You want to train PEC with the highest practical power. This has the effect
of magnifying the corrections to a point where the software can "see" even
the smallest changes. In your case, I would think that the C8 without the
focal reducer would be the best choice. Even if you always image with the
focal reducer, having the best possible accuracy in your recording can't
hurt.

You want to set a guide interval that will avoid any false readings due to
star scintillations due to seeing conditions. Something on the order of 3-5
seconds should be fine. The goal is to avoid any corrections other than
those related to periodic error.

A good polar alignment is a good idea, even though the AP mount has
programming to filter out Dec drift from RA periodic error.

Unless I misunderstand the AP implementation, the guiding rate during PEC
recording is independent of playback; it is recording the magnitude of the
error during the cycle, and will use that information to correct PE during
playback, at whatever guide speed you elect to use. If I am mistaken,
someone please let me know! <g>

However, in order to get a good recording, either .25x or .5x could be good
choices. Neither rate will cause the RA to reverse direction. From my own
experiences with the AP 600 GTO and a 9-micron pixel camera, I successfully
used .5x guiding rate to record PEC, and the end result was guide errors
under 1.5 arc-seconds.

Before recording PEC, of course, you should make sure that your backlash
settings (both in the hand controller settings and in software, if you use
software backlash settings) are correctly set. In addition, the end play of
your worm gears and the mesh of worms and driven gears should also be set
correctly. End play and mesh are complex topics that are just about
impossible to cover in email, so I'll assume that you've got those right for
sake of brevity. <g>

You can set backlash visually, or to higher precision with Maxim/DL using
the Move button in the Guide dialog. This opens a Move dialog, where you can
move the telescope in +X, -X, +Y, and -Y for precise intervals of time. This
allows you to verify whether there is any delay or jump when reversing
direction to a very high level of precision, much better than visual alone.
You may need to use a combination of hand controller and software backlash
compensation to get the highest level of accuracy. The AP hand controller
gives you nine settings to choose from. I recently adjusted backlash
settings for a Vixen SkySensor 2000 hand controller, which allows 999
settings, and I found the extra precision to be very useful. If Marj is
reading, high precision for backlash settings would be a useful upgrade to
think about for the future. Granted, backlash can vary somewhat over
different regions of the sky, and 999 settings is overkill, but for critical
long-focal-length imaging the extra precision is useful if you stay aware of
its limitations.

Ron Wodaski