Acousto-optic modulator (AOM)
Deflects light into a first diffraction order with adjustable efficiency, optional zero order, and square, sine, or sawtooth modulation of the drive.
Open in the canvas →In the real world
An AOM diffracts light off a traveling sound wave launched into a crystal by a piezoelectric transducer driven at an RF frequency. In the Bragg regime, light incident at the Bragg angle diffracts efficiently into a single order, shifted in frequency by exactly the drive frequency (up-shifted or down-shifted depending on propagation direction relative to the sound wave):
In OpticalSetup
Deflection and modulation efficiency are direct configurable parameters rather than quantities derived from crystal or drive properties. The optical frequency shift is not applied, and there is no drive-frequency control: the shift is real — it is the whole basis of heterodyne detection — but at 7.6×10−5 nm for 80 MHz at 532 nm it is a thousand times finer than any wavelength difference this workbench resolves, since every readout here rounds to the nearest nanometre. Carrying it only ever moved a number nothing could report. The AOD had already declined it for the same reason. Gating support lets the modeled RF drive vary in time, which the pulse-timing overlay reads as a temporal gate on the beam. Three waveforms are offered, named for the drive a function generator would supply: square switches the diffracted order fully on and off and is the only one with a duty cycle; sine and sawtooth sweep the drive continuously and are described by a modulation depth instead, swinging between 1−depth and full transmission. Both continuous shapes therefore average 1 − depth/2 over a period, which is the power a detector with no temporal resolution reads.
The ramp carries the symmetry control a function generator puts on its own ramp output. Rise fraction is how much of the period is spent climbing: 1 is the rising sawtooth, 0 the falling one, 0.5 a triangle, and anything between an asymmetric triangle peaking at exactly that point in the period. Sweeping it changes the shape without changing the average, so it never doubles as a brightness control.
A square gate switches the diffracted order fully on and off, so it can be drawn in chunks rather than as a uniformly dimmed line — the same schematic footprint the chopper already uses for gated CW light. Draw gated beam chopped controls it, and it is a drawing choice alone: the traced power stays duty-averaged and every detector reading is identical either way. The continuous waveforms are never chunked, because they have no on/off edges to draw.
With Keep 0th order on, both orders are chunked in opposition: light returns to the undiffracted beam exactly while the RF is off, so one is lit wherever the other is dark. Both beams still carry their duty-averaged power, and the two orders always sum to the incident power.
A detector's time trace shows the levels rather than just the shape. Its vertical axis is absolute for a single beam — full height is one whole source beam, and light lost upstream draws short rather than being rescaled back — so modulation efficiency is visible as the contrast it really sets: at η = 0.5 the diffracted order peaks at half height while the undiffracted one only falls to half, in opposition, and the two sum to the beam at every instant. At η = 1 both swing the whole way. Light lost anywhere upstream shortens the trace in the same way, instead of being normalized back to full scale.
Deflection angle and modulation efficiency are set directly by you, not derived from the Bragg condition, RF power, or interaction length — this is a schematic acousto-optic model, not a Bragg-cell simulator. Modulation efficiency is the crystal's diffraction efficiency under another name: it is the fraction of the beam that can be switched, which is exactly what limits the contrast of both orders. The chunk spacing is schematic too: a real megahertz gate would put its chunks micrometres apart, so a fixed on-screen period is drawn instead, exactly as pulse markers are spaced for legibility rather than to scale, and the two orders share that period rather than each following its own RF timing. The chunks are also idealized in depth: both orders are drawn fully dark between chunks, while a real diffracted order only reaches the configured efficiency and a real zeroth order keeps 1−efficiency of the beam rather than extinguishing. Drawing that residual as its own branch would have let a display setting change a detector reading, which the chunks must never do.