You could have any offset you want if you mount the motor above the yaw pivot.

These things will always tend to have a high resistance as they are designed for high voltage. Remember you are trying to extract power at a fraction of their design rpm, you must expect only a fraction of the original rating.

It so happens that for wind duty with the good thermal contact between the wire and the core and a howling gale blowing over it they can be grossly overloaded compared with the original rating and still survive. This means that you can get a lot nearer the original rating than you would for continuous duty at low speeds in engine driven applications. You must expect the full load efficiency to be low. This lets the prop run clear of stall and you get more power from the prop than the "stalled" air gap machines.

This comes at a greater prop speed than the air gap machines and this higher speed also helps to get more power from a small well overloaded unit.

If you can find the gap flux using FEMM then you can predict the output voltage if you know the turns ( unlikely with an original winding).

I was interested in that last one that Peter wrote about. The gap flux is surprisingly low ( not much above an air gap machine) so it looks as though you could throw a lot more magnet at it although the 4 pole arrangement would be very difficult without proper curved magnets. These strings of little magnets don't seem to get the flux up as far as I would have expected, but with all the gaps between them, approximations to the stator curve and the relative ineffectiveness of small magnets then I suppose this is not surprising.

At least you folks are less likely to have starting problems with iron loss compared with the commercial Chinese machines that have well designed magnets to fit properly. They no doubt get more watts for a given size machine with lower low wind efficiency and possibly better high wind results.

The last motor (3kw 3ph) I rewound was this style. the coil span is 12 10 08  and the no of slots will be 36.

One leg per slot because it is consequent wound (half the poles don't physically exist as wire poles, but as phantom poles). This gives only 18 coils (3 per pole (the other three don't need to exist as wire wound poles) x 6) and then  2 legs per coil x 18 = 36 slots.... 1 per slot.

The diametrically opposite coils are of the same phase and pole orientation (ie. N facing N..=U 120 degrees around finds N facing N =V and another 120 degrees around finds N facing N =W, the phantom S poles are created between them to give 4 poles per 2 coil sets (2x3 actual coils).

Notice a coil span of 12 slots for effectively 4 poles but not 48 slots, only 36. The phantom poles crowd the ends of the 12 span, and  the whole lot pretends to be 4 alternate poles of 9 slot span instead of 2 like poles of 12 span on opposite sides of a 36 slot stator.

What this means to your magnet spacing is I think Zubbly says your magnet width = pole coil span... here that can't happen, as you would need 48 slots worth of space and we don't have it, so a span of 9 slots for mag width should be Maximum.

Your Femm should bear this out I hope

Hope you can understand my ramblings on this.

Nice motor and will follow your adventures with interest. Reading your, Flux and Oztules' responses has taught me a new thing or two again :)

I'm still looking for an (affordable) copy of Rosenberg myself. Parts of it can be read online and they certainly seem very interesting.

All the motors I've taken apart so far had only one coil leg per slot, so I suppose that method is really more common over here (Europe). In fact, until your post, I didn't know any better than that this was 'normal'. Another thing that continues to surprize me (and makes me jealous) is the fact that North-American motors tend to have a larger diameter-to-length ratio; the rotors are shorter but stubber. This should make for a better slow-running multipole genny. If you have a look at my 4-pole 3 hp conversion, you'll see that the rotor is much more slender: http://www.anotherpower.com/gallery/3HP-induction-conversion/motor_rotor_frontside

Was a bit surprised by Flux's remark that flux density of the 500 W conversion was low. It's nearly 1 T in the airgap above the magnets, whereas I understand that axial fluxes operate more in the 0.7 T range. (http://www.anotherpower.com/gallery/album69/blue_boy_FEMM_detail_average_flux_over_one_active_tooth)
But I must admit, I could have made airgap smaller, just that I have limited faith in my machining/building capabilities :) I don't recall having done a definitive FEMM simulation of the 10 hp conversion; that's another thing on the 'to-do' list for this weekend then...

You haven't explicitly stated so in your story but I understand you will use the stock winding and not rewind ? I forgot, but do you use 24 or 48 V system voltage ?

Peter.