I test out the AD9914 evaluation board.
I use this DDS with direct low noise 2.5 GHz clock and I use output frequency ranges where the FTW are exact integer to the system clock and where these last produce no phase truncation spurs.
However I observe unknown spurious in output spectrum around the output frequency and around harmonics of the output frequency, and the narrow-band SFDR is greater than -70 dBc (Attached are examples of output spectrum plot).
How can I explain these spurs that are too close to output frequency? How can I reduce them?