I am using evaluation board EVA-ADF4350EB1Z revision B and Agilent E5052B to measure phase noise. I have found that measured phase noise is very close to phase noise presented in ADF4350 datasheet, but only when output divider is set to 1. For higher division coefficients phase noise doesn't match to reported results. According to figures 7-9 in datasheet phase noise decreases with increasing output divider coefficient. In practise I cant see this effect. For example for PFD=25Mhz, VCO=4.4GHz, N=176 phase noise versus different output divider settings are in the table below:
|Output divider||output frequency MHz||phase noise at 10kHz offset dBc/Hz||phase noise at 10MHz offset dBc/Hz|
|1||4400||-89 (-90)||-148 (-148)|
|2||2200||-94 (-96)||-148 (-150)|
|8||550||-93 (-108)||-144 (-154)|
Number in brackets are taken from figure 9 in datasheet.
Unfortunately I don't have result for divider 16 because output frequency 275 MHz significantly attenuated by output matching circuit in evaluation board and signal analyser uses 3th harmonic 825MHz as a crier.
Does anybody have any ideas why difference so obvious, especially for output divider 4 and 8.
Thank you for any help.