iron eagel

One other thing I would try is to run it without the power meter in series with the batter, believe it or not the voltage drop through insertion loss caused by the meter can play with you, Then try it with the highest data rate (shortest duration) selected for data logging and you may see some abnormality in the data output. As I said before if you have your data logging set up for long duration you may not be able to see any transient abnormality that may be occurring and the higher rate would give you a better view of what is happening in the minute or so of testing you are doing...
Perhaps MJD could look through his setting with the same motor and have you set yours up the same way. Where his is working, it would give you some know good setting with that motor to work with, and see what type of results you get.

Also try it with a smaller or no prop if it works I would say it isn't the timing (per say) it will spin at a higher rpm with no or less of a load, that would indicate that it is voltage or current related.
I know this may be frustrating, but in electronics that's par for the course.