# Scope and limitations The trace stops at the camera entrance. The effects below are outside it. Their impact depends on the instrument and its model parameters. ## Misalignment of aspheric (dual-mirror) mirrors The primary and secondary of the Schwarzschild-Couder telescopes (SST, SCT) are even polynomial surfaces (`mirror_class` 2) built from segments. In the real telescope each segment, and the secondary as a whole, can be tilted or shifted from the design. The alignment errors (`mirror_align_random_distance`, `_horizontal`, `_vertical`) and `random_focal_length` act only on facets of segmented dishes (`mirror_class` 0). Dual-mirror surfaces are traced at their design position; `primary_segmentation` and `secondary_segmentation` define gaps between segments. `mirror_reflection_random_angle` tilts the surface normal at each reflection (a convention, not a ray angle: see [mirror scatter conventions](optics.md#mirror-scatter-conventions)). For the resulting PSF comparisons, see [Validation](VALIDATION.md). ## Curvature and refraction of the camera window (LST and others) A camera window is a transparent plate, flat or curved, in front of the pixels. Refraction in it shifts each photon sideways, moves the focus backwards and delays the photon. A curved window also acts as a weak lens. The window and its transmission belong to the downstream camera simulation; window geometry and refraction are not traced. **Size of the effects** for a flat plate of thickness t and refractive index n (PMMA n = 1.49 as an example; take each camera window's thickness from the camera documentation): | t | lateral shift at 10 / 16 / 25 deg incidence | focus shift t(1 - 1/n) | extra delay (n_g - 1) t / c | | --- | --- | --- | --- | | 3 mm | 0.17 / 0.28 / 0.46 mm | 1.0 mm | 5 ps | | 5 mm | 0.29 / 0.47 / 0.77 mm | 1.6 mm | 8 ps | | 10 mm | 0.58 / 0.95 / 1.5 mm | 3.3 mm | 17 ps | The lateral shift is t sin θ (1 - cos θ / sqrt(n² - sin² θ)); the delay uses a group index n_g ≈ 1.5. The mean incidence angle on the LST focal plane is 16 deg ([validation](VALIDATION.md)). For comparison: the LST D80 is about 3 cm, its pixels about 5 cm wide, and a focus error δ blurs the image by about δ D / f = δ / 1.2 for the LST. A shell of uniform thickness t and radius of curvature R has optical power of order (n - 1) t / R²: a focal length of kilometres for millimetres and metres, negligible against 28 m. The focus shift is the only one that is not negligible. It belongs with the camera position: `focus_offset` (6.55 cm for the LST) was tuned to a measured PSF, so the window is already included. Adding the window to the ray tracing would count it twice unless `focus_offset` were re-derived at the same time. The origin of each telescope's `focus_offset` must be audited before any window refraction is added. ## Other effects not traced - **Telescope structure of LST, SST and SCT.** Only the MST has a `telescope_structure` table in the models examined. The other telescopes use `telescope_transmission`, a parametrisation of shadowing. - **Pixels, light guides, photodetectors:** downstream camera simulation. - **Refraction in air** between an illuminator and the telescope: a few arcsec over hundreds of metres. - **Diffraction and polarisation:** irrelevant at the size of IACT mirrors and for unpolarised Cherenkov light.