I've noticed that my 900GTO's coordinates, as reported by ASCOM (latest version of the driver) change over a session where I am autoguiding. This occurs whether using MaximDL or NINA/PhD. In both cases, I am configured to send autoguider commands via the ASCOM driver rather than a separate ST4 cable.
This confounds automated meridian flips because the mount "thinks" it is pointed somewhere other than where it actually is. Often, the difference is too slight to be significance, but sometimes not.
As an example, last night I was doing 30 minute shots of the Crescent Nebula. Over one hour, as reported by the RA/DEC fields written into the FITS header, I saw a difference of 6 minutes, 4 seconds in RA and 1 minute 7 seconds in DEC.
If I leave the mount running with no autoguider connected at all, RA and DEC remain constant, as I would expect.
Is this expected behavior or am I possibly missing some configuration option? It seems as though the autoguider commands are being interpreted as user initiated slews.