Concave mirror
Focuses reflected rays off a real spherical surface of radius 2f — marginal rays cross ahead of the paraxial focus, which is the aberration a parabolic mirror exists to avoid.
Open in the canvas →In the real world
A concave (converging) spherical mirror focuses light by reflection the same way a lens focuses it by refraction. For a mirror of radius of curvature R, the paraxial focal length is half the radius, and object and image distances obey the same mirror equation as a lens:
That formula is only exact for rays close to the axis. A real sphere brings marginal (off-axis) rays to a focus slightly closer to the mirror than paraxial rays — spherical aberration — which is why fast astronomical mirrors are ground as parabolas instead (see the parabolic mirror page).
In OpticalSetup
The mirror is a real spherical surface: radius R = 2f, vertex at the element's origin, centre of curvature in front of it. Rays are intersected against that circle analytically and reflected off its true normal, with no paraxial correction applied anywhere. Nothing about the focusing is imposed — it falls out of the geometry, and so does the aberration.
That means this element behaves like a sphere rather than like an idealisation of one. Put a point source at the focus and the returning beam is not collimated: marginal rays leave at a different angle from paraxial ones, and the beam widens as it travels. How badly depends entirely on how fast the mirror is:
f/3.9 — 0.07°, near enough to collimated to use. f/2.0 — 0.14°. f/0.5 — 5.7°, useless for the purpose. Same source, same focus, only the aperture-to-focal-length ratio changing. That steep dependence is the whole reason a fast system is built round a parabola instead, and the two elements are worth putting side by side to see it.
The surface is a 2D cross-section of a sphere, so only aberrations that live in the meridional plane can appear — spherical aberration and defocus do, while astigmatism and coma, which need the third dimension or a full off-axis field, do not. There is no coating model, so reflectivity does not vary with wavelength or angle of incidence. And a mirror cannot be wider than its own sphere: a short focal length with a wide aperture is limited to what the radius allows, and the panel reports the aperture actually used rather than the one requested.