Presets
Playback
Parameters
Camera
Layers
Quality
Model check
The science
What is modelled
Two tuning forks stand on wooden resonance boxes with their open ends facing each other. Hit one fork and the other starts to sound by itself if, and only if, it has the same natural frequency. Each fork is a pair of steel cantilever tines, so its pitch follows the tine length; a clip weight on one tine lowers the pitch. The air between the boxes couples the forks weakly, which the page models as a coupling stiffness that falls with the gap. The result is two coupled damped oscillators: with matched forks the energy sloshes completely from one fork to the other and back; with mismatched forks only a little gets across. The bench, boxes, forks (with tines made of bendable segments), mallets and props are Blender-built models. Time is slowed so that 440 Hz can be seen, and the tine bending is drawn larger than real.
Equations
- Cantilever tine (Euler-Bernoulli):
f = (λ1²/2π) (t/L²) √(E/12ρ)with λ1 = 1.875, thickness t = 4 mm, steel E = 200 GPa, ρ = 7850 kg/m³. For L = 86 mm this is 440 Hz. - Added mass at the tip:
f = f0 / √(1 + mclip/(0.2357 mtine)). - Coupled forks:
ẍ1 + 2ζω1 ẋ1 + ω1² x1 + K (x1 − x2) = 0, and the same for fork 2; ζ = 1/(2Q). - Normal modes of identical forks without damping:
ω− = ω0,ω+ = √(ω0² + 2K); the energy changes from fork 1 to fork 2 in the timeπ/(ω+ − ω−). - Free decay: the amplitude envelope is
e−ωt/2Q, the energye−ωt/Q. - Energy of the pair per unit mass:
½(v1² + ω1²x1² + v2² + ω2²x2² + K(x1 − x2)²), constant when ζ = 0.
Checks run in the page
- Simulated energy-exchange time of two matched forks against the normal-mode prediction.
- Single-fork ring-down against the analytic envelope at t = 2Q/ω.
- Energy conservation of the coupled pair after 300 periods.
Honest simplifications
The fork is treated as one effective oscillator per fork with the cantilever frequency (the real U-shaped fork has slightly different end conditions and a second, much higher overtone that is not modelled). The acoustic coupling is a single stiffness that scales as 1/gap, tuned so that the transfer is visible; it is not a calculation of the sound field, and the resonance boxes are represented only by the coupling and a fixed Q. The sound is a sine at the model frequency with the model envelope, played at real pitch while the picture is slowed. The air, the table and the boxes are not simulated.
What to try
- Strike fork 1 with matched forks: fork 2 grows to the same amplitude while fork 1 dies down, then they swap back.
- Add a small clip weight to fork 2: the transfer drops sharply, even though the pitch changes by only a few hertz. The curve under the graphs shows how narrow the resonance is.
- Move the boxes close together for faster, stronger exchange, or far apart for slow, weak exchange.
- Raise Q to make the forks ring longer, lower it to see the resonance widen.