Reduce FM PLL noise
Please state which chip you are talking about, and also quantify the phase noise measured ?
Yes, if you use higher clock frequency it will decrease the overall phase noise.
The phase noise of a 38 MHz crystal oscillator is slightly higher than the phase noise of a 32 kHz crystal oscillator, but the overall phase noise of the crystal oscillator is reduced by the number (N) of the PLL reference dividers on a rate of 20*LOG(N).
As a thumb rule, would it be correct to say that the higher the feedback divider ratio N (for a lower frequency ref clock), the higher the phase noise ? Add to this that presumably the PLL must be capable of generating multiple outputs with that same fixed crystal, so presumably a fractional-N divider is used, which adds to the noise spurs output ?