Math Methods · Learn
Two PDEs that look almost identical — one with a first time-derivative, one with a second — and that tiny difference gives them opposite souls: one diffuses and forgets, the other oscillates and remembers.
Both equations yield to the same idea: guess that the solution factors into a shape times a time,
Substitute, divide, and the PDE splits into two ordinary ODEs — one for the spatial shape $X$, one for the time evolution $T$ — linked by a separation constant.
Fixed ends, $u(0,t)=u(L,t)=0$, force the spatial part to be the eigenfunctions
Any initial state is first broken into these modes — a Fourier sine series (see the Fourier page) — and then each mode evolves independently. All the difference between heat and waves lives in how that mode evolves in time.
Each mode decays, and the exponent grows with $n^2$ — so high-frequency wiggles vanish almost instantly while the smooth fundamental lingers. That's why diffusion smooths: a jagged initial profile melts to a gentle hump and fades to flat. The sharp detail is gone forever.
Now each mode oscillates and never decays, so the high modes keep ringing alongside the low ones. Sharp corners survive, and the whole shape periodically reassembles itself — a plucked string returns to its starting form every period. (Equivalently, d'Alembert's solution is two copies of the initial shape traveling in opposite directions.)
Load a square pulse and run Heat — the corners melt as the spectrum collapses from the high end. Switch the same pulse to Wave — the corners persist and travel. Same modes, opposite fate. Open the lab →
Bjorn Poonen's free MIT 18.03 lecture notes (PDF) derive the heat equation in §29 and the wave equation in §30, working through the same separation-of-variables steps this lab animates.
EngineeringCandy · Learn · the theory behind the playground