An almost flat plate can perform precise optical work
At the front of the telescope sits a transparent circular plate. At first it seems merely to close the tube, while the large mirror behind it collects the light. Replacing it with an ordinary flat window, however, changes the image. The important detail is a weak profile almost invisible from the side: different parts of the plate add different amounts of optical path.
In a Schmidt system, an aspheric corrector compensates spherical aberration in the mirror system. The classical Schmidt camera uses this to obtain a wide field; a Schmidt–Cassegrain adds a secondary mirror and folds the path into a compact tube. Corrector location, surface shapes and additional lenses determine the remaining errors. Glass in front of a mirror does not by itself guarantee a flat field. Celestron: SCT components and limitations.
The Newtonian reflector article explains why a paraboloid is useful for parallel axial light. Here we follow another route: retain a spherical substrate and add correction. The numbers and figures below are teaching examples; the author has neither fabricated a corrector nor physically tested an SCT.
Schmidt was solving the problem of photographing a large area of sky
A survey needs more than a sharp star at the centre of a frame. It needs to record many stars across the field at once. A paraboloid works well on axis, but develops coma away from the axis. Bernhard Schmidt built a wide-angle camera at Hamburg Observatory in 1930; his own account shows how to separate spherical aberration from off-axis errors. Hamburg Observatory’s chronology.
Schmidt’s account, signed in January 1931, describes a spherical mirror, an aperture stop at its centre of curvature and a correcting plate. Richard Schorr’s January 1932 addendum describes a working instrument: a corrector with a 36 cm clear aperture, a 44 cm mirror and a 62.5 cm focal length. The mirror diameter therefore exceeds the entrance aperture. Schmidt: original, pages 15–17 and addendum.
A major application was Palomar’s 48-inch camera, now the Samuel Oschin Telescope. It took its first official photograph in September 1948 and began large surveys in 1949. The photographic plate was held inside the tube, between the corrector and primary mirror. This was a field-recording camera, rather than a familiar tube with an eyepiece at the back. Palomar: construction and history.
Modern digital successors retain this aim: compare large sky regions at different times to find flares, supernovae, variable stars and moving objects. Other instruments can then investigate a detected event in detail. Wide survey fields and long focal lengths for fine detail are different engineering priorities.
The corrector adds the required path difference before the light converges
A spherical surface has the same radius of curvature throughout, making it convenient to work and compare against a reference sphere. Parallel rays falling on different zones of that mirror, however, cross the axis at different locations. The marginal bundle does not wait for the central one to focus. At the detector plane, a blur replaces a single small geometric intersection.
Optical path in glass depends on refractive index and traversed thickness. Refractive index compares light’s speed in vacuum with its speed in the material; the same geometric distance through glass adds a different phase from that through air. If a weak aspheric element varies the thickness with radius, it changes the relative phase of different pupil regions and slightly changes their directions. Its profile is selected for the mirror system. This is deliberate compensation, rather than an arbitrary extra lens to “boost” a mirror.
The corrector reduces spherical error without making this example perfect
The figure compares exact spherical reflection with our model of a very thin phase corrector. It removes the leading error while leaving a smaller residual. The calculation omits finite glass thickness, dispersion, coatings and diffraction. The illustrated instrument is not a Palomar prescription.
Branch · Optical pathOptional · may be skipped
Deeper: why the fourth power of radius appears
Take a hypothetical sphere of radius R=1000 mm, with its vertex at x=0 and centre of curvature/corrector plane at x=R. Parallel light travels towards decreasing x; the paraxial focus is at x=R/2. Paraxial means an approximation for rays near the axis with small angles. Let r denote distance from the axis. The geometric path from the corrector plane through a fixed point on the sphere to the paraxial focus has the leading departure:
W_(sphere)(r) = − ((r)^(4))/(4(R)^(3)) + O (((r)^(6))/((R)^(5))).
This compares optical paths. Segments forced from every mirror point to the chosen focus are not yet actual reflected rays. That is why the model separately calculates the sphere intersection and reflection law.
A thin plate adds the opposite fourth-order term. For a hypothetical constant n=1.5 and thickness change relative to the axis:
W_(plate) = (n − 1)Δt, Δt(r) = ((r)^(4))/(4(n − 1)(R)^(3)).
At r=100 mm this gives 0.05 mm of thickness change and 0.025 mm of extra optical path. A profile small compared with the diameter can still have a large effect compared with the wavelength.
At r=100 mm the transverse ray residual in the x=500 mm plane is approximately −0.5115 mm without the corrector and −0.01154 mm with it. At r=50 mm the residuals are approximately −0.06285 and −0.000354 mm. These values do not promise a zero-size real star: fourth-order correction leaves higher orders. Halving the aperture reduces the leading cubic ray error by about 8, and the corrected fifth-order residual by about 32. This checks error order independently of matching a single picture.
A quadratic profile term changes the chosen focus and zonal balance. The pure fourth power above therefore does not describe every practical plate. The index n also depends on wavelength: a corrector can leave chromatic variation of spherical aberration, despite the mirrors providing the main focusing. Schmidt himself discusses this limitation.
Plate shape, stop position and curved field are three different things
A technical illustration must exaggerate the corrector’s thickness variation. Otherwise its weak asphericity can be mistaken for a flat window. A profile with equal centre and edge thickness can have a thinner annular zone between them. The exact shape depends on the optical solution; a plate from another instrument cannot be selected by outer diameter alone.
A weak profile is enlarged so it can be seen
In the classical camera the stop and corrector lie near the spherical mirror’s centre of curvature. This helps make off-axis bundles symmetric relative to the sphere; the corrector separately compensates spherical aberration. Moving these parts changes the system. Not every telescope with a sphere and front glass inherits a Schmidt camera’s wide field. The original geometric explanation.
Even a well-corrected classical camera forms its best images on a curved surface. Historically, photographic material was bent to follow it. A flat electronic sensor does not follow that sphere: a sharp centre need not mean sharp corners. Special optics, suitable detector geometry or an acceptable field restriction is required.
The Zwicky Transient Facility on Samuel Oschin provides a real example. Its digital camera is more than a large CCD substituted for film. The technical publication describes a cryostat window, local field-flattening lenses and an additional aspheric trim plate compensating the changes. A camera window thus also belongs to the optical problem. Dekany et al.,2020, section III.
The Schmidt–Cassegrain folds the path and puts the focus behind the primary
In a conventional SCT, light passes the front corrector, reflects from the primary to the convex secondary, then returns through the primary’s hole to the eyepiece or camera. The secondary intercepts a still-converging bundle, changes its convergence and displaces the focus. This mechanism is familiar from the Cassegrain article; here spherical surfaces work with a front corrector.
Three tube traversals and a focus behind the primary
The folded route suits a compact tube. Effective focal length, however, is not the sum of the drawn segments. It determines image scale for a small object angle. Our first-order diagram below has only 300 mm between mirror vertices but an effective focal length of 1600 mm. The full housing, cells, focuser and rear working space add mechanical dimensions.
The SCT drawing uses paraxial rays and vertex planes rather than exact intersections with spherical surfaces. It explains component order and path folding, rather than predicting residual aberration. The Schmidt-camera diagram above uses a different, explicitly stated modelling level.
Branch · First orderOptional · may be skipped
Deeper: how two mirrors produce a long effective focal length
Let the primary focal length be f₁=400 mm and vertex separation s=300 mm. Without the secondary, the light would travel another 100 mm to focus. The convex secondary has a hypothetical radius R₂=800/3 mm. In a calculation unfolded along propagation its focal length is f₂=−R₂/2. Its object is virtual, p=−100 mm:
(1)/(p) + (1)/(q) = (1)/(f_(2)), − (1)/(100) + (1)/(400) = − (2)/(800/3).
We obtain q=400 mm from secondary to image:300 mm back to the primary, then 100 mm behind it. The secondary magnification has magnitude m=4, so:
F = mf_(1) = 1600 mm.
Changing mirror separation changes the conjugates, focus location and F. Moving the primary can make the centre sharp again without proving restoration of the original field performance. This is our first-order worked example, not a Celestron prescription or a complete aberration calculation.
Correcting spherical aberration, coma and field curvature are separate tasks
A standard SCT can work well centrally while showing coma and out-of-focus stars near the edges of a large sensor. Its corrector alone does not make the field flat. Special derivatives add or change optics to improve a wider field. Manufacturer documentation: SCT and EdgeHD.
EdgeHD adds lenses inside the primary baffle. Their task includes off-axis images and the sensor plane. This differs from merely reshaping the secondary to remove coma: absence of coma does not itself remove field curvature. Comparisons need named designs, focal planes and fields. EdgeHD engineering paper, sections 3–4.
RASA addresses another task: fast digital imaging with a camera in front of the tube. It is not an SCT from which unscrewing the secondary automatically produces a corrected frame. Its front focusing group and camera have their own geometry, working distance and obstruction limits. This photographic instrument is not intended for eyepiece viewing. The camera body and cables also interact with incoming light. Celestron: RASA and camera position.
For an observation, this means choosing the task: a long image scale for small details, a wide survey field or dedicated photographic speed. Calling one telescope “best” hides those different requirements.
The asphere is made and tested, rather than simply cast as a glass lid
The materials and manufacturing article distinguishes substrate, surface and coating. A corrector also depends on homogeneity and light travelling through its bulk. A mirror’s reflective surface is tested; a corrector cannot be judged merely by the shine of one side.
One historical method elastically deforms a thin blank during working. Its surface changes shape when the load is released. Edgar Everhart described a vacuum method in a 1966 technical publication. Its available abstract explicitly notes breakage risks for excessively fast systems; it is not a home recipe for pressure and thickness. Everhart: publication and limits.
Another disclosed variant appears in Thomas Johnson’s patent. A master block is first made with the inverse contour, checked using a test piece and corrected. Vacuum holds the blank against that contour while the exposed side is worked; the blank is then released. The patent’s variant polishes that side flat. Contact accuracy and cleanliness affect the result, so a dust particle is more than a cosmetic blemish. US 3889431A, method description.
The 2013 EdgeHD paper describes a related production principle, but polishes the free side to a weak concave sphere. “Vacuum method” therefore does not mean one unchanging process. The corrector is subsequently tested in an optical bench with a double pass: light returns from a reference flat and traverses the optics again. Greater sensitivity helps expose errors; it does not mean every surface and wavefront shares one universal “fraction of a wave” specification. EdgeHD: fabrication and testing, pages 8–9.
Reflection and transmission require different coatings. Primary and secondary reflect; the corrector’s unwanted reflection is reduced. After coating and mounting, the set must be checked again: correct individual parts do not establish correct mutual orientation. The published principle does not disclose every current factory parameter, and a patent does not establish that its exact variant is still used today.
Cells, baffles and camera must agree with the optics
The corrector needs support and retention without unwanted stress. The secondary assembly must direct the light where required. Central and secondary baffles reject stray light, but an overly narrow passage can clip useful off-axis bundles. The primary hole and rear equipment position also belong to this geometry.
Optical parts need mechanical supports and defined distances
- Corrector in its cell
- Secondary and alignment holder
- Primary on a moving support
- Central baffle and exit hole
- Focus mechanism
- Camera and sensor plane
- Tube and rear cell
Many SCTs focus by moving the primary mirror. This accommodates different rear equipment, but the motion can cause slight image shift. Bringing central light to focus and placing the sensor at the specified mechanical working distance are separate checks. A field-corrected system needs its own spacing maintained.
For example, Celestron documentation dated 30 December 2025 specifies 133.35 mm for an unreduced EdgeHD 8, measured from the stated rear lock-ring surface to the sensor. Its T-adapter example combines 78.35 mm of adapter with 55 mm of agreed camera-plus-T-ring distance. Do not measure 55 mm from the front corrector or an arbitrary tube end. Other models and reducers change both value and reference surface. Celestron: spacing tables and worked example.
Branch · Field geometryOptional · may be skipped
Deeper: why centre and edge are not always sharp together
For a simple spherical focal surface of radius R_f, its axial departure at height y from the tangent plane is:
Δz = R_(f) − √((R_(f))^(2) − (y)^(2)) = ((y)^(2))/(R_(f) + √((R_(f))^(2) − (y)^(2))).
The second form is more convenient numerically near the axis. The technical account of Samuel Oschin’s original optics gives R_f≈3062 mm. From the centre to the middle of a 355 mm plate edge, y≈177.5 mm, giving a sag of about 5.15 mm. This explains why a large flat detector is not a direct replacement for curved film. It is not a model of the modern corrected ZTF. Technical description of the original geometry.
For a local cone of focal ratio N, the geometric blur diameter from a small axial defocus is approximately:
c ≈ (∣Δz∣)/(N).
With hypothetical N=10 and Δz=0.5 mm, c≈0.05 mm, or 50 µm. This includes neither diffraction, coma, astigmatism nor an actual field-corrector model. It does not mean that a 0.5 mm spacing error in every SCT always produces that circle: refocusing changes the whole system, rather than merely moving a sensor within a fixed cone.
Sensor spacing and defocus require different checks
Dew can form on the front corrector. Humidity and the temperatures of the optics and surroundings affect the outcome; dew protection must suit the instrument. The system also needs to approach its operating thermal state. If the image changes during cooling, altering collimation before it settles can be premature without separately checking the cause. Celestron: SCT operating limitations.
Correctness is checked across the required field
A useful check starts with the conditions: telescope design and model, camera, added optical elements, spacing reference, focus, field position and thermal state. Otherwise two images can show different conditions rather than different optical quality.
Compare the centre, several edges and corners. A symmetric sharpness change from the centre can point towards field or spacing issues; one bad side also requires checking sensor tilt and alignment. Changes after refocusing, reversing focuser travel or changing tube position supply further clues. A single blurred star shape does not uniquely diagnose every possible error.
Engineering acceptance separates surfaces, corrector transmission, matching of the optical set, mechanical stability and the final image. A computer model can test a formula or geometric law; it does not replace measurements of an actual unit.
A thin corrector lets a convenient spherical mirror substrate remain while changing how parts of the wave combine. Stop, secondary, additional lenses and sensor locations then determine the instrument: a wide camera, compact SCT or dedicated astrograph. These relationships explain its capabilities more reliably than similar-looking front glass.