tumeda
|
Hello Experts, Thanks for your answer to my previous questions! It is a new question. :P
According to the paper of Ken, to fast simulation, the VCO and divider(DF) are writte in a behavioral block. So I can get the "Jdf=std(DF_Output_periods)". With different numbers of the sample points, the "Jdf" are also different. How many samples could be enough to let the "Jdf" stable for a PLL Circuit?
In addition to, I also use the VCO and Divider in individual Blcok, the I can directly get the "Jvco=std(VCO_ouptu_periods)". But I can not get the right relationship between the "Jvco" and "Jdf", which should satisify the form "Jvco=Jdf/sqrt(N)" (Table in page42).
How many sample points are enough to calculate the jitter performance of the whole PLL circuit? 1000, or 10000, or even more? ???
|