We propose an experiment to test whether the one-way speed of light is isotropic, without relying on clock synchronisation between separated stations. Two identical stations separated by a fixed distance exchange laser beams through complementary shutters driven by independent atomic clocks. Because the shutters are complementary, the two detector readings always move in opposite directions as the phase offset at one station is swept — they never reach maximum simultaneously. The correct observable is the phase offset at which the two detectors read equal intensity: the crossing point. In the isotropic case this crossing point is stable as Earth rotates. In the anisotropic case it drifts cyclically with the sidereal period. No synchronisation signal is required and the measurement is not circular.The experiment provides two independent observables recorded in parallel. The first is the optical crossing point described above. The second is the timestamp time series δ_WE and δ_EW recorded independently at each station. In the isotropic case both series show no variation. In the anisotropic case both series show a cyclic variation locked to the sidereal day. Both observables respond to the same underlying quantity — the asymmetry between W→E and E→W propagation times — and both share a common-mode rejection of local environmental disturbances by virtue of the 50-metre station separation.Frequency identity between the two stations rests on the constancy of the atomic transition — a physical fact, not an engineering convention. The only signals exchanged between stations are timestamp pulses transmitted at regular intervals to build up the time series described in Section 5.2. No synchronisation signal of any kind is used. The phase relationship between the two shutter drives is never imposed externally; it is found empirically by the crossing-point procedure. The experiment is sensitive to any preferred-frame velocity component along the baseline. A null result constrains preferred-frame theories. A non-null result warrants careful independent replication.