Parabolic mirror
Reflects off the true parabola, so a source at its focus leaves exactly collimated at any aperture — no spherical aberration, unlike a spherical mirror.
Open in the canvas →In the real world
A parabola has an exact geometric property a sphere only approximates: every ray traveling parallel to its axis, at any distance from that axis, reflects through a single focus. There is no spherical aberration to correct for, which is why fast telescope primaries, off-axis paraboloid (OAP) mirrors in ultrafast laser labs, and satellite dishes are all parabolic rather than spherical. In this 2D side view, the mirror profile is the parabola with vertex at the origin and focus a distance f behind it:
In OpticalSetup
The mirror is traced as its real curve. Short flat facets locate where a ray lands, but the reflection uses the parabola's own normal at that point, found analytically: the surface x = −y²/4f has gradient (1, y/2f), and the exact ray–curve intersection is solved rather than taken from the facet chord. The facet count therefore sets positional accuracy only, never angular — which is what makes the defining property hold at any aperture rather than only at gentle ones.
The consequence is worth checking against the spherical mirror directly. With a point source at the focus of each, f = 25 and a 100 mm aperture, the parabola returns a beam 98 mm wide at 400 mm and still 98 mm wide at 1200 mm — collimated, exactly. The sphere returns 468 mm widening to 792 mm: about 11°. Neither number is put in by hand; both come out of the two surfaces.
This is closer to first-principles optics than most elements in the library, but it is still a 2D on-axis cross-section — a real OAP is typically an off-axis section of a 3D paraboloid, which this side view cannot represent. Being exact in reflection also means it is exact in a way no manufactured mirror is: there is no surface figure error, no roughness, and no coating model, so reflectivity does not vary with wavelength or angle.