Interesting analysis fom the PHD2 forum Jose. I can't speak directly to the MESU mk II because I don't have one. With my MESU mk I, I know it is out of balance, simply because it is almost impossible to balance it effectively due to the motors being constantly engaged on that model. It tracks and guides just fine, with guiding corrections in E and W . Given the clutches arrangement on the MESU mk II, it will be a doddle to balance, just use the process in the Starizona link for balance across the range of movement - that did catch me out a few years back when my mount was significantly out of balance, it tracked for a while then stopped, rebalance then it was fine.
Re flexure, differential flexure is a significant problem if it's present. I also had this when trying to guide with a separate guidescope which wasn't fixed absolutely rigidly to the main scope. Subsequently I had this again when trying to guide a C11 sct with an external guide scope, which I susbsequently found out is not possible, you have to use OAG to guide SCTs because of the movement of the primary mirror as the scope tracks. (there is a body of evidence for this assertion). So one good way to remove the potential influence of differential flexure between the main scope and the guide scope is to temporarily move to an OAG setup if you can. Just thinking out loud here 'cos i've never tried it, but what about using your main imaging camera as the guide camera for PHD2? This would be a no cost solution. You could use the PHD2 viewer to select a star, let it track and guide and then inspect the log.
Also, going back to Yann-Eric's idea of using PHD2's guiding assistant, the PHD2 helpfile on this utility states that you need to use your imaging camera (not the guide camera) for that process anyway.
I think you're getting to an interesting point with things now - good luck.