Took a little time away from the new racer project and made a cheap and easy to build electric 2WS demonstration bike. I've known about the effect of 2WS on lean angle for decades, but never effectively demonstrated it. The end of the video has frame grabs that illustrate the 2WS effects on lean angle and front wheel steering angle.
Numbers:
- Weight with battery 62 pounds.
- Wheelbase 53 inches.
- Head tube angle at both ends 0 degrees.
- Trail at both ends 1.4 inches.
- Weight distribution with rider 60% F | 40% R.
- RWS settings 0%, 50%, and 66%.
- Hubmotor 48V 1000W.
- Battery capacity 20 Ah.
- Top speed 24 MPH.
I've never tested 2WS beyond 50% RWS - was afraid it might become unstable past that point. But not knowing something is supposed to be good motivation to learn something rather than rationalize FUD - do the research! At 66% RWS, it worked far better than I expected - it never did anything evil, even if it felt weird at first. While good low speed behavior doesn't always correlate with good high speed behavior, good research results encourage development work.
The rear hubmotor was sold as a front hubmotor for an electric bicycle "Pedal Assist" system - it won't go unless the rider is pedaling. No crank means no pedaling, so the sensor trigger ring was chucked in the lathe, bored out, and pressed onto the hubmotor bearing boss. The sensor bracket was cut down and welded to the fork dropout. It works very well - not sure how legal that is! The 3 speed settings give a maximum indicated 14, 19, and 24 MPH. After the motor controller is turned on and bike is rolled forward a few inches, it takes off from stops with just the throttle.
I wanted the steering ratios to be as accurate as possible. A simple jig that clamped around the fork crown made that easy. It only takes a few seconds to change the RWS settings by repositioning the link's rod ends on different bracket holes. Careful bracket placement made wheel realignment unnecessary.
The initial plan for observing changes in lean angles was to use a common digital speedometer and an old JVC action camera aimed to the outside of the turn and bolted to a welded-on frame bracket - low risk of inconsistent results with a consistent and repeatable perspective. But then the idea of using a Speedangle Apex lap timer (With lean angle indicator, GPS speedometer, and data logger!) sounded like an even better idea. Experience ended up proving otherwise - the Speedangle unit gave very inaccurate readings with 2WS engaged - even worse at 66% than 50%. It is otherwise a very nice unit, but obviously not developed for unconventional motion analysis. So back we went to the first layout: Consistent speed (15 MPH) over a consistent radius (42 feet) gave the consistent results needed. I can't tell you the lean angle difference measurements, but those differences are obvious.
While this is obviously not an SAE level technical publication or doctorate level dissertation, the methods and results are clearly reproducible. There's enough information and breakthrough potential here to get really excited about it all if you can comprehend the possibilities. Meanwhile, it was very satisfying to roll something finished out of the shop, test something unproven with better than expected results, and actually ride.
