# Is a mirror cell only necessary for collimation?

I'm building a truss Newtonian, and I'm planning on achieving collimation using turnbuckle-style trusses, as demonstrated in this thread. If I'm collimating this way, is there any reason to support the mirror on a mirror cell assembly instead of just rigidly fixing it to the bottom of the tube - i.e does the cell serve any purpose beyond being a way to angle the mirror? Obviously, it's a lot easier (and cheaper) to just mount the mirror directly to the tube; not to mention that it eliminates the sagging patterns that cell designs tend to suffer from, allowing the mirror to be thinner and lighter. Is there another purpose that the cell serves that I am missing - like, say, "protecting the mirror from whatever" - or can I safely discard it?

The primary mirror needs to be on a support (a mirror cell) that lets you adjust the tilt of the primary mirror.

The turnbuckles you are describing change the tilt of the secondary mirror. It does not guarantee that the primary mirror is pointing in the correct direction.

In other words, the primary and secondary mirrors need to be collimated and therefore need some means of adjusting the tilt.

(Edit) I took a closer look at this based on comments. The collimation can be done using the 6 trusses to create a "Stewart Platform". The left-to-right alignment of the secondary with the axis of the primary can be done, and the tilt of the secondary to the axis of the primary can be done. One thing to keep in mind is that the primary may need to be aligned "accurately" to begin with so that the optical axis is close to the center of the "tube", or the trusses need to have enough extension capability to be able to translate the secondary off to the side if the optical axis is not near the center. For example, if the primary tilt is off by 1 degree, over a distance of 60 inches (1.5 m) the secondary would need to be offset to the side by 1 inch (25 mm).

A minor benefit of a mirror cell is to allow better circulation of the air around the primary. This may not be as much of a factor in a truss "tube" scope compared to a regular newtonian with a solid tube.

• It's surprising, but that's actually not the case: the use of 6 adjustable truss members means that they form a fully functional Stewart Platform - and so are perfectly capable of translational motion in 3 axis, in addition to the 3 degrees of rotation you mentioned. Here's a cool demonstration. So, the truss can translate the secondary into the optical path of a fixed primary, then it can rotate the secondary to be correctly aligned. They're amazingly useful for such a simple mechanism.
– T.S
Commented Jan 29, 2022 at 23:14
• @T.S though it might be hard to actually do so quickly, by one person, in the field, at night, shivering and in a hurry to see something before the clouds set in. Having separate adjustments might be easier for a person, even though a computer actuated system could do it with just the hexapod.
– uhoh
Commented Jan 29, 2022 at 23:45
• That's true - but there is the upside that you can do it while looking through the eyepiece; even if the movements are a bit more unintuitive at first, you've got direct visual feedback. It makes me think of those experiments people have done inverting car steering; it's amazing how quickly these things become muscle memory. Besides, it's already going to be quite a computerized scope, so that's very much in the realm of possibility. With a robust enough computer control system, you could even eliminate the focuser and just use Z-axis translation to control focus. Food for thought indeed.
– T.S
Commented Jan 29, 2022 at 23:55
• @T.S yes I see just what you mean. Okay I'm sold!
– uhoh
Commented Jan 30, 2022 at 0:13
• @T.S Good points and video. I hope you will post an answer, too. One question I have is how accurate would the primary need to be installed versus the adjustment available in the truss. I imagine most turnbuckle systems have less adjustment than the hydraulic system from the video. Commented Jan 30, 2022 at 1:35

The mirror needs a cell that supports it in a way that maintains its shape (especially less than 6:1 diameter:thickness). It is best to think of the mirror as a sheet of jello that you need to hold in a very particular way so it slumps into the shape you want. see PLOP https://duckduckgo.com/?q=plop+finite+element+analysis&t=ffsb&ia=web

That the "wiffle-tree" support mechanism lends it self to tilting the primary as a means for colliminating is ... well ... secondary.