Thanks, Terry. I had much the same thought.
Using the method from the GetBatteryVoltage sketch I saw an overlap between the power measured without a battery and that when a small lipo was connected, so voltage doesn't seem to be a very good discriminator.
Interestingly the battery connected (and fully charged) voltage was up to 4808mv, higher than the no battery voltage of 4768-4778mv.
Any such voltage is a disaster if applied to a single cell lipo. Maybe it'll burst out the back of the case ...
Henry