Hey EZ, i really hope you can help me out with this. I have the following issue:

i have an AD9910 at frequencies of 200-300 MHz to drive some acousto-optics for the lab. I detect a phase noise, not periodic, with a peak noise deviation of 0.005 pi radians. My question is now if that could be spur-related. The phase resolution of the AD9910 is about 0.1 milliradians, so thats 50 times less than what i can see on my signal. Could that be spur related?

Also, how does the AD9910 - or a DDS in general - handle the phase accumulator overflow? If, lets say, my phase increment is N and my 32-bit phase accu has a current value of 2^32 - N + 1 is the next increment then leading to an overflow and the accu starts at value 1 (would lead to some phase noise due to truncation) or does it start at 0? Is the carry handled in any particular way?

Best regards,

Mike

Regarding accumulator rollover, if the current accumulator value is 2^32 - N + 1 an increment of N would yield an unbounded result of 2^32 + 1, so the final answer is 1. That is, the accumulator counts modulo 2^32, so its maximum count value is (2^32) - 1.

Phase noise can come from a variety of sources, so it is not possible to answer your first question definitively.