A three-part sling
The support of the back of a telescope mirror has been subject to much study, and many good designs exist. The finite element analysis and optimizing program PLOP (by David Lewis) is an incredibly powerful tool for designing a cell to measure. However, at 45 degrees of altitude, the edge of the mirror needs to be supported by the same force as the back, and a poor design of an edge support can cause considerable astigmatism - particularly with large, thin mirrors. For some theory behind this, see my sling page.
An altaz mount has the advantage that there is one defined "down" side to the mirror cell, simplifying the design ("up" and "down" etc will refer to the mirror cell as seen with the telescope at zero altitude). One simple solution is to support the mirror edge at two points, at + - 45 deg from the vertical. This placement means that low-ordered astigmatism by out-of-plane forces will be cancelled (even if other defomations are possible).
The other traditional support is the 180-deg sling supporting the lower half of the mirror edge. Theoretically, a well-designed sling is close to the ideal edge support - perhaps somewhat surprising, as you would expect a thin mirror to yield to "potato-chipping", deformation by its own weight causing astigmatism. However, the force from the sling in the radial direction (towards the center of the mirror) is constant along the whole lower half of the mirror edge, and the side-to-side pressure would largely cancel the "potato-chipping" and the astigmatism.
Still, I think there are possibly serious but usually ignored potential problems:
The sling supports should of course not press against the mirror edge - this means that there must be a small part of the sling that is free between the support posts. Thus, the mirror can swing a little side-to-side causing some possible miscollimation, but this problem may not be so severe.
Another problem is that the commonly used "seatbelt" type of wide sling will not have its forces applied at a well-defined level. If applied at the plane of the center of gravity, the support will be well balanced - I don't believe minor deviations from this ideal is a serious practical problem, either.
The third problem is when the sling "grips" the mirror edge (by friction) near the "sides" where it leaves the mirror. pulling the edge forward or back. This can cause noticeable astigmatism even with tiny misalignments, and is, I believe, the most serious of these problems.
The 3-part sling
So, here is a sling design that I believe can keep the potential problems (i.e. astigmatism) at bay, even for large, thin mirrors. Some practical details are illustrated by my cell for a 13.1" diameter, 1" edge thickness f/4.5 mirror, (this is likely "overkill" for such a relatively small mirror). Anyone who would try to make a cell using the same concept will no doubt find other and better ways of implementing it.
My mirror cell is built with a square frame of 25 mm aluminium square tubing, with the mirror actually within this frame (there are reinforcing corners at the back - also at the front of the frame, although these are removed on the images). The 6-point back support has 3 pivoting bars (T-profile) on a triangle of 20 mm square tubing - the lower one is fixed, the other two are free at the upper end where the collimating bolts are (allowing me to adjust the collimation from the "right" side, with a wrench on a piece of plastic tubing).
Instead of the conventional wide sling, I use a wire (bicycle brake wire - it is quite strong and does not stretch!), supported near the correct level of the COG (see the sling page). To make the design as insensitive to "edge pulling" as possible in case of minor misalignment, I want to use a fairly long free piece of sling between the support and the mirror edge - a small difference of level (as e.g. caused by collimation) will then cause minimum "pulling". This would of course make the design susceptible to sideways swinging - so I divide the 180 deg sling into 2 parts, each covering 90 of the original 180 degrees. This arrangement is perfectly stable against "swinging". However, during transport, the mirror can move up and out of the sling - to prevent this, I use an "upside-down" 180 deg wire sling, spring-loaded with some small force, but after one or two millimeters of stretching the spring, the sling will "bottom" and prevent the mirror from moving more than a millimeter in any direction (in its own plane).
Thus we have 3 slings - one going from the upper left corner of the mirror cell passing below the mirror to the lower right corner, the second is a "mirror image" of the first, going from the upper right to the lower left corner. and the third from the lower left to the lower right corners in an inverted "U" fashion. In the photo above, you see the cable slings, with fastening screws near the corners (for the inverted "U" part, see later). Where the slings touch the mirror, they are covered with the original Teflon sleeving (the outer metal sheath was removed!). The lower left corner magnified to show the screws with washers holding the cable ends (at approximately the correct level)
One end of the inverted U part is held by a screw as the others are, the other end is spring-loaded as shown below. The cable passes through a slit in the square profile, and at its end is the end stop and a washer. Between the washer and the profile is a conical spring - if the wire is pulled, the spring will be fully compressed after some 2-3 mm. This is enough to allow prying off the slings to remove the mirror when desired, but the tension of the spring keeps the slings in place during use and transport.
When the telescope is pointed at the horizon (something it will never be during actual use!), it will be suspended only by its sling, with no force pressing it against the back supports, and it is of course necessary to use some kind of retaining clamps to keep it from accidentally falling out of the cell.
I believe that ideally, the sling should be able to move freely against the mirror edge, thus preventing it from gripping the edge. The Teflon sheath is slightly bent and will not roll - perhaps if I had cut it up in many short pieces, it would - or some other beads, or even bearings, would work.
If you want to better define the level of the slings, perhaps 4 clamps like this, placed at the ends of the "diagonal diameters", would work well.
A final word - this concept is still experimental, and anyone else will find personal solutions for the construction details - I would of course appreciate hearing of your experiences, if you try something along these lines. And I'd like to hear of omissions and errors that you may see.
Nils Olof Carlin - first (and very preliminary) version, July 4th, 2004.