Help me find out what I'm doing wrong in this project

My assignment is to make a monte carlo simulation of a binary code of +/- 3 volts going through Gaussian noise and then to determine the bit error probability and signal to noise ratio and plot them vs each other. However when I run my program and get my plot it looks completely random and there is no noticeable trend or curve. I have my code done and I know it is at least half right because my BEPs were working fine before. If anyone sees what I am doing wrong I would love some input.

close all;clear all;clc;

sigma = 1;

Tb = zeros(1,100000);

Tv = zeros(1,100000);

Rb = zeros(1,100000);

BEP = zeros(1,100);

MSTv = zeros(1,100);

MSN = zeros(1,100);

SNR = zeros(1,100);

SNRdb = zeros(1,100);

n = 1;

while n <= 100

Tr = rand(1,100000);

i = 1;

while i <= 100000

      if Tr(1,i) >= 0.5

          Tb(1,i) = 1;

      else

          Tb(1,i) = 0;

      end

      i = i + 1;

end

Tbbins = [0 1];

Tbcounts = hist(Tb,Tbbins);

j = 1;

while j <= 100000

      if Tr(1,j) >= 0.5

          Tv(1,j) = 3;

      else

          Tv(1,j) = -3;

      end

      j = j + 1;

end

N = sigma * randn(1,100000);

R = Tv + N;

k = 1;

while k <= 100000

      if R(1,k) > 0

          Rb(1,k) = 1;

      else

          Rb(1,k) = 0;

      end

      k = k + 1;

end

errors = 0;

m = 1;

while m <= 100000

      if Rb(1,m) ~= Tb(1,m)

          errors = errors + 1;

      end

      m = m + 1;

end

BEP(1,n) = errors / 100000;

MSTv(1,n)= var(Tv) + (mean(Tv)*mean(Tv));

MSN(1,n) = var(N) + (mean(N)*mean(N));

SNR(1,n) = MSTv(1,n)/MSN(1,n);

SNRdb(1,n) = 10 * log10(SNR(1,n));

n = n + 1;

end

figure(1)

semilogy(SNRdb,BEP,'m-<', 'linewidth' ,2.0);

axis tight

[Image Analyst]

Look at this to format your code properly: http://www.mathworks.com/matlabcentral/answers/13205-tutorial-how-to-format-your-question-with-markup Or else just attach your m-file with the paper clip icon.

[Walter Roberson]

Have you considered, e.g.,

Tb = Tr >= 0.5;

instead of looping?