Hi Tony,
The process you follow looks OK to me, its fundamentally what I do, the only exception is I do continue on after a 2 star alignemnt and make the minor adjustments to bring the error as close to zero as I can. Drift alignment is only as good s how well you can differentiate the 2 path lines. Errors can still be there.
I too still have issues with the telescope not slewing to the target accurately but it is almost always in the FoV of a long EP (>15mm). I attribute it to stacked up errors during setup and aligning. I don't think it is ever possible to get the setup perfect every single time if you constantly have to tear down and setup. I envy the guys with piers and observatories.
You'll find that because most of the imaging people use guiding, polar alignment accuracy is not as important. It's only when you are doing visual that it becomes critical. However, most visual observers probably aren't too fussed as it is easy to make adjustments as you go, and use star hopping to get around the sky. I do understand your frustration with the GOTO function and it should work if everything else is perfect.
I also just thought of another contributor, the mounts mechanical movement will have play in it. So even if you get the PA perfect, the 2 star alignment might add errors if you don't finish the alignment of the star with and "up" and "right" movement. This keeps the gears meshed in the forward RA motion. One of the best mods you can do is replace the gears with belts (
http://www.rowanastronomy.com/produc...tm#heq5beltkit.)
Hoping this all helps a bit...