Every introductory physics course teaches that a pendulum's period is
T₀ = 2π√(L/g), independent of amplitude. That formula is only the
first term of a much longer story: it comes from replacing
sin θ with θ in the pendulum's equation of motion, which
is only accurate for small swings. Release the bob from a wide angle and the true
restoring force is weaker near the top of the swing than the linear approximation
assumes, so the real period grows longer than T₀.
θ'' = -(g/L) sin θ.θ(t) = θ₀ cos(ω₀t) with ω₀ = √(g/L), released from rest at the same instant and amplitude.T = 4√(L/g) · K(sin(θ₀/2)), evaluated live via the arithmetic-geometric mean.At θ₀ = 90° the true period is already about 18% longer than the small-angle formula predicts, and it diverges to infinity as θ₀ approaches 180° — a pendulum released exactly upside-down would, in a frictionless world, take forever to complete its first swing.
Two pendulums released together from the same amplitude: one integrated exactly from the nonlinear equation of motion, one following the textbook small-angle formula. Watch them fall out of sync as amplitude grows.
The small-angle approximation sin θ ≈ θ underestimates the restoring force at wide swings, so the real pendulum's period is always longer than T₀ = 2π√(L/g) predicts — and the gap grows nonlinearly with amplitude.
Drag the amplitude slider from a few degrees toward 170° and watch the red pendulum drift behind its pale ghost. Change length or gravity to rescale the timescale, and compare the live period and phase-drift readouts.
At 90° amplitude the true period is already about 18% longer than the small-angle formula; the period diverges to infinity as the release angle approaches 180°.