Right Ascension is hours:minutes: seconds
Declination is degrees:minutes: seconds
Angles are degrees:arcminutes: arcseconds (sky-linear by definition)
15 seconds of RA is 1 arcsecond (at the equator)
1 second of Dec is 1 arcsecond (anywhere)
Easy so far, now down to the nitty-gritty
If D is the declination then
RA/cos(D) is 1 arcsecond sky-linear (at Declination D)
So the question is: PHD2 just considers correction in sky-linear arcsec and isn't concerned with RA seconds or Dec seconds right? If we want to do, say, comet tracking we need to enter into PHD2 correction of arcsec per hour.
So if we go to JPL Horizons which gives us RA and Dec corrections we need to convert those to PHD2 arcsec per hour. Right?
Dec is easy - 1 second = 1 PHD2 arcsec.
But RA is tricky. JPL gives us "d(RA)/dt is multiplied by the cosine of declination
to provide a linear rate in the plane-of-sky." So I would expect to enter the JPL values into PHD2 because the tracking history is sky-linear (right?) So corrections should be sky-linear (right?). But PHD2 appears to want non-linear values here - it appears to want us to divide the JPL values by cos(D). Now I have a headache. The only agreement between JPL and PHD2 is that they both use arcsec / hour.
PHD2 documentation: "If you are getting the rates from the MinorPlanetCenter web site, you should choose the option for 'Separate RA and Declination coordinate motions'. PHD2 will automatically adjust the rates to compute the apparent motions in the sky." I couldn't find anything useful at
https://www.minorplanetcenter.net/. It may be there but I couldn't find it. I use JPL.