PDoA crystal question

I see there is a comment in PDoA node source code:
/* For best PDoA performance the clock offset between PDoA node and the tag

  • should be >2ppm.
  • In case Tag’s crystal is very off, it also will be trimmed to stay in the [2 … 4]ppm range,
  • as defined below:
  • */

so my question is : why we need at least 2ppm clock offset between tag & node to get a good PDoA performance?
I thought the clock offset should be as small as possible.

DW Remark : this product is now known as Phase Difference of Arrival (PDoA). The end result is the same as it gives the relative position of the tag using the node as a reference.

[color=#222222][size=small][font=Arial, Helvetica, sans-serif]The main reason is the signal-correlated quantization noise - to measure a very small values of phase differences of arrival this small clock offset is enough to get rid of this noise.[/font][/size][/color]