Who says you're limited to one? For that matter, if you're just trying to find the direction of the beam, why bother with "good sensor" at all?
Set up a bunch of not-quite-parallel tubes with simple and durable light-sensors in the "deep" end, and whichever sensor measures the highest must be pointed at the light-source.
> The Sidewinder's seeker used an ingeniously clever optical arrangement, with a Cassegrainian mirror fitted with a tilted secondary mirror. The secondary mirror rotated in unison with a reticle, projecting the whole instantaneous field of view of the mirror through the reticle onto a filter/detector assembly. Because the mirror secondary was tilted, rotating it about the missile's axis swept the cone of the mirror's field of view about the missile's axis in a fashion analogous to a conical scanning radar seeker (see diagram).
