When the rotor is vertical as illustrated most of the flux will prefer to go through the bridging stator piece due to a shorter flux path which has a lower permittivity. When the rotor spins 90° there will be no reason for the flux to still go through the bridging piece, and hence no flux would be going through it. Going from maximum flux to 0 flux will of course cause the coil to generate a voltage.

In the illustration a tiny air gap can be seen as well. I noticed in experiments and femm simulations that having the right permeability is key. Put simply the toroid must have a lower permeability than the bridging piece. If this is not the case the flux will mostly travel through the toroid and not the bridging piece.
The permeability of the toroid can be artificially lowered by introducing an air gap. The air gap need not to go all the way through, a small cut will do however it needs to be symmetrical on both sides or you'll create an asymmetric setup which is not advantageous.

The back torque associated with most electrical generators can hardly be seen in this setup even the cogging due to the air gap is minimal. This has to do with the fact that the field interactions happen mostly transverse to the motion of direction.

Sadly I destroyed my silicon steel toroidal core which I was experimenting with when trying to saw a small airgap through it however the second one is on its way and thought to share the concept in the meantime.

Hi broli,

Thanks for showing this setup drawing, seems good to me too.     I mean I tend to agree with your reasoning on the back torque.

Greetings,  Gyula

I wonder if rotor has more then one pair of permanent magnets around the circle

As you rotate the rotor you will be generating eddy currents in the ring stator... this will cause a drag on the rotor, heating of the stator, and is of course a consequence of the changing flux in the stator as the magnets sweep past. Isn't this also a consequence of Lenz's Law?

Hi TinselKoala,

Yes but either using laminations or ferrite ring the eddy current issue could be minimized.  Of course machining either the air gap in the ferrite or forming the stator from laminations surely makes tinkerer's life harder...   

Gyula