About Pendulums
Two classic lecture-bench demonstrations. PENDULUM WAVE: a row of 12 to 24 pendulums, each a little shorter than the last, pulled back together and let go. They start in line, drift into travelling waves and snakes, split into two, three and four rows, look like a random mess, and then all come back into line at the end of the cycle. Press RUN (PAUSE stops it, RESUME carries on), or drag the time line to go straight to any moment; click a bob (or use ‹ ›) to see its own sums. DOUBLE PENDULUM: a pendulum hung from the bottom of another. Drag either bob to set the start, press RUN, and a twin started a hair's breadth away runs alongside it. Space plays or pauses; COPY LINK copies the current setup.
THE STORY
The story, told by his pupil and first biographer Vincenzo Viviani, is that the young Galileo Galilei, in Pisa Cathedral, timed a swinging lamp against his own pulse and found that wide and narrow swings took the same time. Later experiments showed him that the time of a swing does not depend on the weight of the bob, only on the length of the string, and that four times the length takes twice the time. His belief that every swing, however wide, takes exactly the same time is only nearly true: a swing out to 90° takes about 18% longer than a small one.
Christiaan Huygens built the first pendulum clock in 1656 and patented it the next year; it made clocks many times more accurate almost overnight. In his Horologium Oscillatorium (1673) he gave the time of a small swing, 2π√(L/g), the formula this app uses, and showed that a bob made to follow a cycloid instead of a circle keeps exactly equal time however wide it swings.
The pendulum wave has been a lecture-theatre favourite for a long time; its exact origin is unclear. Harvard's Natural Sciences Lecture Demonstrations describe a version with 15 pendulums in which the longest makes 51 swings in 60 seconds and each shorter one makes one more, up to 65. That's the setup this page opens with.
The double pendulum is one of the simplest machines that is chaotic. Its equations are short and nothing about it is random, yet for big swings a tiny difference in the start grows roughly exponentially until the two runs have nothing to do with each other. Henri Poincaré met this kind of behaviour in the motion of three planets in the 1880s and 1890s; Edward Lorenz's weather model of 1963 made it famous as the "butterfly effect". For small swings the double pendulum is tame: with equal arms and bobs it has two steady patterns (normal modes), the lower arm swinging √2 times as far as the upper one, either with it (TOGETHER) or against it (OPPOSITE).
HOW IT WORKS
The wave. If the longest pendulum makes N swings in the cycle time C, pendulum k makes N + k, so its period is T = C ÷ (N + k) and, from Huygens' formula, its length is L = g (T ÷ 2π)², with g = 9.81 m/s². At time t pendulum k has made (N + k) t ÷ C swings, so each bob is exactly t ÷ C of a swing ahead of the next longer one, the same for every neighbouring pair: that's why the row makes a wave. At a quarter of the cycle neighbours are a quarter of a swing apart (rows), at half they're half a swing apart (two rows), and at the end every one has made a whole number of swings and they're in line again. Every bob's position is worked out straight from the time, not stepped forward, so the time line can jump anywhere exactly. The sums use the small-swing formula; the bench shows how much a real pendulum at that swing would run slow, which is why real demonstrations keep the swings small.
The double pendulum. Two bobs on light rigid rods with no friction (the upper arm 1 m and its bob 1 kg; the lower arm and bob can be changed). The equations of motion come from Lagrange's method and are stepped forward by the fourth-order Runge–Kutta method (RK4) in steps of half a thousandth of a second (2,000 a second). With no friction the energy should never change, so the energy readout is a check on the arithmetic: with RK4 it stays the same to millionths of a joule or better. Switch the stepper to EULER (the simplest method, one slope per step) to see an honest-looking pendulum quietly gain energy. The twin starts with its lower arm turned a little further (1°, 0.01° or 0.0001°); the top chart plots the angle between the two lower arms on a log scale, where steady exponential growth is a straight line. Because the motion is chaotic, after a while no computer can say exactly where the real pendulum would be, and a steady energy only shows the arithmetic isn't drifting, not that the path drawn is the exact one: halving the step and comparing is the stronger check, and even that only holds until the chaos takes over.
FURTHER WATCHING AND READING