Then any reflection from the focal point goes right back into that point. LEDs and LEPs are not point sources but they mostly symmetrical and an off-axis reflection just hits the opposite side of the emitter.
Well the parabolic version could well be doable… Im still not convinced on the hemispherical one maybe im just being thick here I cant see why they would choose a difficult costly material like glass unless it was needed , and the only reason iv got is refraction ? my understanding is that curving the mirror does not change the angle of reflection , it would still be reflective not retro reflective , having it spherical will however correct for the divergence and refocus the reflected beam
The W30 is marketed as a dive light (claimed 100m water resistance), and it just “fell apart?” That doesn’t give me much faith in their water resistance claims.
This “concave mirror” is parabolic not spherical.
With a spherical mirror, rays coming from outside wouldn’t focus on a single spot. But ones originating from the focal point would go right back where they started. That’s because regardless of where they land on the mirror, they would be orthogonal to the mirror’s tangent plane at that point. That orthogonality is literally the definition of a sphere….and light rays hitting a mirror orthogonally are reflected right back.
ADDED:
I think the misunderstanding might be the focus on rays that exit the mirror.
With a full, properly positioned hemisphere there should be none.
Reflective collars work because the hemisphere has a cutout (typically right in the middle) through which light can escape towards the lens.
See the marinebeam drawings.