See this Comm Center post by Andre Ioda. He says the FWHM measurement is unreliable and inaccurate compared to CCD Inspector.

He noted that CCDI uses many more stars. Maybe change PP to do that, or just use every one of the stars, maybe with outlier rejection?