To illustrate the problem, I have attached two AberrationInspector images. The first one is a 1200 second sub via an Ha filter which shows elongation at the lower left. The elongation is not that bad but this is because it has been partly adjusted out (I took this sub thinking that the elongation has been adjusted out but it has not fully). I then adjusted focus some more and took another 1200 second sub. The corner elongation is now gone. The amount of adjustment needed is not constant, so I cannot use the offset feature.
Left lower corner elongated (half way through adjustment):
Elongated mostly gone:
The problem with autofocus is that, if you look at the centre panel the stars are very similar. I suspect as far as the autofocus routine is concerned, the two focuser positions produce similar quality stars. Because most stars are in the centre, the centre is therefore more heavily weighed and the corner stars do not contribute much to the star quality measurement. For autofocus to work with the FSQ85, one needs to exclude the centre so only the corner and edge stars are evaluated by the autofocus routine.