Visual walkthrough — Gyroscope in spacecraft attitude control — preview

We build every symbol before using it. If you have never met an arrow-that-means-spin, start at Step 1.

---

Step 1 — What "angular momentum" looks like as an arrow

WHAT. A spinning wheel carries a hidden arrow. Its length says how hard it is to stop the spin; its direction says which way the wheel spins, decided by the right-hand rule: curl your right fingers along the spin, your thumb points along the arrow. We call this arrow the spin angular momentum and write it $\vec{L}$.

WHY an arrow and not just a number? Because the whole story is about direction changing, not speed changing. A plain number "how fast it spins" could never tell us the axis has tilted. We need something that carries a direction — a vector. (See Angular Momentum and Cross Product for the right-hand rule.)

PICTURE. Look at the red arrow — it stands straight up out of the spinning disc.

$L$ (no arrow) means the length of that arrow, a plain positive number: $L=I\omega$.

---

Step 2 — What a torque does: it hands the arrow a small nudge-arrow

WHAT. A torque $\vec{\tau }$ is a "twisting push." The one rule we assume — the rotational version of Newton's second law — is: $\vec{\tau }=\frac{d\vec{L}}{dt}$

WHY this and not something new? Just as an ordinary force is "rate of change of momentum" ($\vec{F}=d\vec{p}/dt$), a torque is the rate of change of the spin arrow (see Torque and Newton's Second Law for Rotation). This is the only law we need — everything else is geometry.

Rearranged for a tiny time-slice $dt$: $d\vec{L}=\vec{\tau }dt$

• $d\vec{L}$ ::: the tiny change in the spin arrow during $dt$ — itself a small arrow.
• $\vec{\tau }$ ::: the twisting push; its own direction (right-hand rule again) is where the nudge-arrow points.
• $dt$ ::: a sliver of time; making it small keeps the nudge small, so we can watch the arrow move step by step.

WHY read it this way? The equation says: the torque doesn't move the arrow instantly — it feeds it a small extra arrow $d\vec{L}$ every instant. The new arrow is the old one plus this nudge, tip-to-tail.

PICTURE. The black arrow is the old $\vec{L}$; the red arrow is the nudge $d\vec{L}$ that torque adds.

---

Step 3 — The one choice that changes everything: push perpendicular

WHAT. Gravity (or a control torque) tries to tip the axis. Such a torque is perpendicular to $\vec{L}$. Let us split any torque into two allowed jobs and see which one this is.

WHY split it? Because a vector can only do two things to another vector: change its length or change its direction.

• A nudge $d\vec{L}$ parallel to $\vec{L}$ lengthens or shortens it → the wheel spins faster/slower.
• A nudge $d\vec{L}$ perpendicular to $\vec{L}$ cannot change the length at all (a right-angle push adds nothing along the arrow) → it can only swing the direction.

The tipping torque is perpendicular. So it is physically incapable of tipping the arrow over — it can only rotate it sideways. That single fact is precession.

PICTURE. Two nudges compared: the parallel one (grey, stretches) versus the perpendicular one (red, swings). Only red survives here.

---

Step 4 — Turn the nudge into an angle: why we divide by $L$

WHAT. After time $dt$ the arrow has swung by a tiny angle $d\varphi$. How big?

Because the nudge $d\vec{L}$ is perpendicular, the tip of $\vec{L}$ travels a tiny straight step of length $|d\vec{L}|=\tau dt$ at distance $L$ from the pivot. For any arm swinging by a small angle: $\text{arc length}=\text{radius}\times \text{angle}$

WHY this relation, and why call $L$ the radius? The pivot is where the arrow starts (the tail). The tip sits a distance $L$ away — that is the swing radius. The sideways step $|d\vec{L}|$ is the arc. So: $d\varphi =\frac{\text{arc}}{\text{radius}}=\frac{|d\vec{L}|}{L}=\frac{\tau dt}{I\omega }$

• $|d\vec{L}|=\tau dt$ ::: length of the sideways step (from Step 2).
• $L=I\omega$ ::: length of the swinging arrow = the radius.
• $d\varphi$ ::: the tiny turn angle, in radians (radians are made for arc/radius — that's why we use them).

PICTURE. A thin wedge: radius $L$, arc $|d\vec{L}|$ (red), opening angle $d\varphi$.

---

Step 5 — The precession rate falls out

WHAT. Divide that turn angle by the time it took: $\Omega =\frac{d\varphi }{dt}=\frac{\tau dt/(I\omega )}{dt}=\frac{\tau }{I\omega }$

The $dt$ cancels — the axis sweeps around at a steady rate, independent of the sliver we chose.

WHY does bigger spin slow the precession? A longer arrow needs the same sideways step to count as a smaller angle (Step 4: big radius, small angle). So a fast gyro barely turns — it looks rigid. That is exactly the "stubborn reference" a spacecraft wants.

PICTURE. The full circle the tip of $\vec{L}$ traces — one steady sweep at rate $\Omega$.

The tidy direction-aware version (from Cross Product) is $\vec{\tau }=\vec{\Omega }\times \vec{L}$: the axis walks perpendicular to both the push and the spin.

---

Step 6 — Edge and degenerate cases (never get surprised)

WHAT. Test the formula at its extremes.

Case	What happens	Why
$\tau =0$	$\Omega =0$ — arrow frozen	no nudge, no turn: pure rigidity
torque parallel to $\vec{L}$	no precession; $L$ grows/shrinks	Step 3: parallel changes length, not direction
$\omega \to 0$ (spin dies)	$\Omega =\tau /(I\omega )\to \infty$	denominator vanishes — the "walk" becomes a fall; a slow top topples
$\omega \to \infty$	$\Omega \to 0$	infinitely stubborn: axis truly fixed

WHY the $\omega \to 0$ case matters most. Our whole picture assumed the spin arrow dwarfs everything, so the nudge only rotates it. When spin fades, that assumption breaks: the nudge is no longer tiny compared to $L$, the "swing" outruns the geometry, and the top wobbles then falls. This is why real gyros must spin fast. (Compare Precession of a Spinning Top.)

PICTURE. Two towers of the same push: a fast arrow barely tilts; a slow arrow flops over.

---

The one-picture summary

Everything above is one loop: torque feeds a perpendicular nudge → nudge turns the arrow by $d\varphi =\tau dt/L$ → repeat → the axis walks a circle at $\Omega =\tau /(I\omega )$.

Recall Feynman retelling — the whole walkthrough in plain words

A spinning wheel carries a secret arrow standing along its axle; how long the arrow is tells how badly the wheel wants to keep spinning. Now I give the axle a sideways twist. Newton's rule says a twist doesn't shove the arrow instantly — it just glues a tiny extra arrow onto the tip every moment. Here's the magic: I twisted sideways, so the tiny arrow points sideways too, and a sideways add-on can't make the main arrow longer or shorter — it can only swing its tip a little to the side. Do that over and over and the tip walks in a circle: the axle refuses to fall and instead slowly turns. How fast it walks is the sideways step divided by the arrow's length, per second — that's $\Omega =\tau /(I\omega )$. Spin the wheel harder, the arrow gets longer, the same sideways step counts as a tinier turn, so it walks slower and looks rock-steady. Let the spin die and the arrow shrinks to nothing — now the little sideways add-on is huge compared to it, the "walk" becomes a "fall," and the top topples. That single seesaw between the sideways nudge and the arrow's length is the entire trick a spacecraft uses to keep its bearings.

Recall Predict-first checks

Push the axis — which way does it move? ::: 90° sideways (precession), not the way you pushed. Torque parallel to $\vec{L}$ — precession? ::: No; it only speeds up or slows the spin. Why does a faster gyro look rigid? ::: Longer $L$ makes each sideways step a smaller angle, so $\Omega =\tau /(I\omega )$ shrinks. What happens as spin dies ($\omega \to 0$)? ::: $\Omega \to \infty$ — the walk becomes a fall; the top topples.

Related: Conservation of Angular Momentum (how the actuating side stores this same arrow).