Analysis of Guiding - 4h run, reasonable guiding...but...

[CKemu]

So I recently imaged M82, and fairly pleased with the results, having captured both RGB and Ha, but I like to autopsy my results, nit pick and generally try to learn more with my rig.

So this data is from March 24th, whilst I imaged M82 in ha, using 1800s exposures, when stacked it comes out like this:

(I was using an Red frame from a prior run as a "reference" to make sure my stars aligned, hence the tilt/black border).

Looks okay, pretty round stars, though pixel study seems to indicate a slight movement.

So I then stacked each frame on top of each other in Photoshop, using a "lighten" mode per layer, but without aligning the stars, so "as is", and I noticed this movement to the left/up.

]

Now there was a little wind that night, that made DEC go a little haywire and the end of the run I clunked the scope in to the tripod...ooops. But below is my calibration and a snippet of the guide log, any ideas as to why I am getting this slow drift, I thought PHD2 would guide "any" movement out.

For extra information, I am using an OAG (ASI120mm as the camera), PHD2 for both drift align and guiding, and the mount is an NEQ6. I think I forgot to re-calibrate after the drift align, so calibration was based on initial rough polar align.

[lukebl]

The rotational nature of the movement suggests field rotation to me. i.e. Polar Alignment not quite right. 

I have a similar problem with my images inexplicably drifting over time, despite perfect guiding. However, mine are in a straight line indicating a different problem.

[almcl]

The angle between the calibration axes (never been able to understand the figures that PHD lab gives - so just looking at the graph) also hint at a slight polar alignment error.

How did/do you PA for that run?