Greg,
Have you tried setting Star Mass Tolerance to 100? Setting this to 100 disables the star mass tolerance feature. You also have control over the tracking box size and can increase this to larger than 15x15 however the risk is you end up with multiple stars in the box and this can impact centroid determination. Using darks also helps with guide star S/N if you aren't already doing so.
"'Search region' - specifies the size of the "tracking rectangle", in units of pixels. You may need to increase this value if your mount does not perform well or, more commonly, if it's not well-aligned on the celestial pole.
'Star mass change detection' - tells PHD2 to monitor the brightness and size of the guide star compared to the sky background.
'Star mass tolerance' - if 'star mass change detection' is checked, PHD2 will trigger a 'lost star' error if the measured brightness and size vary by more than this percentage. This might be useful if you have two stars inside the tracking rectangle and you want to be sure that PHD2 doesn't mistakenly switch stars. It can also prevent errors caused by thin clouds, high camera noise, or alpha particle "hits"; but it may be unreliable if you are guiding on a faint star. If you are getting too many 'lost star' errors when the star is plainly visible on the display, try increasing the value of this setting. Setting the value to 100 effectively disables the warnings entirely."
http://openphdguiding.org/man-dev/Ad...tm#Guiding_Tab
I'm a bit surprised you are having issues with this - what system are you guiding with?
Cheers,
Peter