applets.

Spinning Things

← All apps

About Spinning Things

Every solid object has three special axes at right angles to each other, its principal axes. When its three moments of inertia are all different (a book, a phone, a racket; not a ball or a plain cylinder), one axis is easiest to turn about (smallest moment of inertia), one is hardest (largest), and one is in between. Thrown spinning about the easiest or the hardest axis, it keeps spinning that way, wobbling a little. Thrown spinning about the middle one, it would only stay put if the spin were exactly about that axis; the tiniest wobble grows until the object turns right over, then back again, over and over. Nothing twists it: there is no air and no hand, and although gravity still pulls on it in orbit, it puts no twisting force (torque) on it about its centre of mass worth counting. This page shows it happen, and why.

IN ORBIT: pick an object and an axis, and press SPIN. The numbered rods are the three axes (1 easiest, 2 middle, 3 hardest; "spun axis" means the one you picked); the dashed arrow L is the angular momentum, which never changes. The strip of pictures under it shows moments through the first flip (for the middle axis; the same moments for the others), and the screen below plots how fast the object turns about each of its own axes. When it flips, the rate about the axis it was spun on goes through zero and comes out the other way. The card counts flips and gives the time between them, measured and Expected.

THE MAP: the same motion seen from the object. Its angular momentum, as the object sees it, can only move along these curves on a sphere. Round axes 1 and 3 they are small closed loops, so a wobble stays small. Through axis 2 run two special curves (the separatrix) that lead away from it and round to the other side. A flip follows a curve just beside them: the separatrix itself would take for ever to arrive, and a start exactly on axis 2 never moves at all. Drag to turn the sphere; click (tap) on it to start from anywhere, and press TRACE. Tick Floppy to let the object lose energy the way a slightly bendy body does.

WHY: a block whose three sides you set. The chart shows how a small wobble behaves for a spin about each axis: about axes 1 and 3 it goes up and down and stays small; about axis 2 it grows by the same factor every turn, until the object flips.

TRY IT: how to see it with a phone, a book or a tennis racket, and a thousand pretend tosses with a hand that isn't perfectly steady.

THE STORY

In June 1985 the cosmonauts Vladimir Dzhanibekov and Viktor Savinykh flew Soyuz T-13 to the Salyut 7 space station, which had lost power and frozen, and brought it back to life. Dzhanibekov noticed that a nut spinning off its bolt (it is described as a wing nut or a T-handle nut) flew on spinning, but turned itself over every few seconds. The effect is now often named after him.

The physics was old. Leonhard Euler wrote down the equations for a turning solid body in the 1750s, and in 1834 Louis Poinsot, in his Théorie nouvelle de la rotation des corps, gave a geometric picture of the motion: the curve the angular velocity traces on the body's ellipsoid of inertia, which he called the polhode. THE MAP shows a close relative of his picture, the angular momentum seen from the body, moving on a sphere. That spinning about the middle axis is unstable was in the textbooks long before 1985, often as the "tennis racket theorem": toss a racket so it turns end over end with its face up, and it usually comes back with its face turned over. In 1991 Ashbaugh, Chicone and Cushman published a paper on that half twist, "The twisting tennis racket".

Real objects bend a little. Flexing turns some of the spin's energy into heat, but cannot change the angular momentum, and with less energy for the same angular momentum a body ends up spinning about its hardest axis. Explorer 1, the first American satellite, launched in January 1958, was meant to spin about its long axis, its easiest one; it soon began to wobble, more and more, because its flexible whip aerials were draining energy. Spacecraft designers have planned for this ever since.

The Earth will not flip over. It is slightly flattened, and spins about its axis of largest moment of inertia, the stable one. It does wobble a little: the Chandler wobble, with a period of about 14 months, moves the poles by a few metres. Stories that the Earth could suddenly turn over like the nut mix this up.

HOW IT WORKS

Each object is built from boxes and cylinders with real sizes and masses (the book 23.4 by 15.6 by 3 cm and 750 g, the phone 14.7 by 7.15 by 0.78 cm and 170 g, the racket 68.5 cm long and about 320 g with frame, throat, grip and strings; the T-handle's sizes are made up, 8 cm across, in steel). The page adds up their moments of inertia, finds the three principal axes, and numbers them 1, 2, 3 from easiest to hardest.

The motion comes from Euler's equations for a body with no twisting force (torque) about its centre of mass. The page splits the energy into three parts, each of which on its own just turns the body steadily about one of its axes, and does those turns exactly, in a fourth-order pattern (Yoshida's). So the length of the angular momentum is kept exactly, and the energy to better than 1 part in a billion: the card shows both. The Expected time between flips comes from Jacobi's elliptic functions (the formula in Landau and Lifshitz's Mechanics, section 37), and the run matches it.

For a small wobble about axis k, the two other spin rates feed each other. Their growth (or swing) rate is the spin rate times the square root of (Ib − Ik)(Ik − Ia) / (IaIb), where a and b are the other two axes. That product is positive only when Ik is in the middle: then the wobble grows exponentially. Otherwise it swings round and stays small.

The floppy body on THE MAP is a simple stand-in, not a model of a real satellite: the angular momentum, seen from the body, is nudged a little towards lower energy at every step, keeping its length. The tosses on TRY IT tip the spin axis by a random angle from a bell curve with the spread you choose, in a random direction, then follow the object for the time in the air with the same equations; there is no air drag, and the object is perfectly rigid (a book that opens in the air won't do this).

FURTHER WATCHING AND READING

FLIGHT DATA FILESPIN TEST

Object
Spin about
Wobble
Spin
Show at
Time in orbit0.00 s
Flips0
Time between flips, measurednone yet
Expected time between flipsnever
Spun axis tips away from L by up to0°
Energy kept to
Angular momentum kept to