Circular motion and inward acceleration

In brief

Swing a ball on a string in a steady circle and you feel the string pull taut the whole time, tugging the ball inward toward your hand. The ball's speed never changes, yet the string is doing something to it every instant, and that something is acceleration. This confounds a common assumption, that acceleration only means speeding up or slowing down. It does not: acceleration is any change in velocity, and velocity carries a direction as well as a size. A ball tracing a circle at unchanging speed has a velocity whose direction is changing every instant, and by the ordinary definition of acceleration, that is enough to require a force, one that always points toward the center of the circle, however steady the ball's speed appears.

The full treatment

First look: the taut string and the false push outward

Whirl a ball on a string overhead and let the string suddenly break or slip from your fingers. The ball does not fly outward, away from the center, as the phrase "centrifugal force" seems to promise. It flies off in a straight line, tangent to the circle at the instant of release, exactly the direction it happened to be moving at that moment. What the string was doing the whole time was not resisting an outward push, it was continuously bending the ball's straight-line tendency into a curve, and it takes a real, inward-directed force to do that bending, moment after moment.

Building the idea: velocity changing in direction alone

Consider the ball's velocity vector at two nearby instants, separated by a short time delta-t, as it moves along its circular path of radius r at constant speed v. Both velocity vectors have the same length, v, since the speed is not changing, but they point in slightly different directions, because the path has curved slightly between them. Draw these two velocity vectors starting from the same point; the small vector connecting their tips is the change in velocity, delta-v, over that interval. Because the two vectors have equal length and are separated only by a small angle, delta-v is very nearly perpendicular to them both, and since the velocity vectors themselves are tangent to the circle, a vector perpendicular to the tangent at that point on a circle points directly toward, or away from, the center. A short geometric argument, following how the ball actually moves rather than assuming an answer, shows it points toward the center: the ball is always curving in, not out.

The formal model: deriving a equals v squared over r

Make this precise. Over a short time delta-t, the ball sweeps out a small angle delta-theta at the center, related to the arc length it travels by delta-theta equals v times delta-t divided by r, since arc length equals speed multiplied by time, and angle equals arc length divided by radius. Because the velocity vector is always tangent to the circle, it rotates by that same small angle delta-theta over the same interval. For two vectors of equal length v separated by a small angle delta-theta, basic geometry (approximating the short arc connecting their tips by a straight chord) gives the length of the change vector as very nearly v times delta-theta. Substitute the expression for delta-theta found above: the size of delta-v is v times, v times delta-t divided by r, which is v squared times delta-t divided by r. Divide by delta-t to get the rate of change of velocity, the acceleration: a equals v squared divided by r. This is the centripetal, meaning center-seeking, acceleration, and the derivation shows it emerges purely from a changing direction, with the speed v held constant throughout. Using the angular rate of rotation, omega, defined as v divided by r, the same acceleration can also be written a equals omega squared times r.

What supplies the force, and why "centrifugal" misleads

Newton's second law says a net force is needed to produce any acceleration, so an object moving in a circle needs a net inward force of size mass times v squared over r, called the centripetal force. This is not a new, separate kind of force the way gravity or friction are; centripetal is a description of the direction a force must point and the job it must do, and the actual physical agent supplying it varies case by case: the string's tension for a whirled ball, gravity for an orbiting satellite, friction between tires and road for a car rounding a level curve, the normal force from a wall for a rider pressed against the side of a spinning carnival drum. The sensation of being "flung outward" in a turning car is not an outward force acting on the passenger at all; it is the passenger's own body continuing in a straight line by inertia while the car curves inward beneath them, an effect that only looks like an outward push if you insist on describing motion from inside the turning car itself, a non-inertial frame of reference in which a fictitious "centrifugal" term has to be invented to make the bookkeeping close.

Lineage

Christiaan Huygens derived the centripetal acceleration formula geometrically and published it in Horologium Oscillatorium in 1673, using an argument close to the one given here, several years before Newton's Principia appeared. Newton had also worked out the relationship privately in the 1660s and gave his own derivation, along with the crucial further step of connecting it to Kepler's laws of planetary motion, in Book I of the Principia in 1687, a connection that let him argue that whatever force holds planets in their orbits must weaken with the square of distance. The two derivations, arrived at independently by two of the founders of modern mechanics using different methods, is part of why the result was trusted and built upon so quickly, and it became the geometric hinge connecting orbital astronomy to terrestrial mechanics for the first time.

The strongest case for it

The centripetal acceleration formula sets the required force for anything moving along a curved path at known speed and radius, and from there it drives an enormous range of calculation: the banking angle needed on a highway curve or a railway bend so that a vehicle can round it without relying on friction alone, the minimum speed for a roller coaster car to maintain contact with the track through a vertical loop, the tension needed in a tether for a given orbit or swing, and the balance between gravity and orbital speed that keeps a satellite at a given altitude, a connection the next entry in this sequence develops directly. It also cleanly separates a genuine physical force, whatever supplies the inward pull, from an observer's frame-dependent sensation of being pushed outward, resolving a confusion that trips up nearly everyone encountering circular motion for the first time.

The strongest case against it

The formula a equals v squared over r describes only the acceleration due to changing direction, and it applies exactly as derived only when speed is constant, uniform circular motion. The moment speed also changes along the path, a pendulum swinging faster near the bottom of its arc, a planet moving faster near perihelion in an elliptical orbit, there is an additional tangential acceleration component, along the direction of motion, and the total acceleration is the vector sum of this tangential part and the centripetal part derived here, not either one alone. The formula also generalizes beyond literal circles, applying locally to any curved path if r is taken as the local radius of curvature at that point, but using a single, fixed r for a path that is not actually circular anywhere will give the wrong answer. The most persistent misconception is treating "centrifugal force" as a real, outward-acting force on the object in the ground frame; it is not, and including it alongside the real inward force in an inertial-frame calculation double-counts and reverses the physics.

Where it stands now

This derivation and its result are broad-consensus, unmodified since the seventeenth century, and remain in direct daily use in vehicle dynamics, orbital mechanics, and rotating machinery design. At speeds approaching that of light the description is absorbed without contradiction into general relativity's more complete treatment of curved motion, but the classical formula derived here remains exact for every ordinary application.

Test yourself

A car of given mass rounds a flat, unbanked curve of known radius at a given speed. Find the minimum coefficient of friction between the tires and the road needed to keep the car on the curve without sliding outward, and then find the speed at which a curve banked at a given angle would require no friction at all to negotiate safely. Finally, take a satellite orbiting at a known altitude and orbital speed, and using only the centripetal acceleration formula, determine the acceleration that whatever holds it in orbit must supply, then explain what additional fact about gravity, not yet established in this entry, would be needed to confirm that gravity alone is capable of supplying exactly that acceleration at that altitude.