previously crashed due to a low-level graphics error. 
% To prevent another crash in this session, MATLAB is using software
% OpenGL instead of using your graphics
% hardware. To save this setting for future sessions, use the opengl('save', 'software') command. 
% For more information, see Resolving Low-Level Graphics Issues. 
% > In matlab.graphics.internal.initialize (line 15) 



      end

% Beta(:,1)=[]; % 1st round left log header empty, but when calculating for
% different [t c] I am putting [t(s2);c(s2)] in each Beta header, to be
% able to tell what fill-out corresponds to each pair of curves.


%% Displaying beta(H0) for particular [t c] fill-up


   B(:,:,s2)=Beta;

   b1=Beta(1,[2:end]);
   b2=Beta(2,[2:end]);
   hf1=figure;plot(H0,b1/k0,H0,b2/k0)
   ax2=gca;
   grid on;
   grid minor;
   ax2.PlotBoxAspectRatioMode='manual';
   ax2.PlotBoxAspectRatio=[1 1 1];
   grid on;
   title('beta(H0)/k0');
   str1=[ ' t/a = ' num2str(t1/a) ];
   text(300,2.4,str1,'FontSize',14)
   xlabel('H_0');ylabel('\beta/H_0');

   legend({' \beta1','\beta2'},'Location','northeastoutside')

   str2=['graph_' num2str(s2) '_chap_09_ex_10']
   saveas(hf1,str2,'jpg')
        
end   


 

graph_1_chap_09_ex_10.jpg
graph_2_chap_09_ex_10.jpg

t/a<<1 is very thin ferrite, and t/a=.25 means half area of waveguide cross section is covered with ferrite.

graph_3_chap_09_ex_10.jpg
graph_4_chap_09_ex_10.jpg
graph_5_chap_09_ex_10.jpg
bottom of page