Gyroscope — steady precession derivation: Question bank

Definition The symbols and coordinate frame used on this page

Set up a fixed frame at the pivot $O$ : $\overset{z}{^}$ points straight up (opposite gravity), and $\overset{x}{^}, \overset{y}{^}$ span the horizontal plane, forming a right-handed set ( $\overset{x}{^} \times \overset{y}{^} = \overset{z}{^}$ ).

$ℓ$ = ==distance from the pivot $O$ to the center of mass== (CM), measured along the spin axis. It is the length of the "arm" on which gravity acts. If the CM sat exactly at $O$ , then $ℓ = 0$ and there would be no gravity torque at all.
$θ$ = tilt (nutation) angle = the angle between the spin axis and the upward vertical $\overset{z}{^}$ . $θ = 0$ means the top stands straight up; $θ = 9 0^{\circ}$ means its axis is horizontal.
$ϕ$ = precession (azimuthal) angle = the compass-bearing of the axis in the horizontal plane, i.e. how far the axis has swept around $\overset{z}{^}$ . Its rate of change is the precession rate $Ω \equiv \dot{ϕ}$ (how fast the axis walks around).
$L$ = angular momentum about $O$ (see Angular Momentum). Its dominant part lies along the spin axis.
$I_{3}$ = axial moment of inertia = resistance to spinning about the symmetry axis itself. $I_{1} = I_{2}$ = transverse moment of inertia = resistance to being tipped sideways, i.e. rotation about any axis perpendicular to the symmetry axis through $O$ (see Moment of Inertia Tensor). For a symmetric top these two transverse values are equal, hence the single symbol $I_{1}$ .
$L_{s} = I_{3} ω_{s}$ = the magnitude of the spin angular momentum along the symmetry axis, where $ω_{s}$ is the spin rate.
$τ = r \times m g$ = gravity's torque about $O$ ; with the axis tilted in the $x$ – $z$ plane, $τ$ points along $\pm \overset{y}{^}$ (horizontal), which is why $L$ moves sideways.

The geometry behind the balance. Everything on this page rests on one picture. Look at Figure s01: the axis is tilted by $θ$ , so the CM sits at horizontal distance $ℓ sin θ$ from the vertical line through $O$ — that horizontal offset is the lever arm of gravity, giving torque magnitude $m g ℓ sin θ$ . Meanwhile the spin angular momentum $L_{s}$ , tilted by the same $θ$ , has horizontal projection $L_{s} sin θ$ — and only that horizontal part gets swept around $\overset{z}{^}$ . Figure s02 puts the two $sin θ$ 's side by side so you can see why they cancel.

Figure — Gyroscope — steady precession derivation

Figure s01 — the fixed frame at $O$ : the tilted spin axis carries $L$ (lavender), gravity $m g$ (coral) acts at the CM a distance $ℓ$ out, its horizontal lever arm $ℓ sin θ$ is marked, the resulting torque $τ$ (mint) is horizontal along $- \overset{y}{^}$ , and the axis sweeps the dashed butter-yellow precession circle at rate $Ω$ .

Figure s02 — the two right triangles side by side. Left: gravity's arm projects to $ℓ sin θ$ . Right: the spin angular momentum projects to $L_{s} sin θ$ . The torque balance $m g ℓ sin θ = Ω L_{s} sin θ$ has the identical $sin θ$ on both sides, so it cancels — leaving $Ω = m g ℓ / L_{s}$ , independent of $θ$ .

These questions assume you know the steady-precession result $Ω \approx \frac{m g ℓ}{I _{3} ω _{s}}, L_{s} = I_{3} ω_{s}$ and the master law $τ = \frac{d L}{d t}$ . If either feels shaky, revisit the parent before testing yourself.

True or false — justify

The precession rate

Ω

increases as you tilt the gyroscope further from vertical.

False (in the fast-top limit). The torque grows as

m g ℓ sin θ

, but the horizontal angular momentum being swept grows as

L_{s} sin θ

too — the

sin θ

cancels (Figure s02), leaving

Ω

angle-independent.

A gyroscope spinning twice as fast precesses twice as fast.

False.

Ω \propto 1/ ω_{s}

, so double the spin halves the precession rate. Fast spinners are more stubborn and walk around more slowly.

Gravity does positive work on the gyroscope during steady precession.

False. With

θ

constant the center of mass stays at fixed height, so gravity does zero net work — the motion is horizontal sweeping, not falling.

In steady precession the angular momentum vector

L

has constant magnitude.

True. Its tip traces a horizontal circle at constant radius; only its direction changes, driven by a torque perpendicular to it, exactly like circular motion of a position vector.

The torque from gravity points straight down.

False.

τ = r \times m g

is a cross product of two vectors in the

x

–

z

vertical plane, so it points along

\pm \overset{y}{^}

(horizontal) — that horizontality is precisely what forces sideways precession.

If you remove gravity entirely, a spinning symmetric top keeps its axis fixed in space.

True. With no torque,

d L / d t = 0

, so

L

(and hence the spin axis) is conserved — the top just spins in place with a frozen axis.

The exact Lagrangian treatment always gives two valid precession rates.

False. Two real roots of the quadratic

I_{1} Ω^{2} cos θ - I_{3} ω_{s} Ω + m g ℓ = 0

exist only when the discriminant

(I_{3} ω_{s})^{2} - 4 I_{1} m g ℓ cos θ \geq 0

; below that spin, no steady precession is possible at that angle.

For a top standing exactly vertical (

θ = 0

), the steady-precession formula predicts an infinite family of allowed

Ω

True-ish / degenerate. The torque vanishes (

sin 0 = 0

), so any

Ω

trivially satisfies the balance;

θ = 0

is the "sleeping top" special case, not ordinary precession.

The sense of precession does not depend on which way the top spins.

False. Reversing the spin flips the sign of

L_{s}

, and gravity's fixed torque then rotates

L

the opposite way, so

Ω

reverses sign — precession direction is locked to spin direction.

Spot the error

L = I_{1} ω_{s}

along the axis, so use the transverse moment of inertia."

Wrong inertia. The dominant spin angular momentum is about the symmetry axis, so it is

I_{3} ω_{s}

I_{1}

only governs the transverse and precession contributions in the exact theory.

"Since gravity pulls down,

d L

must point down, so the axis tips lower each instant."

Wrong direction.

d L = τ d t

and

τ

is horizontal (along

\pm \overset{y}{^}

), so

L

shifts sideways. The axis sweeps around, it does not drop.

"The

sin θ

cancels, therefore tilt angle plays no role anywhere in the theory."

Overreach. It cancels only in the fast-top approximation. The exact quadratic

I_{1} Ω^{2} cos θ - I_{3} ω_{s} Ω + m g ℓ = 0

keeps a

cos θ

through the

I_{1}

term, so the two roots do depend weakly on angle.

"The pivot reaction force contributes a big torque that we forgot."

Wrong about the moment arm. Torque is taken about the pivot $O$ , and the reaction acts at

O

with zero lever arm, so its moment vanishes — only gravity's torque survives.

Ω_{+}

(the plus root) is the physical answer because it's larger and tops precess fast."

Mislabeled. The

-

root is the one that reduces to

m g ℓ / (I_{3} ω_{s})

in the fast-spin limit; it is the ordinary slow precession.

Ω_{+}

is the fast precession branch, a distinct genuine solution.

"Nutation and precession are the same nodding motion."

Conflated. Precession is the steady horizontal sweep of the axis about vertical (

\dot{ϕ}

); nutation is the up-and-down bobbing of the tilt angle (

\dot{θ} \neq = 0

). Steady precession has

\dot{θ} = 0

, i.e. no nutation. See Nutation.

Why questions

Why does the

sin θ

cancel between torque and

d L / d t

Because both the lever arm of gravity (

ℓ sin θ

) and the horizontal part of the spin angular momentum (

L_{s} sin θ

) are the horizontal projections of the same tilted axis — the same geometric factor (Figure s02) appears on both sides.

Why does a faster-spinning top precess more slowly?

A larger

L_{s} = I_{3} ω_{s}

means the fixed horizontal torque produces a smaller fractional rotation of

L

per unit time; the same sideways push turns a bigger vector through a smaller angle.

Why is precession a rotation about the vertical rather than about the horizontal torque direction?

The torque nudges

L

's tip horizontally and perpendicular to the axis; integrating these perpendicular nudges makes the tip chase its own tail in a horizontal circle centered on the vertical, so

\overset{z}{^}

is the axis of the resulting circular sweep.

Why must the spin be about the symmetry axis for the simple formula to hold?

The clean cancellation and the identification

L_{s} = I_{3} ω_{s}

rely on angular momentum being aligned with the symmetry axis; off-axis spin mixes in

I_{1}

contributions and breaks the tidy fast-top result.

Why does a dying (slowing) top precess faster and eventually fall?

As friction lowers

ω_{s}

Ω \propto 1/ ω_{s}

climbs and the discriminant

(I_{3} ω_{s})^{2} - 4 I_{1} m g ℓ cos θ

shrinks; once it goes negative, steady precession no longer exists and nutation/toppling takes over.

Why doesn't the gyroscope simply fall like a non-spinning object?

Gravity changes angular momentum, not the center-of-mass position directly; with a large horizontal

L

present, the horizontal torque reorients

L

sideways instead of building downward angular momentum.

Edge cases

What happens to

Ω

ω_{s} \to \infty

Ω = m g ℓ / (I_{3} ω_{s}) \to 0

: an infinitely fast top barely precesses, its axis nearly frozen — the ideal "stubborn gyroscope" limit.

What happens at

θ = 9 0^{\circ}

(axis horizontal)?

The fast-top formula still gives

Ω = m g ℓ / (I_{3} ω_{s})

since

sin θ

cancels; the torque is maximal but so is the swept horizontal

L

, so precession proceeds at the same rate.

What happens for an over-tilted top,

θ > 9 0^{\circ}

(axis dips below horizontal, so

cos θ < 0

In the exact quadratic the discriminant

(I_{3} ω_{s})^{2} - 4 I_{1} m g ℓ cos θ

becomes

(I_{3} ω_{s})^{2} + 4 I_{1} m g ℓ ∣ cos θ ∣

, which is always positive — so real steady precession exists at any spin. Physically the CM is now below

O

, gravity restores rather than topples, and both roots stay real.

What is the minimum spin for steady precession at a given

θ

Set the discriminant to zero:

ω_{s, m i n} = 4 I_{1} m g ℓ cos θ / I_{3}

, valid only for

θ < 9 0^{\circ}

. Below this spin the roots are complex and only nutating or falling motion is possible.

Does the minimum spin depend on tilt in the exact theory?

Yes — through

cos θ

. Near vertical (

θ \to 0

cos θ \to 1

) the required spin is largest; as

θ \to 9 0^{\circ}

(

cos θ \to 0

) the threshold drops to zero, and past

9 0^{\circ}

there is no threshold at all.

What happens if

I_{1} \to 0

(a "needle-thin" top)?

The quadratic's

I_{1} Ω^{2} cos θ

term vanishes and the equation linearizes to

I_{3} ω_{s} Ω = m g ℓ

, recovering the exact fast-top formula with no approximation needed.

At exactly

θ = 9 0^{\circ}

in the exact quadratic, what special thing happens?

cos θ = 0

kills the quadratic term, so

Ω = m g ℓ / (I_{3} ω_{s})

becomes the single exact root — the two-root structure collapses to one at the horizontal orientation.

Recall One-line summary of every trap

Precession is $L$ being reoriented (not moved down) by a horizontal torque along $\pm \overset{y}{^}$ ; $Ω = m g ℓ / (I_{3} ω_{s})$ is angle-independent only because $sin θ$ cancels (Figure s02), uses $I_{3}$ not $I_{1}$ , slows with spin, has a sign tied to the spin sense, and exists only above a spin threshold (which disappears once $θ > 9 0^{\circ}$ ).

See also: Euler Angles · Angular Momentum · Lagrangian Mechanics · Moment of Inertia Tensor · Nutation · Symmetric Top.