The trouble with the lab frame
Rotation obeys Newton's rotational law: the rate of change of angular momentum equals the torque. In an inertial (lab) frame that reads simply as follows.
The rotational Newton's law — exact, but awkward, because in the lab frame the inertia tensor keeps changing as the body reorients.
The awkwardness is that \mathbf{L}=\mathbf{I}\,\boldsymbol{\omega}, but as the body turns, the components of \mathbf{I} in the lab frame keep changing. The cure is to sit in the body-fixed frame — the frame spinning with the object. There the inertia tensor is constant (just the three principal moments), at the price of working in a rotating, non-inertial frame.
The transport theorem and Euler's equations
Any vector seen from a rotating frame has an extra piece in its time derivative: the transport theorem says (d/dt)_{\text{lab}} = (d/dt)_{\text{body}} + \boldsymbol{\omega}\times. Apply it to \mathbf{L}:
The same physical law carried into the rotating body frame; the extra \boldsymbol{\omega}\times\mathbf{L} term is the price of leaving an inertial frame.
Now choose the body's principal axes, so \mathbf{L}=(I_1\omega_1,\,I_2\omega_2,\,I_3\omega_3) with the three moments constant. Writing the equation out component by component gives the celebrated set:
Euler's equations of rigid-body motion — three coupled, nonlinear equations for the body-frame angular velocity.
The torque-free top: wobble without a push
Set \boldsymbol{\tau}=0 — a body spinning freely in space, or a planet with no external couple. Astonishingly, the axis can still move. For a symmetric top (I_1=I_2\neq I_3), Euler's equations show \omega_3 stays constant while \omega_1 and \omega_2 rotate steadily: the spin axis traces a cone in the body frame. This is torque-free precession, and its rate is
The rate at which the spin axis circles in the body frame of a freely spinning symmetric top — zero only if the body is spherically symmetric (I_3=I_1).
The Earth is a real example. It is slightly oblate (I_3 a touch larger than I_1), so this formula predicts a free wobble of its rotation axis with a period of about 305 days — the 'Euler period'. The wobble is genuinely observed, but with a period of roughly 433 days: the Chandler wobble. The discrepancy is honest physics, not error — the Earth is not perfectly rigid, and its elastic yielding lengthens the period.
The heavy gyroscope: precession under gravity
Now hold the spinning top by a point below its centre, so gravity applies a torque. For a fast top the angular momentum is nearly all along the spin axis, \mathbf{L}\approx I_3\omega_3\,\hat{\mathbf{n}}. Gravity's torque \boldsymbol{\tau}=\mathbf{r}\times M\mathbf{g} points horizontally, perpendicular to \mathbf{L}. A perpendicular torque cannot change the length of \mathbf{L}, only swing its direction — so the axis sweeps sideways around a cone instead of falling.
The steady gyroscopic precession rate — remarkably, it does not depend on the tilt angle \theta, and it slows down as the spin \omega_3 increases.
Real tops also nod up and down — a fast, small tremor called nutation — superimposed on the smooth precession. Friction usually damps the nod away within moments, leaving the steady sweep the formula describes.
Worked example: how fast does a bicycle wheel precess?
- Set up an illustrative case. Suppose a bicycle wheel of mass M=2\ \text{kg} and radius r=0.33\ \text{m} is spun at $10$ revolutions per second and hung from one end of its axle, a distance \ell=0.10\ \text{m} from the wheel's centre.
- Model it as a hoop. Nearly all the mass sits at the rim, so the spin moment is I_3\approx Mr^2 = 2\times0.33^2 \approx 0.22\ \text{kg·m}^2.
- Spin angular momentum. \omega_3 = 2\pi\times10 \approx 63\ \text{rad/s}, so L = I_3\omega_3 \approx 0.22\times63 \approx 14\ \text{kg·m}^2/\text{s}.
- Gravitational torque about the support. \tau = Mg\ell = 2\times9.8\times0.10 \approx 2.0\ \text{N·m}.
- Precession rate. \Omega = \tau/L \approx 2.0/14 \approx 0.14\ \text{rad/s} — about one slow revolution every 45 seconds. Spin the wheel faster and it precesses slower, which is why a rapidly spinning gyroscope seems almost frozen in place.
The specific numbers are round assumptions, but the scaling \Omega\propto 1/\omega_3 is exact and is the whole principle behind gyrocompasses, camera stabilizers and the reaction wheels that point spacecraft.