Why i am getting matrix misdimension.

 i want to implement simple equations.

S=cos(2*pi*fc*t)

Here t[n] is number of samples. Second equation i want to implement is

S1t[n]=cos(2*pi*fc*t[n]+ ϕ(t[n])

    Where ϕ(t[n])= ϕ0(t[n])+∑_(k=0)^n▒ϕ1(t[k]) ), here summation goes from k = 0 to n

Where ϕ0(t[n]) is random variables with gaussian distribution

Where ϕ0(t[n]) is also random variables with gaussian distribution

 ts = 0.01;
  fs = 1/ts;
  nsamples = 100;
  t = 0:ts:nsamples-ts;
  fc = 50;
  S = cos(2*pi*fc*t);                    %%%equation 7.1 noisless
  sigma = 1/1000;
  phi_sync = sigma * rand(1,nsamples);   %%%%random variable with gaussian distributed

for k = 1: nsamples
phi_acc(k,:) = rand(1,nsamples);       %%%%random variable with gaussian distributed
phi(k, :) = phi_sync + phi_acc(k,:);   %%%%equation 7.11
S1(k,:) = cos(2*pi*fc*t + phi(k,:));
end

    %%%%equation 7.10, signal with noise
  figure, plot(fftshift(abs(fft(S1))));
  return