First, make sure the focal length of the guide scope is correctly set. If you are uncertain what it is, temporarily use your guide scope and camera as the Main camera in ASIAIR, set the focal length for the main camera to zero, capture an image, and do a plate solve. ASIAIR will report the correct focal length of the Guide scope. If ASIAIR fails to plate solve, submit the image to astrometry.net and it will solve for the image scale for you, and from there, you can figure out the focal length of the guide scope, knowing the pixel pitch of the camera).
How many steps did the RA axis and the DEC axis move when the Guide calibration process was running?
Per PHD2 documentation (read the PHD2 documentation on the web before trying to use auto-guiding on ASIAIR), you need about a dozen steps per axis. If you have to wait for too many steps, the step size is too small and you are wasting time during calibration. If there are not enough steps, the calibration data will be inaccurate.
Adjust the Calibration Step parameters in the Guide Settings window to get between 12 steps and 20 steps per axis. If it takes more than 24 steps, stop the calibration, increase the Calibration Step Size, clear the calibration, and start again. Decrease the step size if calibration completes in fewer than 10.
If it is taking too many steps and the mount is not moving enough, you have other problems. Make sure the mount is balanced (unlock the clutches and the mount should not move) or the mount's motors don't have enough torque for the payload. At most, there should be just a small bias in the RA axis to keep the gears meshed at all times to reduce the effect of backlash.
If, after calibration has finished, you are not seeing any graph plotted, the error might be larger than the Y axis of your graph. Go into the Guide window and click on Y button on the left of the graph at the bottom of the Guide window (not the floating guide window in the Preview window). Y:+/-8 arc seconds gives you the coarsest graph.
There can be many reasons that the errors are large and take a long time to (or never) settle down. Among other things, again, insufficient balancing, a mount that has insufficient torque for the payload, and a mount that has too much backlash. Take the visual payload capacity of the mount that the manufacturer states, divide it by 2 and that should be the maximum imaging payload you try to place on your mount.
Unless your mount uses Harmonic Drive gears, be sure to carefully balance both axes, and also make sure the payload is symmetrical so there is no bias for all RA and DEC positions (some people call this the "third axis balance" or "3D balance"; do a Google search for that).
Good luck,
Chen
