the big thing for me was not witch motors had free wheels but how good was the free wheel
In my first answer
above I was only concerned with whether drag existed, not how much in differing motors. Drag has nothing to do with how good a freewheel is, whether pawl type or more commonly the ball and ramp type, they all have virtually no drag and disconnect the motor. But as I've noted before, they don't disconnect the orbital gears which are the source of the drag on internally geared motorised wheels.
The variations in drag of different motors has varying causes.
1) Larger hubs can often be lower drag since they can have larger diameter and narrower gears which create less drag.
2) Nylon gears are normally used for reasons of noise prevention, but they are quieter because nylon has more "give", and therefore less precision. In turn this means added drag.
3) The orbital gears are a drive reduction from the motor to the hub, and that's the easiest way, a small gear driving a large one. Unfortunately that means when running without the motor it's the opposite, the wheel is driving the gears, geared up, a large wheel (the toothed ring in the hub) driving small ones, so more drag. The amount of that drag is affected by the reduction ratio of the motor, typically from 5 to 8 times gear reduction. Of course on the overrun that means the extreme case of the 8 times reduction motor will have 60% more drag due to the orbital gears than the 5 times reduction one. Most have smaller ratio differences, but that still accounts for big variations in drag between motors.
4) Manufacturers do all they can to improve their motor's drag, so over time they can get better. Older Suzhou Bafang motors had thin orbital toothed rings, the latest one has a much thicker one. That will be both quieter and more efficient since it will be less subject to distortion under load, so in turn the rolling resistance will be lower.
.
