Why a spinning top does not fall over

Tilt a top that is not spinning and it falls over at once. Tilt the same top while it spins and the axis instead swings slowly around in a circle, with nothing propping it up, and the faster the spin, the slower the circling. That slow circling is called precession.

Why it matters

A push that produces motion at right angles to itself feels like a trick. It is just the ordinary rule for rotating things, and once you have it, the trick becomes a tool. A spinning wheel resists having its axis tipped, so it is a reference direction you can carry around. That is a gyroscope. Satellites and space telescopes steer by tilting spinning wheels on board, and the reaction swings the whole spacecraft.

The mechanism

Angular momentum is the rotational version of momentum: for a spinning top, it is an arrow that points along the spin axis, and its length grows with the spin rate and with how far out the mass sits. Torque is a twist. Gravity pulls down on the top's centre of mass, which sits above the pivot and off to one side when tilted, so gravity applies a torque about the pivot.

The rule that connects them is the whole story: a torque changes angular momentum in the direction of the torque. The torque from gravity on a tilted top points horizontally, at right angles to the tilt. Not down. Sideways.

For a top that is not spinning, the angular momentum starts at zero. Gravity's sideways torque creates a small horizontal angular momentum, a rotation about a horizontal axis: the top falling over. Nothing surprising.

For a spinning top, the angular momentum arrow is already long and points along the axis. Gravity adds the same small horizontal nudge to its tip. But a sideways nudge on the tip of a long arrow does not shorten or tip it; it swings it. The tip moves a little way around a horizontal circle, the torque turns with it, and the next nudge is again sideways. The axis walks around the vertical. That is precession.

The rate follows. The nudge per second is fixed by gravity; the arrow it moves is as long as the spin makes it. So the angle swept per second is the torque divided by the angular momentum. Double the spin and the top circles at half the speed. This is the result Bolina derives elementarily, and it holds as long as the precession is much slower than the spin.

One more detail in the figure. Released from rest, the axis first dips, then rises, then dips, tracing a scalloped path around the circle. That bobbing is nutation. The sideways motion must build up from nothing, and the small dip supplies it. Tanriverdi's classification of these motion types shows how the path depends on how the top is started. At high spin the dips are so small and fast that you see only the smooth circle.

In one breath

Gravity's torque on a tilted top points sideways, not down. With no spin, that sideways torque is the top falling. With spin, the same torque swings a long angular momentum arrow around a circle, and the longer the arrow, the slower it swings. The top does not fall because the push that would drop it has been turned into a slow walk around the vertical.

It does fall eventually. Tip friction and air drag bleed off the spin, the precession speeds up and the bobbing grows, until the top topples. Tanriverdi and Erbasan model exactly that slow loss.