Why does a spinning top stay up longer?
A still top tips over after a tiny lean. Spin it quickly and watch again: its axis can keep its direction, helping it stay up longer.
A rotating object resists a sudden change in its axis direction. Gravity acts on the tilted, fast-spinning top, and the axis slowly circles around. This is called precession. For the same top, that circling is usually slower when the spin is faster.
Friction and air resistance slow the top. Its wobble becomes more noticeable, and eventually it falls. Follow two things: how fast the body spins and how its axis circles.
Twist one yourself and time its dance
- Twist the top. Launch the same top once gently, once hard. Count how long each stands, and notice whether the slow one wobbles in faster, wider circles. Slide any weighted ring up and down the stem: the farther the centre of mass sits from the tip, the readier it sways.
- Spin a coin. Set a coin spinning on its edge on the table. Watch it begin to sway widely as it slows, circling as it falls—a coin is a pocket-sized top running through the whole chain in under a minute.
- Bicycle wheel. With a removed wheel, hold the axle in both hands while it spins and try to tilt it—the axle refuses to dip, and instead swings sideways; lift the front wheel and shake the handlebars, then get the wheel spinning fast and shake again: the bars feel stiff. Note: that is a wheel held in your hands; riding a bike stays upright through constant steering corrections and the rider too—the spin only lends a hand (see bike balance).
Predict first: with the experiment’s centre-of-mass height raised, does the precession get faster or slower? Twist again and check. The readout obeys Ω ∝ d and Ω ∝ 1/ω.
The precession formula, and Earth as a giant top
More: a closer explanation
Rest a top upright on the table and it topples at the first nudge: its centre of gravity hovers above the point, and any lean lengthens gravity’s arm until it falls. But twist it into a fast spin and it stands steady on its point—when spinning fast enough, a push usually does not topple it either, merely drifting the axis a little sideways.
The secret is angular momentum. Picture the top as countless pieces: spinning, every piece circles the central axis, each carrying speed, each wanting to run on along its tangent. Their combined insistence on keeping the turning is a vector pointing along the spin axis:
Angular momentum is as stubborn as momentum in a straight line: with no outside torque, its direction refuses to change. A leaning top feels gravity, and gravity exerts a torque about the tip, τ = mgd·sinα. In the fast-top approximation, that torque barely resizes L; it mainly swings its direction. So instead of falling, the axis circles the vertical—precession. Twist faster and L grows, so the same torque turns it more slowly: faster spin feels stiffer. Once the spin slows, the approximation breaks down and the top visibly sways and falls.
Friction at the tip and drag through the air steal the spin bit by bit. L shrinks, the turning grows more urgent, the circles widen, and swaying, the top lies down. From standing to fallen is a ledger of energy slowly collected.
Open: from gravity’s torque to Ω = mgd / Iω
A top leaning by α feels gravity mg acting at the centre of mass, a distance d from the tip, giving a torque about the pivot:
The torque moves the tip of the angular-momentum vector at dL/dt = τ. With the axis tilted α from vertical, the horizontal part of L is L·sinα, and walking that tip around the vertical gives the precession rate Ω. The sinα cancels on both sides (the fast-top approximation, the standard result in OpenStax’s chapter on rotational motion):
So doubling the spin halves the precession, and moving the centre of mass farther from the pivot speeds it up. On screen the spin is slowed about ninefold and the precession about 0.4-fold so both stay watchable at once—a declared teaching exaggeration that leaves every ratio intact.
Open: Earth precesses too, and the pole star changes
Earth is a giant top: its axis leans about 23.4°, and the Sun’s and Moon’s pull on the equatorial bulge supplies a torque, so the axis likewise refuses to fall and instead sweeps a slow circle—one full precession in about 26,000 years. When the pyramids were raised, the pole star was Thuban in Draco; today it is Polaris; in about 12,000 years Vega will take the post. The different constellations seen through one year mainly follow Earth’s orbit around the Sun. That is distinct from the much slower axial precession; see NASA on constellations.
Open: packing that stubbornness into phones and rockets
A fast-spinning rotor whose axis resists changing direction—that property, boxed up, is a traditional mechanical gyroscope. Rockets and satellites hold their attitude reference with gyroscopes and correct with thrusters (NASA’s outreach on spacecraft attitude control features them). The gyroscope inside a phone is usually not a tiny flywheel but a MEMS structure: a small part vibrating rapidly back and forth; when the phone rotates, it feels a tiny Coriolis effect and circuitry reads the change out—sensing rotation by a different route than a spinning wheel (see the full gyroscope exploration). The tabletop top, the held wheel, Earth and the mechanical gyroscope share “spin creates angular momentum”; the phone’s vibrating sensor is another way of sensing rotation.
Next time you twist a top, watch two things at once: where the axis points, and how much spin has been stolen. One axis that resists a sudden change of direction, plus a widening sway as spin drains away—spin fast enough, and it stands.
Model limits: how is falling simplified?
Fast precession uses Ω = mgd/(Iω). Near-upright stability uses the symmetric-top condition I₃²ω² > 4mgdI₁: I₃ is the axial inertia, while I₁ is the transverse inertia about the pivot. They are different quantities, calculated from the disc and pivot offset. The growing lean and final fall are smooth teaching interpolations, not a solution of full rigid-body nutation or tip slipping. Displayed fall times cannot predict a real toy. Reference: David Tong, Cambridge rigid-body dynamics.