In my reseach I came over a document that based out the old 6;1 factor, based out from a 6” in diameter and 1” thick edge for a plano mirror.
Take this 6;1 in a 18” it means a 3” thick edge, but we can’t use that thickness of course but if we take a 12;1 factor it means a 1.5” edge thickness but it will be -36X less stiff than the std factors at 6” in diameter and 1” thick mirror.
Ok a stiffer glass material can reduce it some, but not much.
-Now it’s said we do has very sophisticated mirror cell’s that take care of this.
I want to bring up some thought's around that.
Normally for a 18” it’s a Plop based 18 pt cell design and a sling for the edge support, or a wiffle tree at now recommended roller bearings to edge COG.
By the use of Plop’s automatic cell design it will easy become under the so call ’golden’ number >5 nm RMS even in thinner edge thickness as 12;1 or less and in Pyrex glass.
This is back of a mirror, but at tilted ( z88 ) say at 45 degree a 2 nm on the back - will be 3 nm.
Cruxis program tell on edge support readings.
If we test a normal 18” there even a 2 pt will do way under 5 nm and the wiffle tree is little better and wire sling show the the best readings.
But we are seldom or never observing totally on the edge anyway.
Anyway, it's easy get under the ’golden’ rule here to with a simple design.
-But why the golden >5 nm RMS ?
According from ATM we need a tolerance on the optic's at diffraction limit or rather twice the diffraction ( or better ).
This is a matter of size of course but say doubble diffraction and that is 1/8 PV wf or 10 nm RMS on the surface.
This is a Very tuff tolerance for a 15;1 at 18” and ex f/4 plano mirror.
But the Plop ’golden’ is a recommendation even twice that number to be ’safe’ so that's why the 5 nm.
As I been visit both Zygo and QED etc, and they told me on a 500 mm Newton optics to be at 10 nm RMS, and even on a 6;1 factor thickness it’s a very very tight tolerance ( $$$ ) and at a f/3-4 it's over 40K USD and they need to has the parabola to 1 wave before they start.
Smothness under 10Å is one thing, but the parabola is very tuff ( now we talk MRF, ION works top notch )
Now I will come to my point ;
How much do the glass move in metric number ?
Well, from a test I heard one dropped down 1 lbs ( 0.45 kg ) and that was creating 1/4 wave or 500 nm in deform.
So 1 gram is 1 nm.
1 nm is 1 micron = 0.001 mm ( or 0.0002” )
( if I got this correct )
So what deform can the eye visually detect ?
Say 5 nm ?
If so 5 nm it is 0.005 mm.
A 18” use a 18 pt cell has a wiffle arm at a CC ( triangle/triangle ) at say 145 mm.
That mean in reality each triangle will change half of that, so 0.0025 mm on each.
Now we look at the mechanically aspects.
We forget about sag in wiffle bar and triangle here now and focus on the movement on the wiffle bar.
We has a bearing in centre or in most cases the bar is on top of a 6 hex nut.
To move this 0.0025 mm at each side it’s really nothing in the centre.
And if use bearing, it might be dust caps, drag motion in bearing, environment dust, and other stiction.
From my aspects, it do sounds incredible that this very simple technology can work whit this very tight tolerances on this rather weak plano mirrors that hurt from beside the cool-off issue, it’s sag and gravity.
Add this up in this construction the sag of the moving parts at tilt position and load vs mechanically aspects in the real world.