A wing nut that would not stop tumbling
In June 1985, cosmonaut Vladimir Dzhanibekov was working aboard the Salyut 7 space station when he spun a wing nut off a bolt to send it floating across the cabin. Instead of drifting with a steady spin, it rotated smoothly for a few seconds, abruptly flipped 180 degrees, spun the other way for a few seconds, flipped back, and kept repeating that cycle indefinitely. It was not a trick of zero gravity or a malfunction — it is a real, purely classical consequence of rotating a rigid body that has three different moments of inertia about a specific one of its three principal axes.
Euler's equations for torque-free rotation
Any rigid body has three mutually perpendicular principal axes, each with its own moment of inertia. Label them so that I₁ > I₂ > I₃ — largest, intermediate, smallest. With no external torque, the angular velocity components ω₁, ω₂, ω₃ about those three axes evolve according to Euler's equations:
I₁ dω₁/dt = (I₂ − I₃) ω₂ ω₃
I₂ dω₂/dt = (I₃ − I₁) ω₃ ω₁
I₃ dω₃/dt = (I₁ − I₂) ω₁ ω₂
Conserved: kinetic energy T = ½(I₁ω₁² + I₂ω₂² + I₃ω₃²)
angular momentum magnitude L² = I₁²ω₁² + I₂²ω₂² + I₃²ω₃²
Both T and L are conserved throughout the motion, which means the tip of the angular velocity vector is forced to stay on the intersection of an ellipsoid (constant T) and a sphere (constant L). That intersection is exactly what determines whether nearby rotations stay nearby or run away.
The intermediate axis theorem
Linearise Euler's equations around a pure spin about each principal axis and check whether a tiny wobble grows or stays bounded. Spinning purely about the axis of largest or smallest moment of inertia gives a linearised system with purely imaginary eigenvalues — any small perturbation just oscillates gently around the spin axis and the rotation is stable. Spinning about the intermediate axis gives real eigenvalues of opposite sign — a small perturbation grows exponentially instead of oscillating. That single sign flip in the eigenvalues, sometimes called the tennis racket theorem because you can reproduce it by tossing a racket spinning about its middle axis, is the entire mechanism behind the effect: the energy-and-momentum-conserving trajectory near the intermediate axis is a saddle, not a centre, on the T-and-L intersection curve.
Why it flips instead of just wobbling forever
Because the unstable trajectory is a heteroclinic orbit — a path that starts arbitrarily close to spinning one way about the intermediate axis and ends arbitrarily close to spinning the opposite way about the same axis — the body does not simply drift away from the unstable rotation, it swings all the way around to the mirror-image spin, tracing a solution expressible in terms of Jacobi elliptic functions. Because both ends of that orbit are equally unstable, in the absence of any energy dissipation the process repeats: the wing nut flips back, sits near the (still unstable) intermediate rotation, and flips again, over and over, exactly as Dzhanibekov filmed.
Where it shows up beyond a floating wing nut
Any free rigid body with three distinct principal moments of inertia is subject to the same instability — tumbling asteroids and irregularly shaped moons can show it over astronomical timescales, satellite and spacecraft designers must actively avoid spin-stabilising a craft about its intermediate axis, and books, phones, and tennis rackets all demonstrate it with a well-aimed toss on Earth (where the flip is a single half-turn rather than a repeating cycle, because air resistance and the toss itself dissipate energy).
Frequently asked questions
Why doesn't this happen when spinning about the longest or shortest axis?
Because those two rotations are dynamically stable. Linearising Euler's equations around rotation about the largest or smallest moment of inertia gives purely oscillatory (imaginary) eigenvalues, so a small wobble just stays a small wobble. Around the intermediate axis the eigenvalues are real, so any tiny perturbation grows exponentially instead of oscillating, which is what drives the flip.
Does the object violate conservation of angular momentum when it flips?
No. With no external torque, the angular momentum vector L stays exactly fixed in the outside (lab) frame throughout the flip — what changes is the object's orientation and its angular velocity vector relative to L. The kinetic energy is also conserved; the flip is simply the trajectory that the tip of the angular velocity vector traces on the surface where both conservation laws are satisfied simultaneously.
Why did cosmonaut Dzhanibekov's wing nut flip repeatedly instead of just once?
Because in the absence of gravity, air resistance and friction there is no energy dissipation to settle the object into a stable rotation. Each flip returns the body to a state arbitrarily close to the unstable intermediate-axis rotation (just spinning the opposite way), so the instability triggers again and again, producing the repeating half-turn reversals seen in the famous 1985 Mir footage.
Try it live
Everything above runs in your browser — open Dzhanibekov Effect and change the parameters while it is running. Nothing is installed, nothing is uploaded, the whole model lives in one tab.
▶ Open Dzhanibekov Effect simulation