I learned an algorithm for winding a BLDC motor. This algoithm has the windings crossing several stator teeth similar to this image:

Windings crossing multiple teeth

When I google BLDC stator, I see way more images like this : enter image description here

In this image, the windings mostly go around a single tooth. They do not span multiple teeth awkwardly. This second version looks much easier to wind... So my question is, the second version... Does winding a motor like that make it lose efficiency? If so, how much efficiency? I am building just a practice motor, so I don't really care if its 100% efficient if it would be easier to wind!

Also, can anyone recommend some good modeling software to model the magnetic fields? I saw this program "motorsolve" but it looks expensive!

  • \$\begingroup\$ The ones I have seen only span one "tooth". Where did you learn this algorithm? Are you sure the algorithm was not for induction motors? \$\endgroup\$ – mkeith Nov 18 '16 at 2:40

In short: go for the second motor!

Motor topologies

The second motor has what the litterature refers to as concentrated windings, and the former has distributed windings (and also stator skew encompassing a whole slot).

As I see it, the main benefits for concentrated windings compared to distributed windings are that they:

  • Are simpler to wound => cheaper
  • Have less end windings (more important in short machines)

The main drawback is, as you suspect, losses. The concentrated windings have more harmonic contents in the magnetic flux, due to not having the smoothness gained by distributed windings (and also sometimes stator skew as seen in your first figure above). These harmonics lead to significantly larger magnetic losses in the iron core.

I learned a lot about this by reading Design of Rotating Electrical Machines, by Pyrhönen et al, so that’s my recommendation if you want to dive deeper.

Magnetic Fields

How to simulate magnetic fields is actually very well known. There’s no need for expensive tools. I used the free FEMM solver for a PhD level course in numerical motor analysis. It can compute quasi-static frequency dependent solutions, and get really nice results when compared to reality. It has some examples to get you started.

From a quick inspection of the Motorsolve package: I think you get nice pre-made geometries with boxes to fill in your sizes and shape factors etc. So the drawback of using a free solver is that you have to draw the geometry yourself. But that’s exactly what I did in the course I mentioned earlier: complete reverse-engineering of a professional machine sawed into two. I can recommend it as an exercise.

  • \$\begingroup\$ Thanks a lot your comment is really helpful. Numerical motor analysis, what did you do, run a ton of finite element analysis on motors to see results of more or less magnets/coils/winding combinations and work through that textbook you mentioned? Def about to read that paper. Probably that textbook also. The book I read left a lot for the imagination. I should probably review it on Amazon. \$\endgroup\$ – Frank Nov 19 '16 at 0:10
  • \$\begingroup\$ I used FEMMs ability to connect to Matlab (or octave as a free alternative) to parametrize the topology from the motor we dissected during the course. Then we compared the performance of the machine to a simulated machine based on standard classical analytical induction machine optimization techniques and found that the commercial machine was a lot better. We also tried to vary some parameters and find a better solution, but, it turns out that it is hard to beat an industry that’s over a 100 years old :) If you like my answer, please mark it as accepted, it really helps a lot! \$\endgroup\$ – vindarmagnus Nov 19 '16 at 10:06

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.