Thought Toys · Waves & rhythm · Exhibit 89
One dot released near the center. Another released far outside, wherever you drag it to. Same equation, same rule, two very different starts. Turn the pull up from zero and watch: however far apart they began, both spiral in and settle onto tracing the exact same closed loop.
—
near · far · pull μ — · far start —
Two dots, same equation, same rule, nothing else touching either one. The amber dot starts almost exactly at the center. The cyan dot starts wherever you drag its slider to — right next to the amber one, or way out at the corner. At zero pull, that is the whole story forever: each dot moves along its own closed circle, keeping whatever size it started at, and the two never come any closer together. Nothing pulls a big loop down to a small one, or a small loop out to a big one. That is not a limit cycle — it's just the plain harmonic oscillator, one closed orbit for every possible starting size, all equally stable, all permanently different.
Now turn the pull up. A single rule switches on: push outward whenever a dot is inside a certain curve, brake it whenever it's outside that same curve. Inside the curve the "damping" term is negative — it adds energy; outside it, the same term flips sign and removes energy. A dot that starts small gets pushed out. A dot that starts huge gets reeled in. There is exactly one curve where the two effects balance over a full lap, and every dot, from every starting point, ends up tracing it. Watch the amber and cyan trails long enough and they stop being two different shapes — they become the same shape, retraced by two dots that simply arrived from opposite directions.
Push the pull dial further and the loop itself changes character: from a near-circle at low pull into a stretched, lopsided racetrack — a slow crawl along one side, then a fast snap across to the other. That slow-crawl-fast-snap shape is a relaxation oscillation, and it's why this equation turns up everywhere from vacuum-tube circuits to heartbeats: it describes anything that charges up slowly and discharges in a rush, over and over, always finding its way back to the same rhythm no matter how it was disturbed.
improve/verify/89-van-der-pol.js, ~20 checks): at μ = 0 the
measured period matches 2π to within 0.001; for μ ∈ {0.5, 1, 2, 4, 6}, a trajectory started at (0.01, 0) and
one started at (6, 6) settle onto curves within roughly 1% of each other's scale, and that agreement
genuinely tightens over time — the two curves are 26× closer together after 1200 time units of
settling than after just 30; the period rises monotonically and more than doubles from μ = 0 to μ = 8,
trending toward the textbook large-μ asymptotic T ≈ (3 − 2 ln 2)μ. Negative control:
at that same equation with μ = 0, a trajectory started at radius 0.5 and one started at radius 3.0 stay
exactly where they started — 2000 time units later, still 2.5 apart — proving the convergence above is a
genuine consequence of μ > 0, not just "oscillators tend to end up somewhere similar."
Also in Waves & rhythm: Fourier epicycles →
← the cabinet · Thought Toys — a cabinet of explorable explanations. Exhibit 89.