物理 · Physics

Oscillations & Waves

Day 10 · 2026 · Phase C Waves, Light & Electromagnetism
A guitar string, a bridge, an electron inside an atom — all obey the same grammar, because almost anything pushed off balance rings back the same way.
From here physics changes its face. So far we asked how particles move; the next long stretch asks how disturbances travel — sound, light, electromagnetism, even quantum probability, all of it waves. And the seed of a wave is the humblest thing: vibration. Why does nearly every oscillation in the universe come out as the same sine curve? Why can a gentle push at just the right rhythm shake a bridge apart? And why, once you trap a wave inside a string, does it insist on existing only in whole-number patterns — that last point is the physical embryo of why atoms have discrete energy levels. Four cards: why SHM is everywhere, resonance, waves as travelling vibration, standing waves & superposition.

Simple Harmonic Motion: Why It's Everywhere 简谐运动

Classical mechanics · linear restoring force
Intuition Nudge anything sitting quietly at a stable equilibrium — a pendulum, a mass on a spring, a marble in a bowl — and it swings back, overshoots, swings back again. The remarkable part: they all trace nearly the same curve, a sine. No coincidence. Zoom into the bottom of any energy valley and it looks like a parabola; a parabolic potential gives a restoring force exactly proportional to the displacement — and that dictates simple harmonic motion.
Mechanism The core is a linear restoring force: the farther you pull, the harder it tugs back, in strict proportion.
F = −kx
F is the restoring force, x is the displacement from equilibrium, k is the "stiffness" (how hard it pulls back). That leading minus sign is the soul: the force always points back toward equilibrium — pull right, it pulls left; pull left, it pushes right.
This force law forces a pure sine:
x(t) = A cos(ωt + φ)
A is the amplitude (how far it swings), t is time, φ ("phi") is the initial phase (where it started). The key player is ω ("omega") — the angular frequency, how fast it oscillates — set purely by the system itself: ω = √(k/m), with m the mass. Stiffer and lighter means faster.
The most counterintuitive line: frequency is independent of amplitude. Pull the mass twice as far and one round trip takes essentially the same time — which is exactly what lets a pendulum clock keep time.
U x position equilibrium true potential parabolic fit
Zoom into the bottom of any well and it ≈ a parabola (amber dashed) — so small oscillations are almost always simple harmonic. Only at large amplitude, climbing out of the valley, does the true potential (blue) diverge and SHM break down.
The counterintuitive point SHM is everywhere not because nature loves sines, but as a mathematical inevitability: the Taylor expansion of any smooth potential near a minimum has, as its lowest nontrivial term, a quadratic (parabola). As long as the oscillation is small, higher terms are negligible and what's left is simple harmonic. "Small oscillations are harmonic" is the universal approximation you get by linearizing a complex system — which is why it's the starting point of all wave physics.
Cross-disciplinary · Math / Engineering "Near equilibrium, treat a nonlinear system as linear" is a craft that spans fields —
  • Control engineering: controllers for aircraft and robot arms are mostly designed by linearizing the dynamics about an operating point; the whole stability analysis rides on this step;
  • Economics: small-perturbation analysis around an equilibrium, to judge whether it's stable, uses the same linearizing intuition;
  • Numerical optimization: a loss surface near its minimum ≈ a parabola (second-order Taylor) — precisely the geometry that makes Newton's method and second-order optimizers work.
The same "valley bottom ≈ parabola" idea is called SHM in physics and linearization elsewhere.
In one line: every valley bottom is a parabola, so every small oscillation is a sine.
Ponder: if an equilibrium's potential happens to have its quadratic term vanish (say a quartic bottom), is the small oscillation still harmonic? Would its frequency now depend on amplitude?
No longer simple-harmonic. Harmonicity comes from a potential minimum approximable as a quadratic parabola (restoring force ∝ displacement); if the quadratic term vanishes and a quartic dominates, the restoring force ∝ displacement³ — a nonlinear oscillator. Its frequency then depends on amplitude (larger swings run faster) — “anharmonic.” The amplitude-independence of SHM holds only for a quadratic bottom.

Resonance: Match the Frequency, Amplify 共振

Driven oscillation · natural frequency
Intuition Pushing a swing, you don't shove randomly — you push in time with its own rhythm, each push landing just as it swings back, so a little force accumulates until it's soaring. Anything that can oscillate has a natural frequency (the rate it "wants" to move at). When an external drive hits that frequency, energy is injected in the same direction every cycle and piles up — the amplitude explodes. That's resonance.
Mechanism Drive a harmonic system with a periodic force, and its steady-state amplitude depends on how close the drive frequency sits to the natural frequency ω0. The closer, the bigger; dead-on, it peaks. How sharp and how tall that peak is depends on damping (friction, dissipation): less damping, narrower and taller peak — in the ideal undamped limit the amplitude is unbounded in principle.
amplitude drive frequency ω₀ amplitude spikes low damping high damping
As the drive frequency sweeps across the natural frequency ω₀, the amplitude spikes. Less damping means a narrower, taller peak — the same effect that lets a radio "tune in" a station and that lets resonance destroy a bridge.
Why it matters Resonance is the universal way to amplify a small input into a large response, and both edges are extreme: the good — it makes selectivity possible, a radio picking one station out of a sky full of waves; the bad — soldiers marching in step across a bridge, or a machine spinning up toward a structure's natural frequency, can pour energy into a single mode until it tears apart. The first thing an engineer does with anything that can shake is compute its natural frequency and keep the drive well away from it.
Cross-disciplinary · Biology / Engineering / Physics The same "match the frequency, amplify selectively" shows up everywhere —
  • Biology · hearing: the basilar membrane in your cochlea resonates at different frequencies along its length — high frequencies near the entrance, low ones deep inside. Your inner ear is a hardware frequency analyzer, spreading sound out by pitch before sending it to the nerve;
  • Engineering · communication: tuning a radio is tuning an LC circuit's natural frequency onto a station's carrier; MRI precisely "rings" the resonant frequency of hydrogen nuclei to image the body;
  • Physics · instruments: a guitar's body, a violin's shell are resonant cavities that amplify the string's chosen frequencies — the timbre is set by which resonances get boosted.
Wherever you need to "pick one frequency out of many and amplify it," there's resonance.
In one line: hit the natural frequency, and a tiny force accumulates into a huge swing.
Ponder: the Tacoma Narrows Bridge is often taught as the textbook case of "resonance destroying a bridge," yet engineers largely blame aeroelastic flutter (a self-excited coupling of wind and structure), not simple external resonance. Where do the two differ?
Resonance needs an external drive at a fixed frequency that, hitting the natural frequency, amplifies — energy poured in from an outside periodic force. Flutter is self-excited: a steady wind, modulated by the structure's own motion, feeds energy back to that motion in a positive-feedback phase, needing no external periodic source; past a critical wind speed it diverges on its own. One is “driven on resonance,” the other “a self-instability from structure-flow coupling.”

Waves: Passing the Vibration On 波

Mechanical waves · v = fλ
Intuition One person vibrating in place is just vibrating; but if they tug the next person, who tugs the next, the vibration starts to travel — that's a wave. Flick a long rope and a bump races from your hand to the far end; yet every little segment of rope only bobs up and down in place, it doesn't run along. This is the most counterintuitive thing about waves: what travels is the "shape of the disturbance" and its energy, not the medium itself.
Mechanism A simple wave is pinned down by three quantities: wavelength λ ("lambda," how long one full waveform is), frequency f (oscillations per second), and wave speed v (how fast the shape moves). One simple relation ties them together:
v = fλ
Each oscillation advances the waveform by one wavelength; at f oscillations per second it advances f wavelengths — so speed = frequency × wavelength. The key: wave speed is usually set by the medium (how taut the rope, how dense the air), not by how fast you shake. Shake faster (f up) and the wavelength λ shortens to compensate; the product holds.
One layer deeper, a wave obeys a wave equation, saying "a tiny segment's acceleration is proportional to how much its neighbors bend it" (∂ reads "partial," the rate of change while holding all other variables fixed):
∂²y∂t² = v² ∂²y∂x²
The left side is the up-down acceleration of a point; the right side, ∂²y/∂x², is the curvature of the waveform there: the more the neighbors bend it, the harder it snaps back. Anything satisfying this equation — rope, sound, light — propagates at speed v. That's why "wave" is one concept spanning every medium.
λ wavelength A amplitude wave speed v
A snapshot of the wave at one instant: wavelength λ is the length of one full undulation, amplitude A the maximum displacement. The whole shape slides right at speed v, while each point of the medium only bobs in place.
The counterintuitive point A wave isn't "stuff flying," it's a relay of states. A cork on the sea surface just circles in place as swells pass; it isn't carried toward shore. Sound crossing a room doesn't fly air molecules from your mouth to the listener's ear — one layer pushes the next and passes it along. This insight — "what travels is the pattern, not the matter" — will carry all the way to light, and to quantum probability waves.
In one line: a wave carries energy and shape, not the medium.
Ponder: sound is faster in warm air, and roughly ten-plus times faster in steel than in air. Since v=fλ, when a sound wave crosses from air into steel with its frequency unchanged, is it the wavelength that changes, or the frequency?
The wavelength changes. Frequency is set by the source and doesn't change across media; in v=fλ the speed v is set by the medium (a dozen-plus times faster in steel) and f is fixed, so λ scales up with v — the wavelength stretches on entering steel. Across media frequency is conserved; speed and wavelength change.

Standing Waves & Superposition: Trapping a Wave 驻波与叠加

Superposition · normal modes
Intuition When two waves meet they don't collide — they superpose: the displacements simply add, and they pass through each other and carry on as if the other weren't there. This superposition principle is what sets waves apart from particles. And when a wave is trapped in a string fixed at both ends, it reflects back and forth and superposes with itself; only the waveforms whose "ends meet up" survive, forming a standing wave that undulates in place rather than travelling — some points never move (nodes), others swing hardest (antinodes).
Mechanism The boundary condition of both ends fixed is the crux: the length L must hold exactly a whole number of half-wavelengths for the wave to be self-consistent. So the allowed wavelengths are "sieved" into a discrete set:
λn = 2Ln,  n = 1, 2, 3, …
L is the string length, n the mode number (an integer). The frequencies come in integer multiples too: fn = n f1. The lowest, n=1, is the fundamental (it sets the pitch); n=2,3,… are the harmonics (together they set the timbre). A continuum of possibilities is quantized by the boundary into a discrete string of modes — note that word.
n=1 fundamental n=2 2nd harmonic n=3 3rd harmonic node (never moves) two phases (dashed = half a cycle later)
Nailed at both ends, the string only exists in whole numbers of half-wavelengths: n=1, 2, 3… the modes are discrete. Solid and dashed are the same standing wave half a cycle apart — it breathes up and down, yet the nodes (amber dots) stay perfectly still.
The counterintuitive point A continuous string, yet its allowed vibrations are discrete — the string didn't "choose" integers, the boundary forced them. Anytime you "trap a wave inside a bounded region," you get a discrete series of allowed modes with their corresponding frequencies. Hold onto this: trap a particle inside an atom and its matter wave likewise collapses to a discrete series of standing-wave modes — the corresponding thing is the discrete energy levels. Why an atom emits only specific colors of light — the embryo is right here on this guitar string.
Cross-disciplinary · Math / Quantum / Engineering "Any state = a superposition of normal modes" is a very deep organizing principle —
  • Math · Fourier: any waveform (indeed any signal) can be decomposed into a sum of sine modes. The complex sound of a plucked string is just its fundamental and harmonics at differing strengths — timbre is a "harmonic recipe";
  • Quantum: a bound particle's wavefunction is a standing wave, and the boundary conditions sieve the energy into discrete levels — every line in the hydrogen spectrum is one allowed mode. This reading is the direct springboard to understanding quantization;
  • Engineering · modal analysis: buildings, bridges, aircraft wings all have their own natural mode shapes (a structural version of harmonics); earthquake-resistant design means computing these modes and keeping quake frequencies from exciting them.
Once you learn to see a complex vibration as a sum of normal modes, you view everything — from musical tone to atom to skyscraper — with the same pair of eyes.
In one line: trap a wave and the continuous turns discrete.
Ponder: on one guitar string, fretting shortens the effective length and raises the pitch — use λ1=2L and v=fλ to explain: is it the wave speed on the string that changes, or the wavelength?
The wavelength changes, not the wave speed. Wave speed v is set by the string's tension and linear density; fretting changes neither, so v stays essentially fixed. Shortening the effective length L shrinks the fundamental wavelength λ₁=2L, and by f=v/λ the frequency rises. On one string the speed is fixed — it's the “box holding the wave” that shortens, so the wavelength shortens and the pitch goes up.

Going Deeper

Why is the sine so special? Wouldn't some other periodic waveform do?
Two reasons. First, the sine is the unique solution of simple harmonic motion — the equation of motion for a linear restoring force F=−kx can only be solved by a sine. Second, the sine is the eigenmode of the wave equation: it passes through a linear medium unchanged in shape, merely translated, whereas square and triangle waves spread and distort in a dispersive medium. Better still, Fourier tells us any periodic waveform can be written as a sum of sines — so the sine isn't "a" waveform, it's the alphabet of all of them. Study the sine and you've studied them all.
Can resonance push the amplitude to infinity? Why doesn't a bridge "inevitably" collapse?
Only in the idealized zero-damping model does a drive exactly at the natural frequency grow without bound. In reality there's always damping (internal friction, air drag, radiated energy), so the amplitude settles at a finite peak — more damping, lower and blunter the peak. Engineering adds two safeguards: designing the structure's natural frequency far from common drive sources, and actively adding dampers (like the tuned mass damper atop Taipei 101 — a huge steel ball that swings out of phase, expressly to "drink away" resonant energy). So resonance is a danger that can be tamed, not a fate.
Is superposition a universal truth, or a conditional approximation?
It's a property of linear systems, not an iron law of the universe. As long as a medium's response is proportional to the disturbance (small amplitude), two waves add cleanly without interfering — most everyday sound and light live in this regime. But push to large amplitude, the medium turns nonlinear, and superposition fails: shock waves (like a supersonic sonic boom) steepen and break; a tsunami curls up near shore. Intriguingly, quantum mechanics' superposition principle is thought to be exact — the Schrödinger equation is strictly linear, which is the root of why quantum superposition is so "absolute," a point the "Wavefunction & Superposition" material tackles head-on.
A standing wave "doesn't move" — so where does the energy go? Does it still transmit any?
A pure standing wave transmits no net energy — it's the superposition of two equal, oppositely travelling waves, and whatever energy one sends leftward the other sends back rightward, net flow zero. Energy just sloshes between nodes and antinodes (trading kinetic for potential between the antinodes). That's why it's called "standing." A real instrument does keep leaking a little energy as radiated sound (otherwise you'd hear nothing), so you must keep plucking to top it up — but that's a small leak on top of the ideal standing wave, whose bulk is a mode that stays put.
If everything can be decomposed into sine modes, are "modes" really real, or just a convenient math trick?
Intriguing. On one hand, normal modes are indeed a mathematical choice — you could use a different basis (wavelets, say) to decompose the same vibration. On the other, modes are oddly "real": each has a definite frequency, can be independently excited or damped, and in quantum mechanics corresponds to a countable share of energy (a phonon). Perhaps the criterion is: when the system is linear, modes don't exchange energy and each evolves on its own, and this independence earns them the status of genuine degrees of freedom, beyond a mere math decomposition. Once nonlinearity enters, modes start to couple and feed each other energy, and that "reality" is discounted.

Further Reading