Originally Posted by phidauex
vrkelley, the answer given is basically correct. People argue alot about how much both precession and geometry play into it, but it is true that they both play a part. An experiment was once done where a bike was built that had a second set of wheels attached to the main wheels, and lifted off the ground, that rotated in the opposite direction of the main wheels, to cancel out gyroscopic precession, and the bike retained stability, though it was apparently a little more difficult to control. Both aspects contribute, though geometry of the bike is probably a greater contributor.
Additionally, rider input is NOT necessary for a bike to retain self stability. Take a bike, and get it rolling on a flat surface, it will wobble for a moment, and then enter a stable equilibrium where the bike weaves back and forth in a gently straight line as the front wheel moves back and forth under the center of gravity. When it slows down, it can't adjust the position of the wheel relative to the center of gravity fast enough, the weaving will get more erratic, and will eventually cause the bike to fall over.
Check out the simulations on this site for more detail and a graphic depiction of the process:
http://www.win.tue.nl/dynamo/bike/bike.html
peace,
sam
The self centering of the front wheel is an attribute of the front wheel trail which is a function of head tube angle and fork rake. With a large wheeled bike such a road bike or cruiser the, majority designs will self center enough for the gyroscopic effect of the tires to keep the bike upright. But on smaller wheeled bikes such as recumbents with more fork rake, you find the a hands free bike would not self center. The rider closes the control loop by steering the bike under fall-the inverted pendulum effect.