The Nut That Flipped in Orbit: The Dzhanibekov Effect and the Tennis Racket Theorem

In the summer of 1985, a Soviet cosmonaut floating inside a freshly rescued space station gave a small metal nut a flick and watched it do something that looked like it broke the laws of physics. The nut spun off its bolt and drifted across the cabin, rotating neatly about one axis. Then, with nothing touching it, it flipped over by half a turn. A few seconds later it flipped back. Then it did it again, and again, at regular intervals, for as long as he cared to watch.

The cosmonaut was Vladimir Dzhanibekov, and the strange flip now carries his name: the Dzhanibekov effect. Physicists know it by two older names, the tennis racket theorem and the intermediate axis theorem, and the surprise is not that it happens. The surprise is that almost everyone, on Earth or in orbit, has seen it without noticing. Toss your phone, a book, or a tennis racket into the air with a spin, and you can watch the same flip in about half a second.

It is a cousin of the physics behind Euler’s Disk, the desk toy that screams toward a sudden stop, and of the tippe top that turns itself upside down. All three come back to the same 18th-century equations for spinning solid objects. This one also nearly embarrassed America’s first satellite, and it answers a question that occasionally goes viral: could Earth itself suddenly flip over?

A dead station and a drifting nut

Dzhanibekov was not in orbit to do physics demonstrations. In early 1985, the Soviet space station Salyut 7 had gone silent. When the crew of Soyuz T-13, Dzhanibekov and flight engineer Viktor Savinykh, launched from Baikonur on June 6, 1985, nobody knew whether the station could be saved. They found it rolling slowly about its long axis with its solar panels pointing in random directions. Dzhanibekov, a veteran pilot with four previous flights, flew the Soyuz in by hand, matched the station’s rotation, and docked with it. It was the first time anyone had docked with a dead, uncontrolled space station.

Inside, the station was frozen. Frost covered the walls and equipment, the batteries were flat, and the crew worked in fur-lined hats while they traced the fault to a failed sensor in the solar-array pointing system. Over the following days they recharged the batteries, restarted the heaters and got the attitude-control system running again. The rescue was dramatic enough that Russia turned it into a feature film, Salyut 7, in 2017.

Once the station was alive again, supply ships could bring up spare parts and cargo. As the popular telling goes, Dzhanibekov was unpacking supplies held down by a wing nut. (Some accounts describe a T-shaped nut.) When he spun the nut off its bolt, it kept spinning as it floated away, and it kept flipping. In weightlessness, with no table, hand or gravity to interfere, the flips were obvious and eerily regular.

The story is often told with a dash of Cold War mystery: a widely watched 2019 Veritasium video says the observation was kept quiet for about a decade. Whatever happened to the cosmonaut’s notes, the physics itself was not secret. It had been in print since the 1830s, and it shows up in standard mechanics textbooks such as Herbert Goldstein’s Classical Mechanics. In 1991, three mathematicians, Mark Ashbaugh, Carmen Chicone and Richard Cushman, published a paper in the Journal of Dynamics and Differential Equations with the wonderful title “The Twisting Tennis Racket,” analyzing the extra half twist a thrown racket picks up.

Three axes, and only two of them behave

Every solid object, no matter how lumpy, has three special directions it can spin around, called its principal axes. They are always at right angles to each other. Each comes with a number, the object’s moment of inertia about that axis, which measures how hard it is to get the object spinning that way. Mass far from the axis counts for a lot; mass close to it counts for little.

Take a tennis racket. Spinning it around the handle is easy, because most of the mass sits close to that line: that is the axis with the smallest moment of inertia. Spinning it like a wheel, around an axis that goes straight through the strings, is hardest, because both the long handle and the wide head swing far out from the center: that is the largest. The third axis lies in the plane of the head, crossways to the handle. Its moment of inertia is somewhere in between, which is why it is called the intermediate axis.

Here is the rule. Spin an object about its smallest axis or its largest axis, and the spin is stable: small wobbles stay small. Spin it about the intermediate axis, and the spin is unstable. Any tiny error, and there is always a tiny error, grows quickly until the object swings over and is rotating “upside down” relative to where it started. Then the process runs in reverse and it swings back.

That is what happens when a tennis player flips a racket into the air for one full turn and catches it by the handle: in almost every throw the face has also made a half turn, so the side that was facing up is now facing down. It is what Dzhanibekov’s nut was doing. The nut was simply rigid enough, and the cabin quiet enough, for the flips to go on and on.

Euler wrote the equations, Poinsot drew the picture

The mathematics goes back to Leonhard Euler, the Swiss mathematician whose name is also on that spinning steel disk. In his 1765 treatise on the motion of solid bodies, Theoria motus corporum solidorum seu rigidorum, Euler wrote down three linked equations, now called Euler’s equations, that describe how a spinning rigid body’s rotation changes when no outside force twists it. In those equations the rates of spin about the three axes feed into one another, and the strength of that feedback depends on the differences between the three moments of inertia.

Run the numbers for a spin near the largest or smallest axis, and the feedback acts like a spring, pulling small wobbles back. Run them near the intermediate axis, and the signs work out the other way. The feedback pushes the wobble to grow, and it grows exponentially, doubling again and again until it is no longer small at all.

In 1834 the French mathematician Louis Poinsot turned Euler’s algebra into something you could see. In Théorie nouvelle de la rotation des corps (A New Theory of the Rotation of Bodies), he showed that a freely spinning body must keep two quantities fixed: its angular momentum and its rotational energy. Each of those conditions defines an egg-shaped surface, an ellipsoid. The body’s spin can only travel along the curves where the two surfaces cross. Poinsot even coined a name for the path the spin axis traces inside the body: the polhode, Greek for “path of the pole.”

Poinsot’s picture makes the tennis racket theorem almost obvious. Around the largest and smallest axes, those crossing curves form small closed loops, so a slightly-off spin just circles nearby forever. Around the intermediate axis, the curves form an X, like the roads crossing at the top of a mountain pass. A ball balanced exactly on the pass could stay there, but nudge it and it runs down one side, around the mountain, and back up to the pass from the opposite direction. Each trip around is one flip.

There is a neat detail hidden in that geometry. The closer you start to the exact intermediate axis, the longer the object lingers before each flip, but the flip itself always happens fast. That is why the motion looks so startling. The nut seems to settle down, then turns over almost all at once, then settles again.

The satellite that would not spin straight

For a perfectly rigid object, spinning about the smallest axis is just as stable as spinning about the largest. Real objects are not perfectly rigid, and in early 1958 that difference turned into a real-world surprise for America’s first satellite.

Explorer 1 lifted off from Cape Canaveral on the night of January 31, 1958, on a Juno I rocket. The satellite and its spent final rocket stage orbited together as a single pencil-shaped body, about 14 kilograms (31 pounds) in all. To keep it steady, the rocket’s spinning upper stages left it rotating about its long axis at 750 revolutions per minute. Like a rifle bullet, it was supposed to stay pointed that way. For a long, thin object, that is the axis with the smallest moment of inertia.

Explorer 1 also carried four flexible wire “whip” antennas sticking out from its middle. Radio operators tracking the satellite soon saw a slow rise and fall in signal strength on top of the flicker from its fast spin. By the end of its first orbit it was clear that the spin was dying away. Within a short time the satellite had swapped its rotation for a slow end-over-end tumble at about 9 revolutions per minute, according to a 1960 JPL report on the Explorer electronics.

The culprit was energy loss. As the satellite wobbled, the whip antennas flexed back and forth and turned a little of its spinning energy into heat. Losing energy is allowed; losing angular momentum is not, because nothing outside was twisting the satellite. And for a fixed amount of angular momentum, the lowest-energy way to spin is about the axis with the largest moment of inertia. So the satellite slid, step by step, from its intended spin to a flat spin around the axis it was never meant to use.

Stanford radio scientists Ronald Bracewell and Owen Garriott explained the effect in Nature in 1958, in a short paper called “Rotation of Artificial Earth Satellites.” (Garriott later flew in space himself, aboard Skylab in 1973.) The lesson became a basic rule of spacecraft design, often called the major axis rule: a spinning spacecraft that flexes or contains sloshing fuel will end up spinning about its axis of greatest moment of inertia unless something actively stops it. JPL dropped the whip antennas for Explorer 3, and in 1966 a JPL engineer, Peter Likins, wrote a long report on energy dissipation and spacecraft motion that traced the whole problem back to that first launch. It was, in effect, the first big addition to Euler’s theory of rigid bodies in roughly two centuries.

Explorer 1 still did its job. Its Geiger counter, built under James Van Allen, found the radiation belts that now carry his name. Those belts later played a starring role in Starfish Prime, the 1962 H-bomb test in space.

So could Earth flip over?

Every few years, the Dzhanibekov effect resurfaces online with a doomsday twist: if a spinning nut can suddenly turn over, could the whole planet? The short answer is no, and the reason is the same rule that caught Explorer 1.

Earth is not a perfect sphere. It bulges at the equator, which makes its moment of inertia about the polar axis slightly larger, by roughly a third of a percent, than about any axis through the equator. So the planet already spins about its axis of greatest moment of inertia. That is the most stable choice available, the same flat-spin state Explorer 1 settled into. There is no intermediate-axis instability to trigger. Earth’s day-night spin, which Foucault’s pendulum made visible in 1851, is not about to reverse itself.

Earth does wobble a little, exactly as Euler’s equations say a slightly off-axis spinner should. Using the planet’s known flattening, Euler predicted a free wobble with a period of about 305 days. When American astronomer Seth Carlo Chandler finally measured it in 1891, it turned out to take about 433 days, a gap Simon Newcomb explained by the fact that Earth is partly elastic rather than perfectly rigid. Today this Chandler wobble moves the point where the spin axis meets the surface by about 9 meters (30 feet). It is a gentle circle, not a flip.

Some much smaller bodies do tumble. The asteroid 4179 Toutatis, about 2.5 kilometers across, rotates in a complex tumbling state rather than around one fixed axis, possibly because the YORP effect, a tiny twist produced as sunlight is absorbed and re-radiated as heat, drained and scrambled its spin. In general, though, bodies larger than a few kilometers lose energy internally and settle into rotation about their axis of greatest moment of inertia, just like Explorer 1.

Try it yourself

You do not need a space station. Grab a hardcover book and wrap a rubber band around it so it stays closed. Toss it up spinning about each of its three axes in turn: around an axis parallel to the spine, around an axis that pokes straight through the cover, and around the remaining axis, which lies flat in the plane of the cover and runs from the spine across to the open edge. The first two come down looking just as they went up. The intermediate one comes down with an extra half twist almost every time. A phone works too, if you do it over a bed.

It is a nice reminder that some of the most counterintuitive physics is hiding in plain sight, like the self-steering bicycle that scientists misunderstood for a century. Life in orbit just removes the distractions. Astronauts in weightlessness deal with plenty of stranger effects, from floating tools to bones that thin without gravity, but few are as easy to reproduce at home as a cosmonaut’s flipping nut.

Euler’s equations were written for planets and spinning tops. Poinsot’s ellipsoids were drawn for geometry lovers. More than a century later, a cosmonaut in a fur hat, an American satellite with floppy antennas and a tennis player tossing a racket all ran into the same quiet fact about the universe: of the three ways to spin a solid object, only two are stable, and the middle one always wants to turn over.

Further reading