Hello

I'm puzzled about 2 reasonings about the phase noise/jitter relationship of a VCO and a subsequent divider (e.g in a PLL):

1. As stated in Ken's papers, a divider with N-stages sums (in a squared sense) the jitter from the VCO
   -> Jit_divider = √N×Jit_vco

2. The relationship between the VCO and divider phase is given by:
      Φdiv = Φvco÷N               (1)

    The VCO RMS jitter (time-domain) can be expressed as:
       JitRMS,vco = Φvco÷(2π)×Tvco
     
    Since Tvco = Tdiv÷N and since (1) is valid this expression can be converted to the jitter equation for the divider
    (the 'N' terms cancel), showing that the VCO and divider jitter are equal!

Clearly, both reasonings contradict eachother, but which one is wrong and what is the error in the reasoning?

Thanks in advance!

Visjnoe