Rotation, torque, and angular momentum

A bicycle you are sitting on falls over the instant it stops, yet the same bike stays up on its own once it is rolling. A skater with arms flung wide drifts around slowly, then snaps into a blur the moment the arms come in. And the phone in your pocket somehow knows the instant you turn it sideways. Three everyday mysteries, one hidden set of rules: the physics of spinning.

You already have a feel for straight-line motion from the earlier lessons: a push changes how fast something moves, heavier things are harder to get going, and a coasting object keeps its motion until something interferes. Spinning turns out to obey the very same story, told with a twist. Everywhere the straight-line world said push, the spinning world says twist. Everywhere it said heft, spinning asks not just how much mass but how far out that mass sits. Get that one substitution and rotation stops being a separate subject.

Play with it before any equation

The widget below is that skater seen from above, plus a second tab with a wrench on a stuck bolt. Start with the skater. Drag the arm-radius slider from arms-out to tucked-in and watch the spin rate. Notice the top bar, labelled L, never moves: the simulator holds one quantity dead constant while everything else responds. Your job is to name what changed and what did not.

Pulling the arms in made her spin faster, sometimes several times faster, and no one added any twist. So something was conserved, kept fixed, while the spin rate climbed. That conserved thing is the star of this lesson. Before we name it, we need the rotational vocabulary, built one word at a time from the straight-line words you already own.

Every straight-line idea has a spinning twin

Start with motion itself. A car on a road has a position, a speed, and an acceleration (the rate its speed changes). A spinning wheel has the exact same three, measured as angles instead of distances. Its angular velocity is how many degrees or radians it turns per second, and its angular acceleration is how fast that spin rate itself is changing. Same graphs as a moving car, with angle on the vertical axis.

Now the cause of motion. In a straight line, a force changes motion, and the rule is that acceleration equals force divided by mass. Spinning has a matching rule, but first we have to see why a plain force is not the whole story for rotation.

Torque: a push only counts if it has leverage

Switch the widget to the wrench tab. A bolt is stuck. You push on the wrench, and the same push does far more when your hand is out at the end of the handle than when it is right next to the bolt. Everyone has felt this: to loosen a tight bolt you grab a longer wrench or slide your hand outward. The turning effect of a force is not the force alone, it is the force times how far out it acts, along a line at right angles to the twist.

That turning effect is called torque. Its size is the force multiplied by the lever arm, the perpendicular distance from the pivot to where the force acts.

This is why the wrench in the widget breaks the bolt free once the grip is far enough out or the push is hard enough: only the product has to clear the threshold. A gentle push with a long lever beats a hard shove with no lever. A door handle sits far from the hinges for exactly this reason; push near the hinge and the door barely moves.

Moment of inertia: heft depends on where the mass is

In a straight line, mass is the single number that says how hard something is to get moving. Spinning needs a richer number, because the same lump of metal is easy to spin when it hugs the axis and stubborn when it sits far out. The rotational stand-in for mass is the moment of inertia, written $I$. What matters is not just how much mass but how far that mass sits from the axis, and the distance counts squared.

The $r^2$ is the whole secret behind the skater. Her arms barely change mass when she pulls them in, but they move much closer to the axis, and because distance enters squared, her moment of inertia drops sharply. That is the I bar you watched fall in the widget.

Angular momentum: the thing that was conserved

Now we can name the quantity the simulator held fixed. In a straight line, motion is carried by momentum, mass times velocity, and it is conserved when nothing pushes from outside (the currency crashes conserve). Spinning carries its motion in angular momentum, moment of inertia times angular velocity.

Here is the payoff, and it is a derived result, not something to take on faith. With no outside torque, $L$ cannot change. So if the skater shrinks her moment of inertia $I$ by pulling her arms in, and the product $I\omega$ has to stay the same, then her spin rate $\omega$ has no choice but to rise to compensate:

That is the answer to the predict probe: she spins faster, and no mystery force did it. She simply rearranged her own mass, and conservation did the rest. The very same rule explains a diver tucking to flip faster and a cat twisting to land on its feet.

Why the moving bike stays up, honestly

It is tempting to say the spinning wheels act like gyroscopes and hold the bike upright by their angular momentum. That sounds satisfying, and it is mostly wrong. Researchers have built bikes that cancel the gyroscopic effect entirely and they still balance themselves while rolling.^[3] The real reason is steering geometry: when a moving bike starts to tip, the front wheel automatically steers into the fall, and that curves the bike back under its own center of mass. It is a self-correcting feedback loop that only works while there is forward speed to steer with, which is why a stationary bike simply falls. Angular momentum does play a supporting role, but the honest headline is steering, not gyroscopes.

The CS bridge: the gyroscope in your pocket

Your phone contains a MEMS gyroscope, a chip with a tiny vibrating structure that senses rotation, and it reports angular velocity $\omega$ hundreds of times a second. To know which way the phone is actually facing, the software has to add up all those tiny turns over time, which is numerical integration, the same summing-up-small-steps trick from simulating motion. And it has the same weakness: every reading carries a little error, and integrating error makes the estimated heading slowly drift. That is why phones fuse the gyroscope with the accelerometer and magnetometer to pull the drift back, the same sensor-fusion idea that keeps a drone level.

heading = 0.0            # radians, where the device points
dt = 1 / 200            # gyro reports 200 times a second

for omega in gyro_stream():      # omega = angular velocity, rad/s
    heading += omega * dt        # integrate: add each small turn
    # every omega carries a tiny bias, and summing bias
    # makes 'heading' creep away from truth over minutes.
    # real devices fuse in the accelerometer to correct it.

Check yourself