Bad results after curve fitting a spectrum to a lorentzian curve.

I am trying to fit a power spectrum to a custom equation of a Lorentzian form using the curve fitting toolbox. I have used the NonLinearLeastSquares and LinearLeastSquares methods with the same results. If I don't specify starting points the values are far off what it is expected, if I specify starting points (with the NonLinearLeastSquares method) the results are the same starting points. Am I missing something? I have tried even fitting just one variable with no success.

I have attached two variables with the type of signal to fit and below is my code:

if true
  %Fitting the curve to a lorentzian curve
      f = FitFreq;
      Sx = fittype('((1.3806488e-23*300/k)/(pi*Q*fo))/((1-(f/fo)^2)^2+(f/(Q*fo))^2)',...
                  'indep','f');           
      options = fitoptions('Method','NonlinearLeastSquares');
      [fitmodel,goodness,output]  = fit(f,FitData,Sx,options);
      Qf = fitmodel.Q
      kf = fitmodel.k
      fof = fitmodel.fo
      FittedTo = ((1.3806488e-23*300/kf)/(pi*Qf*fof))./((1-(f/fof).^2).^2.+(f/(Qf*fof)).^2);
  %Plotting of the curves
      plot(FitFreq,FitData,'b','LineWidth',0.5)
      hold on
      plot(FitFreq,FittedTo, 'r','LineWidth',0.5)
      ylabel('Magnitude (AU)')
      xlabel('frequency (Hz)')
      hold off
end