I am interested in using an ADRF6755 in a product under development. I have been testing the eval board and I am seeing something I did not expect. I am feeding a common REF between two PLL chips with Frac-N capability, the ADRF6755 and another chip. When I program the two devices to the same frequency Int-N or Frac-N I am seeing that the output frequency of the two PLLs differ by a minute amount. If I am perceiving this correctly I believe they are off by 1/(2^26) of the comparison frequency. Can the ADRF6755 generate this small frequency offset from the programmed value?
As an example I have compared the other PLL Frac-N device with an alternate AD part (ADF4350) and it appears to synthesize a frequency that is accurate [exact] (Frac-N and Int-N).
Is there something I have configured in the ADRF6755 that will create this minute frequency programming error? This seems like it has something to do with the 25 bit Frac-N modulator, but I cannot manually set the modulator to a value that is an exact frequency lock. It is always either a minute frequency offset higher or lower at best.
Any insight on this would be helpful.