WELL I GOOFED!
I can ride around, and make what looks like DC volts... but it seemed to be measuring pretty low. Even at 10,000rpm I was only reading about 11V on my gauge, with all my lights off and the battery disconnected. I could not for the life of me figure out what I was doing wrong.
Just had an epiphany, pulled out my magnet on a telescoping stick, and popped the flywheel off.
I assumed the flywheel was 4-poles (i.e. N-S-N-S in a 90-degree orientation).
With my coils spaced 120 degrees apart, this would have given me 3-phase, 120deg AC, perfecto for making beautiful full-wave rectified DC voltage with a tiny ripple that could be smoothed out by a battery.
I slid the magnet around my flywheel, and guess what. It has 6 poles (N-S-N-S-N-S in 60-degree orientation).
So, I'm rectifying 3 phases that are completely in phase, it turns out... which doesn't accomplish much in the way of making smooth DC.
The way I see it, I have a few alternatives...
1) Keep the coils at 120 degrees. Wire them in series to have a high-voltage single-phase AC output, rectify it, and use the dirty bouncing wave to maybe charge the battery. It's an AGM battery, it shouldn't care too much?
2) Switch to four coils, in a 60-120-60-120 orientation (like stock 6V Peugeot 5-coil setup). Wire the in-phase coils in series, and have two AC outputs. Rectified, this will provide the exact same shape as option 1, but with a little less voltage and more current.
3) Figure out some way to re-magnet the flywheel?????????? Having a 4-pole flywheel would make everything perfect. Don't most mopeds use 4 poles?
4) Make a custom stator plate with a 90-degree orientation. It's a simple enough shape (just a 2-d cutout with some provision to hold the coils in place). Don't have the tools for this, though.
5) Any of the single-phase setups with rectifying and some sort of smoothing. Maybe if I throw enough volts at it, the battery will be sufficient? (edited)