Hi all,
I was trying to design a charge pump PLL and simulate the phase noise of it. I ran into some convergence problem when I used PSS+PNOISE, therefore I decided to calculate the PLL phase noise with transient nosie analysis discussed in the site below:
http://www.cadence.com/community/blogs/cic/archive/2009/03/26/calculating-large-signal-phase-noise-using-transient-noise-analysis.aspx
Following the setup described in the article, I could obtain the phase noise very close to my calculation. However, the phase noise offset frequency has a very limited range. The article says the phase noise plot will extend from fmin=4/tstop to fmax=fosc/2, where tstop is the transient noise simulation stop time and fosc is the oscillation frequency. In my simulation, fosc=2GHz, and I ran the transisent analysis for 10us, therefore I should get a phase noise plot from 400KHz to 1GHz. However, I only saw the phase noise plot from 2MHz to 1GHz (as shown in the figure). I also tried to increase the transient analysis time, but that didn't help. The other transient setup parameters that I have are
Noise Fmax=20GHz, Noise Fmin=100KHz, Noise Seed=1, Noise Scale=1, Noise Tmin=1ps and errpreset=moderate
Since I am also interested in the frequency offsets lower than 2MHz, I really want to extend the phase noise plot to the lower frequencies. I guess the problem is because of my wrong setups of noise Fmax and noise Fmin, but I was actually not quite sure how to set them. Does anyone have any experience on this? Why the phase noise plot from transient noise analysis cannot extend from 4/tstop to fosc/2?
Thank you in advance!
Faye