I've tried a few methods including the phd drift align, alignmaster, but have settled on the CCD method described here
http://www.observatory.digital-sf.co...nt_CCDv1-1.pdf
I haven't measured how close I am, but unguided exposures seem quite good.
one major reason for me to use this was balance.
you need to drift align using the main scope and if you're using the guide camera in a guide scope, then you need to remove it and fix it in the main scope which throws it out of balance.
so the CCD method works for me since I'm using the main CCD in the main scope and don't need to change anything.
Other two methods I'm to try are using PolarAlignMax that uses plate solving with Maxim and EQAlignmax. Not sure how good either are.
With Alignmaster, I have replaced the sterne.txt file but find that the star pairs are usually very low.