I found that I have some 1" thin-walled tubing from an exercise machine that has a nearly 22mm diameter — the outside diameter of the bearings I bought. I figure I can weld a piece of tubing to the ends of the windmill frame then cut a slot and use a seat-post clamp to grip the bearing.
I also tested the motors I had selected as candidates for a 10-watt generator. None of them performed better than the motor I had originally selected.
Based on my former calculations, the rotors would only run at 300 RPM in a 10 MPH wind, so a 7x gear-up would only yield 2,100 RPM which wasn't enough to generate any appreciable power. To get any power out, I had to run the baby carriage wheel in the drill press at 620 RPM with a 0.33" gear, so the ratio to the 7 3/4" wheel (23.5x) means I ran the motor at 14,500 RPM. If this were 20 mile-per-hour wind, then a 70 mile-per-hour wind would overcrank the motor to over 50,000 RPM and probably blow it up. I'll have to see how slow I can get the motor to run and still get decent power. Well, I know that 3,100 RPM is not enough: the motor, when hitched directly to the drill press, would only generate about 3.4 volts open and 1.9 volts into 3 ohms (only about 1 watt.)
Ok, so let's say I limit the motor to 15,000 RPM at maximum wind speed of 70 MPH, or 2,100 RPM. The ratio, therefore, is 7:1 … bummer. I think I'm going to need to get a different motor. If I use 21,000 RPM as the cap, I can get to a ratio of 10:1, so in a 10 MPH wind, the rotors would only run at 300 RPM driving the motor at 3,000 RPM which isn't enough.
None of the bicycles I have currently have the correct dimensions for using the triangular front-end as-is on the back of the Bike With 2 Brains, so I'll have to custom build those parts. It's probably for the best anyway since I don't like the size of any of the top brackets.