Jump to content




Search Articles


- - - - -

A 7inch F/3 Schmidt Camera


Voice your opinion about this subject in our forums

A 7inch F/3 Schmidt Camera

By David Cartwright

Cloudy Nights members may be interested in a Schmidt camera that I made twenty odd years ago. Although photographic equipment has moved on since then, and Schmidt’s original design will now appeal to few, the novel method of making the corrector plate might still be useful in making a Schmidt Cassegrain telescope or astrograph.

The idea of making the camera came during the machine polishing of a refractor lens where the polishing was concentrated on a zone 2/3 of the way out from the centre.

A Schmidt camera consists of a spherical mirror and a correcting lens placed at the radius of curvature of the mirror. The curved focal surface lies midway between these two components.

If a fast spherical mirror were used alone its images would be blurred by severe spherical aberration. Deforming that sphere into a paraboloid as in a Newtonian corrects the spherical aberration but provides a complete solution only for images at the centre of field, because only the axial beam of light can be symmetric with the paraboloid. All other images are more or less degraded by coma and astigmatism, increasing as the image moves away from the centre of field.

Bernhardt Schmidt’s brilliant contribution was to take the correcting deformation off the mirror and relocate it on an optical window placed at the mirror’s centre of curvature. This lens, being refractive, has a deformation opposite in shape and four times as deep as that of the mirror it replaced. Wave fronts from all objects in the camera’s field of view suffer virtually equal deformation as they pass through the lens. They all have equal symmetry with the mirror and each are brought to an almost perfect focus if the spherical aberration imparted by the lens exactly matches that of the spherical mirror. With increasing camera speed images become degraded by residual aberrations so that a useful limit is usually set somewhere between F/1.5 and F/1. However, in amateur sized cameras the performance at F/2 is excellent and at F/3 virtually perfect across a field of 10 degrees.

If the correcting deformation is placed on a plane surface, the lowest point of the Schmidt curve, the so called neutral zone, is at 70.7% of the radius of the disc. If the surface is slightly convex such that the neutral zone is found at 86.6% of the radius, then residual spherochromatism will be at a minimum although this refinement is hardly necessary in slower cameras.

Good accounts in ATMA and ATM3 are amongst the many references to Schmidt cameras and the aspherising of the correcting plate to be found in amateur literature. Rutten and Venrooij in ‘Telescope Optics’ describe the Schmidt as offering the combination of a fast focal ratio, large field angle and unparalleled image sharpness over the entire field.

Although the fact that Schmidts can’t be used visually has contributed to their scarcity, the greatest stumbling block to many would be builders has been the reputed difficulty of accurately figuring the corrector with its 4th order curve whose profile is convex around the centre but concave towards the edge.

The methods of using rubber backed ring and petal tools and the vacuum deformation technique have been well described but neither appealed to me. It seemed that that the approach that was so unsuitable for polishing the refractor lens might simply and automatically solve the problem of the 4th order paraboloid.

I had some 7” discs of 10mm float glass that had been intended for autocollimation flats and using the three plates method I ground and smoothed both sides of one disc until they were within a wave or so of plane. In testing for flatness prior to polishing and figuring, I used a colourless shoe polish with good results. After applying a small quantity and wiping off any excess and leaving aside to dry for a few minutes, I rubbed it vigorously with a brush until a glaze formed which was then buffed lightly with a soft cloth. When placed in contact with a flat interference fringes formed showing reasonable contrast. Irregularities in the form of a small scale ripple 3 millionths of an inch high could be seen. If the wax was too thick contrast was poor or the fringes didn’t show at all.

Just how faithfully this glaze followed the glass was shown in an experiment in which the ground Schmidt plate with both sides shoe polished was placed in contact with a 6” flat standing on its edge. A 5” achromat was, in turn, in contact with the plate and the whole combination tested with a knife edge. The illuminated slit was viewed with a x100 eyepiece and a bright crisp image seen; formed from light that had reflected off the 6” flat and that had travelled twice through the whole system. This showed, apart from the regularity of the surfaces, the quality of the float glass which was to form the corrector plate. Also seen to one side was a much dimmer though also well defined single image formed by reflection off the two waxed surfaces indicating wedge was virtually absent.

With the float glass established as a suitable material I decided to polish and figure in the one operation, see how far the asphericity could be taken, and finally make the mirror to match the correction on the lens. The technique consisted of a spindle rotating clockwise (I’m right-handed), as seen from above, at a brisk rate of 80rpm or so, and fixed to it an ordinary button lap of medium hard pitch. The facets of this full sized lap were formed by placing a punched rubber mat and netting over warmed soft pitch. It was important to ensure that the facets were not distributed symmetrically about the centre of revolution.

My first attempt was unsuccessful with the result that conspicuous zones showed up in the interference test. It was necessary to remake the lap by heating under a grill until fluid followed by faceting with the rubber stencil and netting. After cooling and cold pressing, I checked the lap by running it on the machine and assessing by eye whether the polishing action would be uniform.

The lens was held against the lap by the index finger and thumb of both hands (Fig 9a). A short diametral stroke was used, the lens being pushed forwards a half to one inch from a centre over centre position and returned to that position, the lens then being allowed to rotate a little before being pushed forwards once again. The lens was never allowed to continue beyond the centre over centre position on the return stroke. During this process most downwards pressure was exerted by the right hand, especially the index finger. It was possible to vary the position of the neutral zone by altering the relative force exerted by the right hand finger and thumb and the length of stroke. Generally, the shorter the stroke and the greater the pressure from the fore-finger, the further out the neutral zone. There was little pressure on the return stroke.

I had read Michael Harlow’s account of a Schmidt corrector in TM27 and followed the progress of the work with a 4” flat, a small neon indicator bulb and a 4” collimating magnifying glass. The asphericity deepened at the rate of 1 fringe every ten minutes until a maximum of 11 fringes was reached after which a further 2 or 3 hours polishing removed the remaining pits without affecting the figure. Both sides of the lens were treated the same way giving a total asphericity of 22 fringes or 11 waves. It was possible to photograph the fringes using the same combination of neon bulb, magnifying glass and 50mm camera lens.

The deviation, D, from a plane of the surface of a Schmidt corrector lens is given by the approximation:

D=(x^4-k.r^2.x^2)/(4(n-1).R^3)

Where x is the radius of the zone (2.47), r the radius of the corrector plate (3.5ins), k a constant, equal to 1 where the deformation modifies a plane surface and where x equals .707r. ( k is 1.5 where x equals .866r). R the radius of curvature of the spherical mirror, and n the refractive index of corrector glass used (1.52).

D had been measured as 11 waves of light (.000262ins), and so R turns out to be approx 41 ins. The depth of lens correction varies with the inverse cube of the mirror’s R.O.C so shortening this from 41 to say 40ins would require a further one and a half fringes asphericity on the plate. I didn’t worry about using the exact refractive index for neon light as the idea was to get an approximate match between the corrector and mirror before final figuring of the plate.

The profile was good except that the neutral zone was 6mm inside the 70.7% position and the centre seemed a fringe or two high. The corrector could perhaps have been left in that condition, the remaining aberration probably being masked by the limited resolution of the emulsion. In Sidgwick’s ‘Amateur astronomer’s handbook’ details are given of maximum permissible speeds for different apertures of spherical mirrors employed without corrector lenses. Taking the example of a 4” F/4 mirror and a permissible image size of .05mm, the fact that the superfluous lens would have had a deviation of nearly 3 waves gives an idea of possible tolerances.

However, .05mm is a mediocre ambition and so to further improve the figure a polythene mask was cold-pressed onto the lap leaving 6 segments in contact with the glass (Fig 9b, button facets not shown). After some further polishing the correct neutral zone position was achieved.

The Mirror

I had a 10” pyrex mirror disc ground to a radius of 60” intended for a Cassegrain instrument. The curve was further deepened with the original 6” tool until the required 41” was nearly reached. The smoothing stages were carried out with this same tool. Because its area was little more than a third that of the mirror many wets were need before pits of the preceding grade were completely removed. It was a mistake not to have made up a 9” tool. Another error was revealed later during polishing when I found out that a turned edge had been ground into the mirror. This was due to the small tool being worked over the edge of the mirror in an attempt to keep its radius sufficiently long. Although prolonged fine grinding was tried the problem was never solved, but the tde was restricted to a 1/8” zone, subsequently covered by a mask.

The mirror was polished by hand with a 9” button lap poured onto a resin base which conformed to the curvature of the mirror. The resin in turn was supported by a water proofed wooden disc. The mirror’s surface was checked by knife edge and figured by extending the stroke a little until a good null was achieved. An F/2 mirror is about the limit that can be tested by a Foucault setup. It was essential to use a small prism after the slit to ensure that outgoing and returning rays were nearly co-axial. I found that I could just view the whole mirror by squinting close to the knife edge and only burnt my forehead a couple of times on the hot bulb holder. The null test was backed up by substituting a 25mm Kellner eyepiece and viewing the slit.

With the mirror finished it was time to check the corrector. For this I made a rough jig from ply and dowel and placed a small torch with attached slit at the focus of the mirror (Fig 5). Between the torch and slit tissue paper acted as a diffuse light source. Just beyond the corrector a 4” refractor was used to form an image which was tested by knife edge or eyepiece. If the lens had been properly made issuing light would be perfectly parallel as if it were light from a star. Residual spherical and zonal aberrations can be measured in terms of wavelengths by reckoning a ¼ wave error at the focus to be equal to 2.25 mm k.e travel in an F/16 system. The mirror without the lens would suffer spherical aberration equal to the required asphericity of the corrector lens divided by two because of refraction. This being 51/2 waves amounts to 48mm k.e travel at the focus of the refractor! I measured over 40mm travel when testing that outer part of the 7” aperture not covered by the torch and slit. With the corrector in place this reduced to 6mm which is still equivalent to 2/3 wave error at the focus or twice that on the lens. Substituting an eyepiece for the knife edge I found the image to be far from sharp.

To fine tune the optics I made a new lap from medium soft pitch on a 7” disc of inner tube rubber which in turn was glued to a glass disc (Fig 9c) The knife edge had shown the figure to be smooth but undercorrected and so I deformed the lap with a petal mask such that the polishing action would be greatest at the edge of the lap (not at the .707 radius) reducing to zero at its centre. This mask was cut from a bin liner. The corrector was figured at 20 rpm for 7 minutes and, as luck would have it, this extra spell of work cured the last errors, the knife edge showing a null and the x60 eyepiece a crisp view of the slit and tissue paper. A slightly raised zone close to the centre was shown by the knife edge and can be seen in the interferogram but, being hidden by the film holder, I left it.

This method of aspherising is well suited to amateurs equipped with a turntable, preferably adjustable for speed. Compared with the making of a Newtonian or refractor it required only ordinary care and time spent, but no greater skill. Anyone who succeeds with the mirror will certainly be able to tackle the lens. It encourages a smooth, zone free figure if the lap has been well made and an excellent edge without turn. Through fine control of figuring, the method lends itself to producing high grade visual systems.

However, the physical effort was considerable and might be a problem in larger sizes. An 8” would be fine but 10” is perhaps the limit.

The neutral zone tends to be restricted to .707 radius, which is a drawback if .866 is wanted to reduce spherochromatism. Some way of reducing any finger heating effects might be worthwhile, but I didn’t bother and there seemed to be no ill effects.

As far as figuring with an ordinary lap is concerned, the lens described represents the limit, but with a flexible base to the lap, the lens could be aspherised a great deal more. An experiment with fine abrasive and lead instead of pitch facets on rubber resulted in an aspheric curve 35 fringes deep which, if on both sides, would be fine for an F/2 system.

An advantage of float glass, apart from its price, is that it’s likely to be free from wedge. However, I should be wary of thinking one can escape without grinding both sides before polishing and figuring. I once bought 10 7” discs of 10mm float glass. Not one of the 20 sides could furnish a 4” surface good enough for autocollimation.

The glass I used was noticeably green, due to absorbed light, and so the camera’s effective speed was correspondingly reduced.

Final null testing relies on a telescope of known high quality, with an aperture at least half that of the Schmidt plate. I used a refractor but, no doubt, a Newtonian would work as well, although I haven’t tried it. Early progress can be followed with the help of a flat. Grind, smooth and semi polish one side of a 10mm float glass disc using the three plates method and the result will probably be fine for this purpose without any figuring. Small, high efficiency, fluorescent light bulbs could be tried as the interference light source.

Tube assembly

The tube, mirror and corrector cells and film holder were all made of fibreglass. Fairly crudely so I won’t go into too much detail. The film holder spider was also made from fibreglass with brass vanes, and was adjustable with four pairs of push-pull screws.

Collimation was straightforward. Firstly, the corrector was squared onto the mirror by centring my own reflection, after which its cell was screwed to the tube.

Secondly, in order to collimate the mirror I marked a cross at the centre of the corrector’s 1st surface and adjusted the mirror’s three supporting screws until the cross’s reflection was positioned immediately behind it but on the 2nd surface. The float disc hadn’t been supplied exactly circular so this was rough, but it seemed to work out O.K.

There was no need to exactly centre the curved surface of the film holder other than by ordinary measurement but in order to collimate it, i.e. make it concentric with the mirror, I attached some graph paper in lieu of film. This was lit up and viewed across opposing diameters with a refractor focused at infinity. The spider was adjusted via push-pull screws until all points on the graph paper were equally in focus. A screw stop could be adjusted to allow for changes in temperature. Next time I’d use invar or build temperature compensation into the film holder assembly. I’ll try low density polythene; it has a high coefficient of expansion.

All of which worked fairly well as did fixing the individual pieces of film to the holder with double sided tape, but it was a fiddle. My mount was hopelessly inadequate and so I never got beyond recording star trails. The finest trails were under 10microns across on 120 tri x film and I could see no difference between the centre and edge of the negative.

The classical Schmidt camera is capable of superb imagery but is also unmatched in its impracticality, due to its inaccessible curved focal surface. A next step up might be to make a flat field, Cassegrain version such as the Slevogt described in ‘Telescope Optics’ by Rutten and van Venrooij. Whilst there is the second convex mirror to be made, it, like the primary, is spherical which eases collimation.

However, I’ve been reading a thread by Kona ‘Large Format Astrophotography’ first post 10.22.13 in the film forum, featuring 4 by 5 sheet film. Maybe I shouldn’t give up on this awkward Schmidt so soon. So I plan to keep most of the present OTA but redo the film holder and assembly. Make it bigger; 90 mm diameter and have a temperature compensating mechanism to fix the focal position. Then think about an equatorial.

These astrographs aren’t a great choice for a first effort but amateurs with a little experience have every chance of producing an instrument whose optics are up there with the best.

David Cartwright


  • wh48gs, Nightfly, brian304 and 1 other like this


0 Comments



Cloudy Nights LLC
Cloudy Nights Sponsor: Astronomics