I like TSX's polar alignment routine (using T-Point) very much: just run 40+ calibrations, get a polar alignment error. Use the Polar Alignment routine (re-center a star) and usually I get a very accurate polar alignment. Furthermore, can do this before it's really dark!!
But when I setup my scope at Likely and polar aligned I was surprised to find that I still had a noticeable polar alignment error (judging by the PHD2 guiding graph that showed the DEC axis drifting). I made sure that I entered the right coordinates (lattitude, longitude, elevation). But still the same. The polar alignment routine worked without any problems though.
... and the next day it dawned on me! The computer on the scope does not have an internet connection (the wireless signal in the area where you setup your scope at Likely is very weak). Which means that Dimension4 can't synchronize the clock on the computer! And when I compared the times between the scope computer and my laptop (which I took up to the club house and connected it there) I could see that they were not in sync :-(
After having used the TSX polar alignment so long, I am not sure how accurate my RAPAS still is. So, I used Alignmaster. I'm pretty sure that it doesn't require an exact timing as it only measure the error when slewing from one star to another. I did that and the alignment error was MUCH smaller (though still not completely gone - but it could be guided out and didn't cause any visible image rotation).