loading...
running...
The science
Superposition. A ripple tank is linear to a good approximation, so the surface height is the sum of the waves from each dipper. Each source i at distance ri sends out a circular wave whose amplitude falls with distance and whose phase depends on the path length:
h(x,y,t) = Σi sin(k ri - ω t + φi) / (1 + α ri), k = 2π/λ, ω = 2π c/λ, φi = i·Δφ
Here α is the decay slider (real circular waves lose amplitude roughly as 1/√r; the 1/(1+αr) form keeps the field finite at the dipper). The surface you see, the buoys and the lamp projection all use this same formula as the 2D companion page.
Constructive and destructive interference. For two sources the phase difference at a point is 2πΔr/λ + Δφ, where Δr = r1 - r2. In phase (Δφ = 0):
Δr = m λ (bright, antinodal lines) Δr = (m + 1/2) λ (dark, nodal lines)
Both families are hyperbolas with the two sources as foci. Since |Δr| ≤ d (the triangle inequality), only m + 1/2 < d/λ can occur, so the number of nodal hyperbolas is 2·ceil(d/λ - 1/2). In the far field the n-th nodal direction satisfies d cos α = (m + 1/2) λ, which is the same condition as for a double slit. If d is less than λ/2 there is no nodal line at all.
Time-averaged intensity. The mean of h2 over one period is exactly half the squared length of the phasor sum:
<h2> = (1/2) | Σi ai ej(k ri + φi) |2, ai = 1/(1 + α ri)
This is what the "time-averaged intensity" colouring and the two amplitude graphs show, and the buoys that barely move mark the nodal lines.
Phased array. With a phase step Δφ between neighbouring sources the nodal and antinodal lines swing sideways without moving the dippers: the principle behind phased-array radar, sonar and ultrasound beam steering.
Lamp projection. A lamp above a real ripple tank projects bright lines where the surface acts as a converging lens (curvature, ∇2h < 0) and dark lines where it diverges. The floor image here is drawn from minus the discrete Laplacian of the displayed surface, a first-order picture of that focusing.
Checks (run in the page). (1) the time average of h2 sampled over ten periods against the exact phasor formula; (2) the first far-field nodal direction on a large arc against acos(λ/2d); (3) the number of intensity nulls on a far circle against 4·ceil(d/λ - 1/2).
Simplifications. Deep-water dispersion, damping, reflections from the tank walls and surface tension are ignored (waves are non-dispersive with c = 1 in model units and pass through the walls). The displayed height is scaled by the Visual wave height slider; the physics is linear, so it does not change the pattern. The 1/(1+αr) amplitude law is a model choice, not a derived result.