Assuming that I have a PSD of random noise on Vcontrol of VCO as discrete points from the simulator, eg. 50 points per decade. Now I have to simply find the phase noise at the output of the VCO. Do I :
A. assume the NTF is Kvco/f, do dot product of PSD and square(Kvco/f) at the corresponding frequency points?
                                                              OR
B. Now that the points are discrete, I assume each bin value is a sinusoid, convert the rms to peak by sqrt(2) and then get the USB/LSB as half the value and then dot product by square(Kvco/f) at the corresponding frequency points?

A and B differ by a scaling factor due to the sinusoid assumption.