Ram,

It is difficult to simulate the jitter of VCO in HSPICE.

I used to measure the jitter of VCO by adding a small series resistor on power supply. But it just introduce a ripple of the same period as the VCO. So the period of VCO is still constant.

I also tried to measure the jitter by tuning the power supply a little, say 100mV. Then I measure the difference of the period (T_p1) before tuning and the period (T_p2) after tuning.

T_jitter = T_p1 - T_p2

Obviously it is not accurate because we don't know the amplitude of power supply noise and ground noise.

So the problem is modeling the substrate noise and other inteference. The real jitter of a PLL/VCO is very senstive to its environment, including swithing noise of digital circuit, SSN noise of pads, even electromagnetic radiation. The real jitter is difficult to predict because the huge digital circuit is not controled by the circuit designer who design the PLL.  

Best regards,
Yawei

Ram,

The T_jitter is just a crude estimation. I don't know how to convert it to rms jitter. sorry.


Best regards,
Yawei