Cymatics Tone Plate

Most standing-wave demos ask you to pick a mode — type in n = 3, m = 5, get a figure. That skips the part real cymatics is actually about. On a physical rig you do not choose a mode; you choose a frequency. A signal generator feeds a loudspeaker bolted to the underside of the plate, you turn the dial, and for most of the dial nothing much happens. Then you cross a resonance and the plate suddenly comes alive.

This simulator works the same way. The only thing you set is the drive frequency. Every mode the plate can support is then excited in proportion to how close the drive is to that mode's own resonant frequency, using the standard driven-oscillator amplitude response:

The displayed surface is the weighted sum of all the modes with meaningful amplitude. Between resonances several modes contribute at once and the picture is a soft, unresolved blur — exactly what a real plate looks like off-resonance. Land on f = fmode and that one term blows up to Q while the others stay near 1, so a single clean figure snaps into focus. The Q slider controls how sharply that happens: low Q (heavily damped, like a plate with a rubber mat under it) gives broad, mushy resonances; high Q (a ringing steel plate) gives razor-thin ones you have to hunt for.

Ernst Chladni published Entdeckungen über die Theorie des Klanges ("Discoveries in the Theory of Sound") in 1787. His method was almost absurdly simple: scatter fine sand on a thin brass plate clamped at its centre, then draw a violin bow across the edge. The plate rings at one of its natural frequencies, the sand dances, and within seconds it settles into a sharp geometric figure. Change where you bow or where you damp the edge with a fingertip and you get a completely different figure.

Chladni toured Europe demonstrating this, and in 1808 he performed it for Napoleon. Napoleon was impressed enough to fund a 3,000-franc prize through the Institut de France for whoever could produce a mathematical theory explaining the figures. It turned out to be a genuinely hard problem. The prize was won in 1816 by Sophie Germain — on her third attempt, after two rejected submissions — who derived a fourth-order partial differential equation for the vibrating elastic plate. Her equation had an incorrect boundary-condition treatment, which Kirchhoff corrected in 1850, but the core insight was hers. Germain had taught herself mathematics from library books and initially corresponded with Lagrange under a male pseudonym, since women were barred from the École Polytechnique.

A standing wave on a plate has two kinds of place. Antinodes are where the surface swings through the largest displacement. Nodes are the curves where the displacement is permanently zero — the plate on either side is moving in opposite directions, so the line between them cannot move at all.

A grain of sand sitting at an antinode is being thrown vertically upward many hundreds of times a second. As soon as the plate's downward acceleration exceeds g (about 9.8 m/s²) the grain loses contact, goes ballistic, and lands somewhere slightly different. Repeat that a few thousand times and the grain performs a random walk that is biased away from every high-amplitude region. The one place it can land and stay is a nodal line, where the plate never launches it. Nodes are not attractive; they are simply the only places that are not repulsive. The sand accumulates there by elimination.

The Sand render mode reproduces this directly: each particle takes a random step whose size is proportional to the local |displacement|, so particles in loud regions wander fast and particles in quiet regions barely move. Change the frequency while sand mode is running and you can watch the whole population physically migrate to the new figure over a few seconds — which is the part that never comes across in a still photograph.

A drumhead and a metal plate both make standing waves, but their restoring forces are completely different, and that changes how their resonances are spaced. A membrane (a stretched skin with no stiffness of its own) is pulled back by tension, governed by the second-order wave equation. A plate resists being bent, governed by Germain and Kirchhoff's fourth-order biharmonic equation.

The plate's outline decides which figures are possible, because the boundary is what quantises the modes. Two shapes are offered here and they look nothing alike.

• Square plate. Each mode is the symmetric combination Z = cos(nπx)cos(mπy) − cos(mπx)cos(nπy). The minus sign enforces the plate's diagonal symmetry and is what bends the nodal lines into curves rather than leaving a plain checkerboard. Note that n = m makes the expression cancel to zero — that combination produces no figure at all.
• Circular dish. Modes separate into radial and angular parts: Z = Jn(αns r/R) · cos(nθ). You get n nodal diameters (straight lines through the centre) and s − 1 nodal circles. This is the geometry of the classic cymatics dish and of a real cymbal, and it is why circular figures look like spokes-and-rings rather than lattices.

The simulator evaluates Jn directly — a convergent power series for small arguments and the Hankel asymptotic expansion beyond x ≈ 12 — against a table of the zeros of Jn′, which are the correct eigenvalues for a rim that is free to move rather than clamped. In a real dish filled with a shallow layer of water, the water simply obeys the plate underneath it; the ripples you see on the surface are the plate's own nodal geometry made visible by refraction and specular reflection, which is what the Water render mode is imitating.

The word cymatics was coined by the Swiss physician Hans Jenny in his 1967 book Kymatik, from the Greek kûma, "wave". Jenny built an apparatus he called the tonoscope and photographed vibrating plates loaded with sand, powders, pastes and liquids. His photographs are genuinely beautiful and did a great deal to popularise the subject.

It is worth being precise about what is going on in those images, because cymatics attracts a lot of claims that the physics does not support. Every figure on this page — and every figure in Jenny's book — is the solution to a well-understood boundary-value problem in classical mechanics, worked out between Germain in 1816 and Kirchhoff in 1850. The pattern is a property of the plate: its shape, size, thickness, material stiffness and how it is supported. The frequency selects which of the plate's pre-existing modes gets excited. Two plates of different geometry driven at the same frequency give different figures, and the same plate driven at the same frequency gives the same figure every time.

Where cymatics-style visualisation earns its keep in practice is instrument making and engineering. Luthiers still tune violin and guitar plates by tapping them and watching Chladni patterns form, shaving wood until the modes fall where they want. Engineers run the same analysis with accelerometers and laser vibrometers instead of sand to map the natural modes of aircraft panels, brake rotors and spacecraft structures — because a component driven at one of its own resonances is a component that eventually fails.