Please refre to the attached DOC file describing the settings and expected output frequencies with plots and setup as described below:
1. In positive correction the algorithm is working correctly.
- If the fractional is different than 0 the algorithm is working correctly.
- In fractional equal to 0 and negative correction:
- There is a shift in the correction value around 34Hz when we are using reg 28e & 28f.
- If we will insert main RF freq of 2599.999984MHz the value of reg 273 is EA.
- If we are using reg 28e & 28f with -16Hz correction the value of reg 273 is F9.
reg 274 is FF and reg 275 is 7F.
- In this case the value of fractional reg (273-275) is 7FFFF9 (h) = 8,388,601 this value is bigger than the modulus 8388593.
It looks like there is a problem in the algorithm of the RF PLL update when using the correction word reg 28E and 28F.
What could be the reason?
Thanks for your support