Other Parts Discussed in Thread: CLOCKPRO
I have seen the question and would like to ask further about the algorithm to use to find the values for the CDCE925.
The strategy for increasing N / M isn't what TI Clock Pro does, neither is trying to get R=0.
For example, if I calculate all the N / M values for 36.800 MHz output given a 27MHz crystal, TI Clock Pro gives N / M of 3312 and 486 respectively, but my own program, which calculates all possible values of N / M I find that 4048 and 495 are the highest values for N / M.
On further investigation there are lots of other instances where this happens. It is not obvious what is being optimized in TI Clock Pro.
I understand that there is a fixed frequency loop filter in the CDCE925 and that TI don't want to give away any IP, but is it possible to provide enough information to at least allow an algorithm that produces the same results as TI Clock Pro so it can be used to generate values "on the fly"?
All suggestions gratefully received.