My failure with toroidal-wound motors.

I want to share with you not only my successes but also my failures. For the past few months, in my spare time, I have been experimenting with unconventional winding methods for BLDC motors. For instance, using toroidal winding, which could allow manufacturing motors on machines designed for winding toroidal transformers. This could potentially reduce their production costs. Modeling in Ansys also yielded positive results, so I decided to manufacture such a motor and test its operation.

I bought a used hoverboard and made a new stator with a new toroidal winding for one of its two motors (photo at the top of the article: left is the motor with toroidal winding, right is with traditional winding). Afterward, I measured their parameters. Despite good fitting and nearly identical winding parameters, the efficiency comparison between these motors showed a significant difference: the toroidal one lags behind the factory one by 28-45%.

After reflecting on the reasons for the failure, I have come to two conclusions:

1. I overdid the number of turns. With each subsequent turn of the wire, reactive resistance decreases while active resistance increases. The optimal configuration is considered to be when the winding window width is half the core width. It seems I exceeded this parameter - too many turns aren't always beneficial.
2. There might be a less obvious mistake: I made the same number of teeth and coils, but the magnetic field closure for each phase on the magnets was uneven, which could have led to decreased efficiency.

In the diagram, you can see that with traditional winding, the magnetic field of three coils closes on three teeth, whereas with toroidal winding, the magnetic field of three coils closes on four teeth.

I will continue experimenting in this area and will definitely share with you what I achieve. Sincerely, Iaroslav Burcov.