This is something I looked at 3 or 4 years ago and then forgot about. I used a CAD-style program that does element analysis and adds up magnetic force vectors. If you look at some of the Perendev-motor”work” by others on the net, there is the issue of finding an arrangements of the magnets so that they don’t find a local minimum in the gradient field and get “stuck there”; the solutions usually involved using multiple “rotors” with a staggered configuration of the outer stator magnets. I thought “why not stagger the magnets on a single rotor/stator pair in such a way that a single pair will never find a local minimum in a full rotation?”. The software (Vizimag) can tell you the total force vector on each magnetic element in the diagram.