You are correct in that I threaded the shaft/axle/whatever-it is.  In my windmill, a 7/16-NC die fit over the shaft, but was rather loose.  This has the excellent effect of cutting threads but not decreasing the diameter of the shaft at all.  I got a nut with the nylon insert and used that for a precise tension adjustment.  At one point, I was considering using another shaft.  I got some threaded stock with the intent of welding it into place.  The threaded stock was REAL sloppy, so I shimmed it with a .38 case that I had sawed (yes, I have other hobbies :-) )  The brass from the .38 case was the perfect size for a shim.

Another suggestion is 2 nylon nuts with a jam nut for each.  I taper them toward the bearing with a mill file so they don't bind quite so bad.

>BTW, the bearings are not very replacable. I totally ruined one set getting them out for a closer look.  I >always start with at least 2, so I still have a spare blade!

The bearings on mine really stink, too.  Lowest bidder, I suppose.

I'm going to try several of the suggestions here; they're good.  Thanks.

Check the wiring. It should be connected 3 phase star / wye.
Test each phase seperately for volts, they should all be the same.
(10 coils and 10v, 1v / coil, on 2 phases. But the 3rd phase has 8v, that means the 8v phase has 1 backwards coil. Or maybe 2 shorted out coils.)

Check the resistance of each phase- they should be close to each other.

Maybe one or more of the coils is connected backwards?
Each backwards coil cancels out a good one, then the whole system gets out of whack.

If a paper clip sticks to the metal behind the magnets, you need more metal back there.

I can't tell in the stator pic's.  What coating on the metal did he use for the laminates?
If electric can get through the lams, that is where the power is going, eddy currents.

I have never seen lams in that way.  That direction.  If they are individually insulated, maybe its OK.  Flux, UGL, anybody but me, would know a lot more.

If the lams are plastic coated coat hangers(?), I wonder how much metal is in there compared to the space. For 1 thing, it's round, leaving space between them, and the plastic is going to use a lot of space too.  Just sound like not much total metal in the space available.

If it is wired right, and none of the coils is backwards or shorted, I would change the lams.

After going over everything 3 times, and not finding a problem, maybe rectify the 3 output seperately, add a large capacitor to the output of each bridge, and connect them in series. It should add the voltages, but reduce the amps. Again, I don't know how well that would work- it's just what I would try.  I think it was maybe Jerry who did this to get more volts, and it worked for him.  Some guys do it that way for low voltage stepper motors too.