Van der Pol Synchronization: How Two Clocks Learn to Beat Together

An oscillator that regulates its own amplitude

The Van der Pol equation was written down in the 1920s by Balthasar van der Pol to describe vacuum-tube circuits, and it is the standard toy model for a self-sustained oscillator — a system that settles into a fixed-amplitude rhythm on its own, rather than needing to be driven or simply decaying to rest:

ẍ − μ(1 − x²)ẋ + x = 0

μ  > 0  controls both the nonlinearity and how relaxation-like the cycle becomes

Look at the damping term μ(1 − x²)ẋ closely: when |x| is small (inside the unit circle in phase space) the term (1 − x²) is positive, so the "damping" is actually negative — it pumps energy in and amplitude grows. When |x| grows past 1 the sign flips and energy is removed, shrinking the amplitude back down. The system therefore converges, from almost any starting condition, onto one specific closed trajectory in phase space called a limit cycle — not a mathematical idealisation but an attractor that nearby trajectories spiral into and stay on.

Coupling two of them: the entrainment problem

A single Van der Pol oscillator has its own natural frequency, set mostly by the linear restoring term. Couple two of them — for instance, letting each oscillator's equation feel a term proportional to the other's velocity or position — and each one's rhythm is nudged by the other's:

ẍ₁ − μ(1 − x₁²)ẋ₁ + ω₁²x₁ = k(x₂ − x₁)
ẍ₂ − μ(1 − x₂²)ẋ₂ + ω₂²x₂ = k(x₁ − x₂)

k = coupling strength,  ω₁ ≠ ω₂ = the two natural (uncoupled) frequencies

With k = 0 the two oscillators run at their own independent frequencies and drift in and out of phase forever. As k increases, something qualitatively different happens once it crosses a threshold: the two frequencies snap together and lock — a phenomenon called entrainment or frequency locking, first documented by Christiaan Huygens in 1665, who noticed that two pendulum clocks hung on the same wall would spontaneously synchronize their swings through vibrations transmitted along the wall.

Arnold tongues and phase locking

For a periodically driven or coupled nonlinear oscillator, locking does not only happen at a 1:1 frequency ratio. Plotting the region of (coupling strength, detuning) parameter space where locking occurs produces wedge-shaped regions called Arnold tongues, each one anchored at a rational frequency ratio p:q — the oscillator can lock so that it completes p cycles for every q cycles of the other. The tongues widen as coupling increases, which is why weakly coupled, mismatched oscillators only lock near exact rational ratios, while strongly coupled ones lock across a broad range of detuning. Between the tongues, for weak coupling, the relative phase instead drifts slowly — a regime called phase drift or, when the drift itself briefly hesitates near-but-not-quite-locked, phase slipping.

Why this is the standard model for cardiac rhythm

The Van der Pol equation's relaxation-oscillator shape — a slow charging phase followed by a fast, spike-like discharge as μ grows large — closely mimics the shape of a real action potential, which is why coupled Van der Pol (and closely related FitzHugh-Nagumo) oscillators are the textbook model for how the heart's natural pacemakers stay in step. The sinoatrial node is the heart's dominant pacemaker, but it is really a population of thousands of individual pacemaker cells, each with its own slightly different intrinsic rate, that must synchronize with each other to produce one coherent heartbeat, and it in turn entrains subordinate pacemaker tissue (like the atrioventricular node) at a 1:1 locking ratio under healthy conditions. Cardiac arrhythmias, including certain forms of AV block, are naturally described as the system falling into a different Arnold tongue — a stable p:q lock other than 1:1 — or losing the lock entirely and drifting.

Reading the demo

The simulation integrates both oscillators' equations with an explicit numerical scheme and plots their phase-space trajectories together with the phase difference between them. At low coupling the two limit cycles rotate at visibly different rates and their relative phase sweeps steadily through all values. Raise the coupling slider past the locking threshold and the phase difference stops changing — the two trajectories rotate together, offset by a fixed angle, exactly the transition Huygens first watched happen between two clocks on a shared wall.

Frequently asked questions

Try it live

Everything above runs in your browser — open Van der Pol Synchronization and change the parameters while it is running. Nothing is installed, nothing is uploaded, the whole model lives in one tab.