I'm using the AD9548 Eval board (AD9548/PCBZ) to evaluate this IC.
The on board reference is an crystal oscillator at 50MHz +- 10ppm of frequency accuracy.
I generate a system clock of 1GHz and 160MHz of free running DDS freq. The output divider of out0 is programmed to divide by 4 then the output frequency is 40MHz.
Measuring the output frequency using a counter (locked to a stable frequency reference) the frequency is 39.999850 MHz. I expect that this value should be more precise because with a stability of +- 10ppm @ 50MHz I expect to have a maximum errror of +- 400Hz @ 40MHz.
Anyone can help me to understand why the frequency error is bigger than expected?