Free Symmetric Top and Body-Cone/Space-Cone Precession

In the Cornell cafeteria, sometime in 1946, a student tossed a dinner plate. Richard Feynman, watching the red medallion on its rim, noticed that the plate's wobble and its spin ran at different rates — and, working out the ratio from "the motion of the rotating plate," rediscovered the equations on this page. He later claimed the whole chain of play that followed led to his Nobel Prize: "the whole business that I got the Nobel Prize for came from that piddling around with the wobbling plate." The wobbling plate is the free symmetric top: a rigid body with two equal principal moments ( $I_1=I_2\neq I_3$ ) and no torque on it at all. Its angular momentum $\mathbf{L}$ is nailed to the sky, its kinetic energy is locked — and yet its symmetry axis refuses to sit still. Instead the axis sweeps a steady cone around $\mathbf{L}$ , the visible wobble of a thrown frisbee, a spiralling football, or a tumbling satellite that has lost attitude control. Nothing torques these bodies. The wobble is pure inertia wearing a disguise. The machinery is a two-frame story. Riding the body, Euler's equations decouple: the spin $\omega_3$ along the symmetry axis freezes, and the transverse angular velocity circles it at the body rate $\Omega=(I_3-I_1)\omega_3/I_1$ . Watching from space, the axis and $\boldsymbol{\omega}$ both precess about the fixed $\mathbf{L}$ at $\dot\phi=L/I_1$ . The two descriptions fuse in Poinsot's lovely geometry — a body cone rolling without slipping on a space cone — and scale from a 175-gram frisbee all the way up to the Chandler wobble of the Earth itself, whose 433-day rhythm is this article's equations played on a planet.

Key takeaways

A torque-free axisymmetric rigid body (a free symmetric top) conserves both its angular momentum $\mathbf{L}$ and kinetic energy $T$ , yet its symmetry axis $\hat{\mathbf{e}}_3$ generally does NOT point along $\mathbf{L}$ — it precesses around it.
In the body frame, Euler's equations give a constant axial spin $\omega_3$ and a uniform circling of the transverse angular velocity at rate $\Omega=\frac{I_3-I_1}{I_1}\,\omega_3$ .
In the space frame, the symmetry axis and $\boldsymbol{\omega}$ precess uniformly about the fixed $\mathbf{L}$ at $\dot\phi=L/I_1$ , with the tilt angle $\theta$ rigorously constant — steady coning, no nodding.
The three vectors $\mathbf{L}$ , $\boldsymbol{\omega}$ , $\hat{\mathbf{e}}_3$ are coplanar at every instant; that plane rotates rigidly about $\mathbf{L}$ .
Geometrically the motion is a body cone (around $\hat{\mathbf{e}}_3$ ) rolling without slipping on a space cone (around $\mathbf{L}$ ), the contact line being $\boldsymbol{\omega}$ : for a prolate body ( $I_3<I_1$ , pencil-like) the cones touch externally, while for an oblate body ( $I_3>I_1$ , disk-like) the thin space cone sits inside the wide body cone.
A thin disk has $I_3=2I_1$ , so the visible wobble runs at about twice the spin — Feynman's cafeteria plate; the Earth's version is the 433-day Chandler wobble.
Key idea: Free precession is the geometric consequence of three vectors — the conserved $\mathbf{L}$ , the instantaneous $\boldsymbol{\omega}$ , and the symmetry axis $\hat{\mathbf{e}}_3$ — being coplanar while $\mathbf{L}$ stays fixed: that plane simply rotates rigidly about $\mathbf{L}$ , and the body cone rolls on the space cone with $\boldsymbol{\omega}$ as the line of contact.

Where this lives today

The Chandler wobble is now a precision observable. VLBI networks staring at quasars, GNSS constellations, and satellite laser ranging track the pole's position to better than a millimetre-per-day equivalent; the International Earth Rotation Service publishes the polar-motion series that every GPS receiver silently consumes. The century-old mystery of what keeps the wobble alive — a free mode should damp within decades — was settled only in 2000, when Richard Gross's angular-momentum budget showed ocean-bottom pressure fluctuations, with a minority share from winds, resonantly pumping it. The 305-versus-433-day gap, meanwhile, has become a tool: the measured period constrains mantle elasticity and the fluid core's decoupling, geophysics extracted from a discrepancy in this article's equation (8). In orbit, free precession is a daily engineering matter. Spin-stabilised spacecraft are launched coning — separation tip-off gives them an unwanted $\omega_\perp$ — and carry passive dampers (a fluid ring or a mass on a lossy spring) that dissipate kinetic energy at fixed $\mathbf{L}$ . For a major-axis spinner that drives the cone angle to zero; for a minor-axis (prolate) spinner the same physics opens the cone, the Explorer 1 lesson. Ground radars routinely fit torque-free-top light curves to tumbling rocket bodies and defunct satellites to recover their inertia ratios, and the same body-frame precession, quantised, shows up as the rotational band structure of deformed nuclei — the free symmetric top running from frisbees to femtometres.

The physical picture

Spin an American football slightly off its long axis and throw it. The one thing space guarantees is the direction of $\mathbf{L}$ — set at launch, conserved forever after. The football's pointed axis $\hat{\mathbf{e}}_3$ does not get that guarantee: it circles $\mathbf{L}$ once per precession period, and that circling is the spiral wobble every quarterback knows. No torque, no force — the wobble is what "constant $\mathbf{L}$ with anisotropic inertia" looks like. Why can't the axis just align with $\mathbf{L}$ and be done? Because $\mathbf{L}$ and $\boldsymbol{\omega}$ are different vectors for an anisotropic body. Decompose $\boldsymbol{\omega}$ into an axial part $\omega_3\hat{\mathbf{e}}_3$ and a transverse part $\boldsymbol{\omega}_\perp$ : the body responds with $\mathbf{L}=I_3\omega_3\hat{\mathbf{e}}_3+I_1\boldsymbol{\omega}_\perp$ — the two pieces are weighted differently, so $\mathbf{L}$ is tilted away from $\boldsymbol{\omega}$ unless the spin is purely axial or purely transverse. Rearranged, $\mathbf{L}=I_1\boldsymbol{\omega}+(I_3-I_1)\,\omega_3\hat{\mathbf{e}}_3$ : the angular momentum is a linear combination of $\boldsymbol{\omega}$ and $\hat{\mathbf{e}}_3$ alone, which is the coplanarity of all three vectors in one line of algebra. A fixed $\mathbf{L}$ then drags the whole plane — with $\boldsymbol{\omega}$ and $\hat{\mathbf{e}}_3$ frozen into it at fixed angles — around itself like a rigid fan:

Figure

The rolling cones — prolate tops touch the space cone externally, oblate tops swallow it

Poinsot's rolling-cone reading makes the kinematics tactile. Attach one cone rigidly to the body around $\hat{\mathbf{e}}_3$ (the body cone, half-angle $\alpha$ from axis to $\boldsymbol{\omega}$ ); hold a second cone fixed in space around $\mathbf{L}$ (the space cone, half-angle from $\mathbf{L}$ to $\boldsymbol{\omega}$ ). They touch along a single line — the instantaneous rotation axis $\boldsymbol{\omega}$ , whose points are momentarily at rest — so the body cone rolls without slipping on the space cone, carrying the symmetry axis around $\mathbf{L}$ as it goes. Which cone wraps which is decided by the shape. For a prolate body (mass hugging the axis: pencil, football, rifle bullet; $I_3<I_1$ ) the transverse spin is weighted more heavily in $\mathbf{L}$ , tilting $\mathbf{L}$ farther from the axis than $\boldsymbol{\omega}$ : the order from the axis is $\hat{\mathbf{e}}_3,\ \boldsymbol{\omega},\ \mathbf{L}$ , and the two cones touch side by side, externally. For an oblate body (mass flung outward: coin, frisbee, dinner plate, planet; $I_3>I_1$ ) the weighting reverses, $\mathbf{L}$ slips between $\hat{\mathbf{e}}_3$ and $\boldsymbol{\omega}$ , and the thin space cone sits wholly inside the wide body cone, which rolls around it from outside. Watch for the trap: several otherwise fine texts scramble this pairing; the half-angle relation $\tan\theta=(I_1/I_3)\tan\alpha$ derived below settles it in one line.

Figure

One rigid fan of three vectors, two precession clocks — the plane turns in space while omega circles in the body

One motion, two clocks. The space clock $\dot\phi=L/I_1$ times how fast the fan — and with it the visible axis — sweeps around $\mathbf{L}$ : this is the wobble you see. The body clock $\Omega=(I_3-I_1)\omega_3/I_1$ times how fast $\boldsymbol{\omega}$ circulates relative to the body — the rate a rider feels the tilt direction migrate around them. They are different numbers with different jobs, and much of this subject is keeping them straight.

The core equations

With zero external torque about the centre of mass, the lab hands us two locked quantities before any dynamics starts:

\boldsymbol{\tau}_{\text{ext}}=0\ \Rightarrow\ \mathbf{L}=\text{const},\qquad T=\tfrac12\,\boldsymbol{\omega}\cdot\mathbf{L}=\text{const} \qquad (1)

where: $\mathbf{L}$ — angular momentum about the CM, fixed as a vector in the space frame; $T$ — rotational kinetic energy. The fixed direction of $\mathbf{L}$ will serve as the reference axis of the whole problem — the one straight line nature draws for us. The dynamics comes from Euler's equations — derived in Euler's Equations of Motion and simply quoted here — specialised to the symmetric case $I_1=I_2$ , torque-free:

I_1\dot\omega_1=(I_1-I_3)\,\omega_2\omega_3,\qquad I_1\dot\omega_2=(I_3-I_1)\,\omega_1\omega_3,\qquad I_3\dot\omega_3=0 \qquad (2)

where: $\omega_{1,2}$ — transverse body-frame components of $\boldsymbol{\omega}$ ; $\omega_3$ — the component along the symmetry axis $\hat{\mathbf{e}}_3$ . The third equation is the gift of symmetry: its coupling coefficient was $(I_1-I_2)$ , which axisymmetry kills, so the axial spin $\omega_3$ is frozen and the remaining pair becomes linear with constant coefficients — the nonlinear rigid body reduced to a harmonic oscillator.

Derivation ABody-frame solution — the transverse spin circles at the rate Omega

We are solving equations (2) completely, to show that the transverse angular velocity keeps a constant magnitude and rotates uniformly around the symmetry axis at the body rate $\Omega=(I_3-I_1)\,\omega_3/I_1$ . The strategy is the standard one for cross-coupled linear pairs: bundle the two components into a single complex variable, which turns the pair into one first-order equation whose solution is a rotating phase. Watch the sign of $\Omega$ — it flips between oblate and prolate bodies, and that sign carries real physics (prograde versus retrograde circulation of the tilt).

Step 1. With $\omega_3$ constant (third equation), define $\Omega\equiv\dfrac{I_3-I_1}{I_1}\,\omega_3$ . The transverse pair becomes

\dot\omega_1=-\Omega\,\omega_2,\qquad \dot\omega_2=+\Omega\,\omega_1

Why the coefficients are constant: they contain only the frozen $\omega_3$ and the fixed moments — this is why the symmetric case escapes the elliptic functions of the general asymmetric top.

Step 2. Form $\eta\equiv\omega_1+i\omega_2$ . Multiplying the second equation by $i$ and adding,

\dot\eta=i\,\Omega\,\eta

Why the complex trick is natural: multiplication by $i$ is a 90-degree rotation in the transverse plane, and the coupled pair says exactly that each component feeds the other a quarter-cycle ahead.

Step 3. Integrate: $\eta(t)=\eta(0)\,e^{i\Omega t}$ , so

\omega_1=\omega_\perp\cos(\Omega t+\delta),\qquad \omega_2=\omega_\perp\sin(\Omega t+\delta),\qquad \omega_\perp=|\eta|=\text{const} \qquad (3)

where: $\omega_\perp$ — the constant magnitude of the transverse spin; $\delta$ — a phase set by initial conditions. In the body frame $\boldsymbol{\omega}$ therefore sweeps a cone of fixed half-angle $\alpha=\arctan(\omega_\perp/\omega_3)$ about $\hat{\mathbf{e}}_3$ — the body cone — at the uniform rate $\Omega$ .

Step 4. Read the sign. Oblate ( $I_3>I_1$ ): $\Omega>0$ , the tilt circulates in the same sense as the spin. Prolate ( $I_3<I_1$ ): $\Omega<0$ , retrograde circulation. Sphere ( $I_3=I_1$ ): $\Omega=0$ , no precession at all — $\mathbf{L}\parallel\boldsymbol{\omega}$ and the body just spins. Verified: for a thin disk $I_3=2I_1$ gives $\Omega=\omega_3$ ; the sphere limit kills the effect; magnitudes and signs agree with Goldstein §5.6 and Morin Ch. 9.

The body-frame answer feeds the space-frame one through the coplanarity identity, worth displaying on its own. Split $\boldsymbol{\omega}=\boldsymbol{\omega}_\perp+\omega_3\hat{\mathbf{e}}_3$ and weight the pieces with their moments:

\mathbf{L}=I_1\boldsymbol{\omega}_\perp+I_3\omega_3\hat{\mathbf{e}}_3 \;=\; I_1\,\boldsymbol{\omega}+(I_3-I_1)\,\omega_3\,\hat{\mathbf{e}}_3 \qquad (4)

where: the second form adds and subtracts $I_1\omega_3\hat{\mathbf{e}}_3$ . Equation (4) exhibits $\mathbf{L}$ as a combination of $\boldsymbol{\omega}$ and $\hat{\mathbf{e}}_3$ only: the three vectors are coplanar, always. Since $\mathbf{L}$ is fixed and the mutual angles are constant (magnitudes $\omega_\perp,\omega_3,L$ all conserved), the entire plane rotates rigidly about $\mathbf{L}$ — free precession in one sentence. How fast does the plane turn? That question deserves its own derivation, because it is where the Euler angles earn their keep.

Go deeper · Derivation B · 1 minSpace-frame rates — steady precession at L over I1 with theta rigorously constant

Derivation BSpace-frame rates — steady precession at L over I1 with theta rigorously constant

The goal is the lab-frame description: the symmetry axis holds a fixed tilt $\theta$ to $\mathbf{L}$ and precesses about it at the uniform rate $\dot\phi=L/I_1$ , while the body spins about its own axis at $\dot\psi$ . The strategy: point the space $Z$ -axis along the conserved $\mathbf{L}$ (the one free choice that costs nothing), express the body-frame spin components in Euler angles, and let the two conserved projections of $\mathbf{L}$ do the work. The subtle step is proving $\dot\theta=0$ — steady coning is derived, not assumed; contrast the heavy symmetric top, where a torque makes $\theta$ oscillate (nutation).

Step 1. With $Z\parallel\mathbf{L}$ , the standard Euler-angle kinematics (see Angular Velocity in Terms of Euler Angles) give the body components

\omega_1=\dot\phi\sin\theta\sin\psi+\dot\theta\cos\psi,\qquad \omega_2=\dot\phi\sin\theta\cos\psi-\dot\theta\sin\psi,\qquad \omega_3=\dot\phi\cos\theta+\dot\psi

Why this frame choice: every projection of $\mathbf{L}$ onto body axes becomes a pure trigonometric function of the tilt $\theta$ — the dynamics collapses into geometry.

Step 2. Project $\mathbf{L}$ onto the symmetry axis and the transverse plane:

L_3=I_3\omega_3=L\cos\theta,\qquad L_\perp=I_1\omega_\perp=L\sin\theta

Why both are conserved: $L$ is constant by (1), $\omega_3$ by (2), so $\cos\theta=I_3\omega_3/L$ is constant — $\dot\theta=0$ falls out immediately. The axis holds its tilt: free precession has no nutation.

Step 3. With $\dot\theta=0$ , the transverse components in Step 1 give $\omega_\perp=\dot\phi\sin\theta$ . Combine with $L\sin\theta=I_1\omega_\perp$ :

I_1\,\dot\phi\,\sin\theta=L\sin\theta\ \Rightarrow\ \boxed{\dot\phi=\frac{L}{I_1}} \qquad (5)

where: $\dot\phi$ — the space-frame precession rate: uniform, independent of the tilt, set only by the momentum magnitude and the transverse moment. This is the wobble rate a lab camera records.

Step 4. The spin about the symmetry axis follows from $\omega_3=\dot\phi\cos\theta+\dot\psi$ :

\dot\psi=\omega_3-\frac{L\cos\theta}{I_1}=\frac{L\cos\theta}{I_3}-\frac{L\cos\theta}{I_1}=L\cos\theta\,\frac{I_1-I_3}{I_1I_3} \qquad (6)

where: $\dot\psi$ — rotation of the body about $\hat{\mathbf{e}}_3$ relative to the precessing plane. Note $\dot\psi=-\Omega$ exactly (substitute $L\cos\theta=I_3\omega_3$ ): the body clock of the previous derivation is the same angle run backwards — the rate at which the coplanar fan migrates around the body equals minus the rate the body turns under the fan. Verified: $\dot\psi=-\Omega$ checks identically using $L\cos\theta=I_3\omega_3$ ; the precession period $2\pi I_1/L$ matches Landau–Lifshitz §33; oblate signs reproduce Goldstein eq. (5.49).

The cone geometry is now one trigonometric identity away. The body-cone half-angle $\alpha$ (from $\hat{\mathbf{e}}_3$ to $\boldsymbol{\omega}$ ) and the tilt $\theta$ (from $\hat{\mathbf{e}}_3$ to $\mathbf{L}$ ) compare as

\tan\alpha=\frac{\omega_\perp}{\omega_3},\qquad \tan\theta=\frac{L_\perp}{L_3}=\frac{I_1\omega_\perp}{I_3\omega_3}=\frac{I_1}{I_3}\,\tan\alpha \qquad (7)

where: the factor $I_1/I_3$ is the entire oblate/prolate story. Oblate ( $I_3>I_1$ ): $\theta<\alpha$ , so $\mathbf{L}$ lies between the axis and $\boldsymbol{\omega}$ — the space cone (half-angle $\alpha-\theta$ ) nests inside the body cone. Prolate ( $I_3<I_1$ ): $\theta>\alpha$ , $\boldsymbol{\omega}$ lies between the axis and $\mathbf{L}$ — the cones (space half-angle $\theta-\alpha$ ) touch externally. Both configurations are drawn in the hero figure, and the two rates complete the kit:

T^{\text{body}}_{\text{prec}}=\frac{2\pi}{|\Omega|}=\frac{2\pi I_1}{|I_3-I_1|\,\omega_3},\qquad T^{\text{space}}_{\text{prec}}=\frac{2\pi}{\dot\phi}=\frac{2\pi I_1}{L} \qquad (8)

where: the body period governs what an on-board gyroscope logs; the space period is the wobble a tracking telescope times. For the Earth the first is the Chandler timescale (months); for a frisbee the second is a twentieth of a second.

Which formula answers which question

you want	use	notes
the visible wobble rate of a thrown body	$\dot\phi=L/I_1$ , eq. (5)	what the lab camera sees
the on-body circulation of the tilt	$\Omega=\frac{I_3-I_1}{I_1}\omega_3$	sign: + oblate, − prolate
the constant tilt of the axis	$\cos\theta=I_3\omega_3/L$	no nutation — torque-free
which cone rolls where	$\tan\theta=\frac{I_1}{I_3}\tan\alpha$ , eq. (7)	oblate: space cone inside
spin relative to the precessing plane	$\dot\psi=-\Omega$ , eq. (6)	body and space clocks linked
wobble of a thin disk	$\dot\phi\approx2\omega_3$	Feynman's plate, $I_3=2I_1$

Limiting cases

Case	Result	Meaning
spherical top, $I_1=I_3$	$\Omega=0$ , $\mathbf{L}\parallel\boldsymbol{\omega}$	no precession; the degenerate polhode is a point
thin disk, $I_3=2I_1$	$\Omega=\omega_3$ , $\dot\phi\approx2\omega_3$ (small tilt)	the 2:1 cafeteria-plate wobble
slightly oblate, $I_3=I_1(1+\epsilon)$	$\Omega=\epsilon\,\omega_3\ll\omega_3$	slow body-frame wobble: Earth, $\epsilon\approx1/305$
prolate rod, $I_3\ll I_1$	$\Omega\to-\omega_3$ , external cones	retrograde tilt circulation: footballs, bullets
pure axial launch, $\omega_\perp=0$	$\alpha=\theta=0$ : steady spin	both cones collapse to the line $\mathbf{L}$
pure transverse launch, $\omega_3=0$	$\Omega=0$ : plain end-over-end tumble	a symmetric top thrown like a majorette's baton

Common trapThe two precession rates are not the same number — and engineering nutation is neither

Three near-synonyms, three different things. The body rate $\Omega=(I_3-I_1)\omega_3/I_1$ is how fast the transverse spin circles the symmetry axis for a rider on the body; the space rate $\dot\phi=L/I_1$ is how fast the axis itself circles $\mathbf{L}$ for the lab — for Earth the first takes 433 days while the second is a sliver over one sidereal day, so conflating them is not a small error. And neither is nutation in the strict sense: the free top's tilt $\theta$ is rigorously constant (Step 2 of the space-frame derivation), whereas true nutation is the oscillation of $\theta$ that appears when a torque acts — the heavy symmetric top's nodding. Spacecraft engineers routinely call free coning "nutation" (hence "nutation dampers"); read that vocabulary as this article's torque-free precession, not the heavy-top phenomenon.

Worked examples

1 · The wobbly frisbee (confidence builder)

A frisbee of mass $m=0.175\ \mathrm{kg}$ and radius $R=0.137\ \mathrm{m}$ leaves the hand spinning at $\omega_3=10\ \mathrm{rev/s}$ with its angular momentum tilted $\theta=5°$ from the symmetry axis. Find both precession rates, the wobble period, and the cone geometry.

Step 1. Thin-disk moments: $I_1=\tfrac14 mR^2=8.21\times10^{-4}\ \mathrm{kg\,m^2}$ (diameter axis), $I_3=\tfrac12 mR^2=1.64\times10^{-3}\ \mathrm{kg\,m^2}$ (symmetry axis) — the perpendicular-axis theorem's $I_3=2I_1$ in action. Oblate: $I_3>I_1$ .

Step 2. Body rate: $\Omega=\dfrac{I_3-I_1}{I_1}\,\omega_3=(2-1)\,\omega_3=\boxed{10\ \mathrm{rev/s}}$ — for a thin disk the tilt circulates the body exactly once per spin.

Step 3. Space rate: with the small tilt, $L\approx I_3\omega_3$ , so $\dot\phi=L/I_1\approx(I_3/I_1)\,\omega_3=2\omega_3=\boxed{20\ \mathrm{rev/s}}$ : the wobble period is $\boxed{0.05\ \mathrm{s}}$ . The wobble flickers at twice the spin — count flashes of a rim sticker against wobbles of the rim and you can measure $I_3/I_1$ from the bleachers.

Step 4. Cone geometry from (7): $\tan\alpha=(I_3/I_1)\tan\theta=2\tan5°$ gives $\alpha=9.9°$ . Since $\theta=5°<\alpha$ , $\mathbf{L}$ lies between axis and $\boldsymbol{\omega}$ : a space cone of half-angle $\alpha-\theta=4.9°$ nested inside a $9.9°$ body cone — the oblate panel of the hero figure with real numbers.

2 · The Chandler wobble — this article played on a planet (exam level)

The Earth is a slightly oblate free symmetric top with dynamical flattening $(I_3-I_1)/I_1=1/305.4$ . (a) Predict the rigid-body free-precession period. (b) The observed polar motion has the rotation pole circling the figure axis with amplitude about $6\ \mathrm{m}$ at the surface; find the tilt angle. (c) Find the half-angle of Earth's space cone and comment on which frame the wobble is visible in.

Step 1. The relevant clock for motion of the pole over the ground is the body rate: $\Omega=\omega_3/305.4$ with $\omega_3=2\pi/86164\ \mathrm{s}$ (sidereal day). Period $T=2\pi/\Omega=305.4$ sidereal days $\approx\boxed{305\ \mathrm{days}}$ — Euler's 1765 prediction of a ten-month free wobble.

Step 2. The observed period is about $433$ days — 40 % longer. The discrepancy is physics, not failure: the Earth is not rigid. Elastic yielding of the mantle and the fluid ocean lets the equatorial bulge partially follow the wandering rotation axis, weakening the restoring effect and stretching the period; Newcomb identified the mechanism within a year of Chandler's 1891 discovery. The rigid-Earth formula brackets reality from below — a textbook case of a model failing informatively.

Step 3. Tilt angle: $\alpha\approx6\ \mathrm{m}/6.371\times10^{6}\ \mathrm{m}=9.4\times10^{-7}\ \mathrm{rad}=\boxed{0.19\ \mathrm{arcsec}}$ — the rotation pole and the figure axis nearly coincide.

Step 4. Space-cone half-angle: $\theta-\alpha\approx\alpha\,\dfrac{I_3-I_1}{I_1}\cdot\dfrac{1}{1}\approx\dfrac{0.19''}{305}\approx6.4\times10^{-4}\ \mathrm{arcsec}$ — about $\boxed{2\ \mathrm{cm}}$ at the Earth's surface. In space, $\boldsymbol{\omega}$ hugs $\mathbf{L}$ to within a coin's width: the Chandler wobble is essentially invisible as a wobble of the axis in the sky and lives almost entirely in the body frame, as polar motion — the ground shifting under the pole. That is why it is measured by geodesy (where is the pole relative to the observatories?) rather than astronomy of the axis direction, and why the body clock $\Omega$ , not $\dot\phi$ , is the star of the geophysics.

Origins · 1 minSiméon Denis Poisson and how this was discovered

Siméon Denis Poisson

1781–1840 · École Polytechnique, Paris

Poisson was the great systematiser of French mechanics — Lagrange and Laplace's student, examiner, then successor at the École Polytechnique, author of the two-volume Traité de mécanique that trained a century of engineers. His 1809 memoir on the rotation of the Earth attacked exactly this article's problem: how a spinning near-symmetric body moves when left alone, and how its axes wander. The kinematic relations that evolve a space-fixed vector's components in the rotating body frame — the workhorse $\dot{\mathbf{A}}\big|_{\text{body}}=\mathbf{A}\times\boldsymbol{\omega}$ used by every attitude propagator — are still called Poisson's equations in the rigid-body literature. His name is scattered across the physics that grew from these methods: the Poisson bracket, born in his 1809 study of perturbed rotation constants, the Poisson equation of potential theory, the Poisson distribution of statistics. "Life is good for only two things," he liked to say, "doing mathematics and teaching it."

History

1758

— Euler solves the torque-free symmetric top, the first exact rigid-body motion beyond steady spin

1765

— Euler predicts a free wobble of the Earth's pole with a period near ten months

1809

— Poisson's memoir on the rotation of the Earth systematises the kinematics of free rotation

1834

— Poinsot recasts the motion as pure rolling geometry: polhode, herpolhode, body and space cones

1891

— Chandler detects the actual polar wobble at 427-433 days, not 305

1892

— Newcomb explains the stretched period by the Earth's elasticity and ocean response

Going further

References. Morin, Introduction to Classical Mechanics, Ch. 9 · Goldstein, Poole & Safko, Classical Mechanics 3rd ed., §5.6–5.7 · Landau & Lifshitz, Mechanics Vol. 1, §33 · Taylor, Classical Mechanics, Ch. 10

Provenance. Equations derived directly from Euler's equations specialized to $I_1=I_2$ (Goldstein §5.6–5.7; Landau-Lifshitz §33; Morin Ch. 9). Verifier should: (1) confirm $\Omega=(I_3-I_1)\omega_3/I_1$ and $\dot\phi=L/I_1$ ; (2) check the disk relation $I_3=2I_1$ via the perpendicular-axis theorem and the resulting 2:1 wobble; (3) check Earth's Euler period $\approx 305$ days from $(I_3-I_1)/I_1\approx 1/305$ and the observed Chandler period $\approx 433$ days. All numbers are standard textbook values; none fabricated. Additional worked numbers on this page (frisbee $\alpha=9.9°$ , space cone $4.9°$ , wobble period $0.05$ s; Earth tilt $0.19''$ , space cone $\approx2$ cm) are illustrative arithmetic recomputed from equations (5)–(8). Note: the cone-nesting assignment (prolate = external tangency, oblate = space cone inside body cone) follows directly from equation (7); a verifier should re-derive it from the half-angle relation, as it is frequently transposed in secondary sources.

Builds on: Euler's Equations of Motion · Principal Axes and Principal Moments of Inertia · Angular Momentum of a Rigid Body · Euler Angles · Classification of Rigid Bodies: Spherical, Symmetric, and Asymmetric Tops Leads to: Angular Velocity in Terms of Euler Angles · Inertia Ellipsoid · Rotational Kinetic Energy

If absolutely nothing torques the body, what physically drives the axis around the cone — where does the "force" come from?

Nothing drives it, and that is the point: uniform precession is the natural inertial motion of an anisotropic body, the rotational analogue of straight-line coasting. The feeling that an axis "should" stay put is imported from spheres, for which $\mathbf{L}=I\boldsymbol{\omega}$ locks the axis to the conserved $\mathbf{L}$ . An axisymmetric body responds to axial and transverse spin with different moments, so $\boldsymbol{\omega}$ and $\mathbf{L}$ point differently; the instantaneous rotation then perpetually carries the symmetry axis around the one direction that cannot move, $\mathbf{L}$ . No energy is exchanged, no work is done — every magnitude in the problem ( $\omega_3$ , $\omega_\perp$ , $\theta$ , $T$ , $L$ ) is constant. Only directions circulate.

Is this the same precession as a gyroscope leaning on a pivot, just without gravity?

No — they are cousins with opposite causal structures, and mixing them is the classic exam disaster. The gyroscope on a pivot (see Gyroscopic Precession) precesses because a gravitational torque continuously rotates its angular momentum: $\mathbf{L}$ itself moves, tracing a cone, at a rate $\tau/L\sin\theta$ set by the torque. The free top precesses with $\mathbf{L}$ nailed in place and zero torque; what circulates is the axis around the fixed $\mathbf{L}$ , at $L/I_1$ , set purely by inertia. Quick diagnostic for any problem: ask what $\mathbf{L}$ is doing. $\mathbf{L}$ moving → torque-driven precession; $\mathbf{L}$ fixed with the axis coning around it → free precession, this article.

My frisbee's wobble visibly dies out after a second or two while the spin keeps going — doesn't that contradict the constancy of the cone angles?

The constancy holds for a torque-free rigid body, and a frisbee in flight is neither. Aerodynamic pressure on a wobbling disk oscillates asymmetrically and exerts small torques that preferentially damp $\omega_\perp$ ; the flexible plastic also dissipates a little energy internally at fixed $\mathbf{L}$ , and for an oblate (major-axis) spinner both effects shrink the cone angle. The clean prediction survives where the perturbations vanish: in vacuum footage of spin-stabilised spacecraft the coning persists for hours exactly as equation (8) says, until an engineered damper — not the physics of this article — removes it. Your dying wobble is confirmation that a stable oblate spinner sheds its cone, which is precisely why frisbees fly well.

Feynman's story says the plate's spin was twice the wobble — the formula here gives wobble twice the spin. Which is right?

The formula. For a thin plate $I_3=2I_1$ , so the space-frame wobble rate is $\dot\phi=L/I_1\approx(I_3/I_1)\omega_3=2\omega_3$ : wobble twice the spin, for small tilt. In Surely You're Joking, Mr. Feynman! the anecdote indeed states the ratio the other way — "the medallion rotates twice as fast as the wobble" — and physicists have amused themselves about the inversion for decades (Benjamin Chao's 1989 note settled the arithmetic under the title "Feynman's dining hall dynamics"). The charitable reading: the book was dictated thirty years after the cafeteria. The physics reading: measure it — film any spun plate and count. The 2:1 the correct way is a two-line consequence of the perpendicular-axis theorem, and getting it from a flying plate is a lovely afternoon experiment.

How can omega change direction at all if angular momentum is conserved and nothing acts on the body?

Because $\boldsymbol{\omega}$ is not a conserved quantity — only $\mathbf{L}$ is. Conservation laws protect momenta, and the rotational momentum is $\mathbf{L}=\mathsf{I}\boldsymbol{\omega}$ , not $\boldsymbol{\omega}$ itself. Since the inertia tensor referred to space axes turns with the body, a constant $\mathbf{L}$ maps to a time-dependent $\boldsymbol{\omega}=\mathsf{I}^{-1}(t)\,\mathbf{L}$ . The energy is no help either: $T=\frac12\boldsymbol{\omega}\cdot\mathbf{L}$ only pins the projection of $\boldsymbol{\omega}$ onto the fixed $\mathbf{L}$ , leaving it free to swing around the space cone at fixed angle. The sphere is the exception that proves the rule: there $\mathsf{I}$ is a scalar, $\boldsymbol{\omega}$ inherits $\mathbf{L}$ 's constancy, and precession dies.

Does the free symmetric top nutate? Spacecraft engineers keep talking about nutation dampers for exactly this motion.

In the strict mechanics vocabulary, no: nutation is an oscillation of the tilt angle $\theta$ , and the free top's $\theta$ is rigorously constant — the space-frame derivation forces $\dot\theta=0$ from two conservation laws. The nodding motion properly called nutation belongs to the heavy symmetric top, where gravity's torque lets $\theta$ oscillate between turning points (see Heavy Symmetric Top and Nutation). Aerospace usage evolved separately: there, any coning of the spin axis — this article's free precession — is called "nutation," and the fluid-ring or mass-spring devices that damp it are "nutation dampers." Both communities are self-consistent; just translate at the border. On this site, precession means the steady sweep, nutation means the nod.

Why does the space-frame rate depend only on L and I1, not on I3 or on how fast the body spins about its axis?

Because the precession of the plane is carried entirely by the transverse dynamics. The rigid fan of $\mathbf{L},\boldsymbol{\omega},\hat{\mathbf{e}}_3$ turns about $\mathbf{L}$ exactly as fast as the transverse angular velocity carries the axis sideways, and the transverse response of the body is governed by $I_1$ alone: $L\sin\theta=I_1\omega_\perp=I_1\dot\phi\sin\theta$ , whence $\dot\phi=L/I_1$ with everything else cancelling. The axial moment $I_3$ and spin $\omega_3$ enter only through how much of $L$ is axial — they set the tilt $\theta$ and the body clock $\Omega$ , but once $L$ is fixed, the plane's sweep rate is blind to them. This is why wildly different bodies with equal $L$ and $I_1$ wobble in lockstep.

Is the polhode from the general rigid-body story consistent with the circles I get here?

Perfectly — the symmetric top is the polhode picture with a symmetry bonus. For a general asymmetric body the intersection of the energy ellipsoid and momentum sphere gives the polhode's non-circular loops and the separatrix through the intermediate axis (see Euler's Equations of Motion). Set $I_1=I_2$ and both surfaces become surfaces of revolution about $\hat{\mathbf{e}}_3$ : their intersections degenerate into perfect circles at constant latitude, which is exactly the body cone traced by $\boldsymbol{\omega}$ at fixed $\alpha$ . The separatrix and the instability vanish along with the intermediate axis — a symmetric top has no middle moment to be unstable about, which is why its motion is steady coning rather than Dzhanibekov flips.