Ken, thank you for looking through the log and yes, you are correct, it does state that all autofocus events were carried out with the OIII filter set. The two instances of AF were as follows:
The first one after the move to the new target and the platesolve using the LUM filter:
[07/09/2016 23:35:22] [DEBUG] [Camera Thread] Auto focus is complete (best fit method: 1947; HFR 0.90)...
I thought that the focuser had moved more that I expected, so I stopped the sequence and reran AF and it did this:
[07/09/2016 23:43:21] [DEBUG] [Camera Thread] Auto focus is complete (best fit method: 1954; HFR 0.89)...
So, slightly curious as the log does suggest that AF was carried out with the correct OIII filter, but the two results are significantly different.
Anyway, my next task will be to fill all the Focus Point settings in the Filters screen and I will then use the LUM filter for focus and the delta settings info for when the filter changes. Hopefully that will give me consistent and correct autofocus in the future.
Thank you for your help.