To analyse the L-shaped stars. Arcturus, M3 and M101 images taken 2011-05-21.
Arcturus and M3 unguided; M101 6-min. images were guided both with and without PEC.
The 5th row of graphs are screenshots of PHD guiding. Autoguiding scale is 0.93as/p. The RMS equates to approx. 1/2 arc-second.

The current hypothosis is that the mount's minimum move is 0.5 arc-second and the imaging scale is 0.60 arc-seconds. An alternate theory is this is another anomoly of the 8300 chip at work on brighter stars.

Unguided images of SAO100801, a short exposure of mag 6 star, to avoid saturation, taken 2011-05-23. Stars appear Gaussian (if that is the correct term), which exonerates the camera and leaves the guiding.
The next step is to test guiding using the supplied Lodestar cable rather than the ASCOM driver. (Tested 24-May-2011 - no difference.)
The eventuality is an AO device.

26-Jun-2011: There has been speculation in a Yahoo group of a phenomenon known as "trefoil" that may be causing the lobing. After re-examining the short exposures of the mag. 6 star, that may be the case. I need another set of eyes to look at this.