Picture a child on a swing. Left alone after one big push, the swing arcs back and forth but each arc is a little smaller than the last — air resistance and friction in the chains steadily bleed away the motion until the swing hangs still. That is damping. Now stand behind the child and give a small shove once every time the swing comes back to you. If you time your shoves to the swing's own rhythm, those little pushes add up and up until the child is flying — you have driven the oscillator on resonance. Mistime them, pushing too fast or too slow, and you fight the swing as often as you help it; it barely moves.

That single playground scene contains the whole physics of this page. A mass on a spring, an electron in a radio's tuning circuit, a molecule soaking up light, a suspension bridge in a crosswind — every one of them is a damped, driven oscillator, and every one obeys the same equation. We will write that equation down, split its solution into a part that dies away and a part that lives forever, and discover why a tiny periodic push at just the right frequency can build a catastrophically large response.

Three forces, one equation

Take a block of mass m on a spring, free to slide along a line. Let x be its displacement from the resting position. Three forces act on it:

• the spring's restoring force, -k x — always pulling back towards x=0, with stiffness k;
• a damping (friction/drag) force, -b\dot{x} — opposing the velocity \dot{x}, with damping constant b;
• an external driving force, F_0\cos(\omega t) — a push that oscillates at our chosen drive frequency \omega.

Newton's second law, m\ddot{x} = \sum F, gathers them into one line. Move everything but the drive to the left and you have the equation this whole page is about:

It is a second-order, linear, non-homogeneous differential equation: second order because of the \ddot{x}, linear because x and its derivatives appear only to the first power, and non-homogeneous because of that F_0\cos(\omega t) sitting on the right-hand side refusing to be zero.

First, switch the drive off: free damped motion

Before we push, let us just let go. Set F_0 = 0 and the equation becomes homogeneous:

How the block returns to rest depends entirely on how the damping b compares with the critical value b_c = 2\sqrt{mk} = 2m\omega_0. There are three regimes:

• Underdamped (b < 2\sqrt{mk}): the block overshoots and oscillates, but inside a shrinking envelope. It rings like a struck bell fading away: x(t) = A\,e^{-\gamma t}\cos(\omega_d t + \phi), wobbling at a slightly reduced frequency \omega_d = \sqrt{\omega_0^2 - \gamma^2}.
• Critically damped (b = 2\sqrt{mk}): the fastest possible return to rest with no overshoot. This is the sweet spot engineers aim a car's shock absorbers at — you hit a bump and settle at once, without bouncing.
• Overdamped (b > 2\sqrt{mk}): so much drag that the block creeps sluggishly back to zero without ever crossing it — like a door on a stiff hydraulic closer. It never oscillates.

In every case, though, the key fact is this: the free motion decays. Whatever wobble the block started with fades under that e^{-\gamma t} envelope and, after a while, is gone. Hold on to that — it is the reason the driven problem eventually simplifies.

Transient + steady state

Now switch the drive back on. Because the equation is linear and non-homogeneous, its general solution is a sum of two pieces — exactly the recipe from solving a second-order non-homogeneous equation:

• x_h(t) is the homogeneous solution — the free damped motion we just met. It carries the memory of how the system started (the initial position and velocity). But it wears the e^{-\gamma t} envelope, so it dies out. We call it the transient.
• x_p(t) is a particular solution — a single steady oscillation that keeps in step with the driving force forever. We call it the steady state.

Wait long enough — a few decay times — and the transient has vanished. All that survives is the steady state. And here is the crucial, non-obvious fact:

You can check the amplitude formula by substituting x_p = A\cos(\omega t - \delta) back into the equation of motion; the algebra is a lovely exercise in matching \cos and \sin terms. What matters for us is the shape of that function A(\omega) — because that shape is resonance.

The resonance curve

Let us plot the steady-state amplitude A(\omega) against the drive frequency \omega. To keep the picture clean we set m = k = F_0 = 1, so the natural frequency sits at \omega_0 = \sqrt{k/m} = 1. Drag the slider to change the damping b and watch the peak.

Everything about resonance is in this one picture. When you drive near \omega_0, the two terms under the square root nearly cancel — the (k - m\omega^2) term vanishes exactly at \omega = \omega_0 — leaving only the small b\omega term in the denominator, so the amplitude soars. The lighter the damping, the taller and sharper the spike. This is why timing your pushes to the swing's own rhythm works so spectacularly, why a wine glass shatters at exactly its ringing note, and why soldiers break step crossing a bridge.

Worked examples

Example 1 — the natural frequency. A block of mass m = 0.5\ \text{kg} hangs on a spring of stiffness k = 200\ \text{N/m}. Its natural angular frequency is

To make it resonate you would drive it at about 20\ \text{rad/s} — roughly 3.2 pushes per second (since f = \omega_0/2\pi).

Example 2 — which damping regime? The same spring and mass are fitted with a damper of constant b = 15\ \text{kg/s}. Is the free motion underdamped, critical, or overdamped? Compare b with the critical value:

Since b = 15 < 20 = b_c, the system is underdamped: it will oscillate a few times inside a decaying envelope before settling. Swap in a stronger damper with b = 25\ \text{kg/s} > 20 and it would be overdamped, creeping back without a single wobble.

Example 3 — amplitude at resonance. Keep the underdamped case (m = 0.5, k = 200, b = 15) and drive it right at resonance, \omega \approx \omega_0 = 20\ \text{rad/s}, with a force amplitude F_0 = 6\ \text{N}. At \omega = \omega_0 the (k - m\omega^2) term is zero, so the amplitude formula collapses to

Halve the damping to b = 7.5 and the resonant amplitude doubles to 4\ \text{cm} — lighter damping, taller peak, exactly as the resonance curve promised.