[ATM] Re: Second Thoughts on Mirror Support

[jonah at eecs.berkeley.edu (Jeff Anderson-Lee)]

Jean-Guy Moreau wrote:

>>That assumes that the two surfaces are stiff.  If they are not, one or
>>both will sag slightly until more and more points touch.  By using thin
>>metal over carpet, we are trying to make it the metal that sags more
>>than the mirror and let the carpet balance the forces.
>>    
>>
>
>
>At this point, it might be useful to relate some experimental
>results i had while investigating ways to make the fused glass
>mirrors i now do.
>
>...
>
>Success only came when i completely fused the joints, at high temperature
>in a kiln, essentially turning the whole thing into one continuous structure.
>Any partial, so called tack fusing of the joints, will led to failure.
>
>These mirrors are nothing more than a very thin meniscus plate, on an
>apparently good glass support, can we consider it this way ?
>the one i am working on now has a 26" by 9/16 " face plate, that's a
>46/1 aspect ratio ! it has no printrough or astigmatism issues while
>polishing/figuring with common thick glass techniques, nor does it display
>these problems on a simple test stand, or mounted on a sealed plywood
>"cell" with multi point RTV, in other words it behave as a full thickness
>mirror.
>
>Does this mean it is possible, with the right choice of materials, if they
>exist,
>to build a cell that will make practical the use of very thin glass ?
>  
>
I'm certain that a cell would work for your mirror if 81 posts were laid 
out as per Plop and a wiffle-tree was constructed with support under 
each post, leading to an 81 point cell for such large mirrors.

I don't know enough [yet] about static analysis though to determine how 
much stiffer than the face plate and back plate the resulting composite 
mirror is.  It is possible for example, that you could get away with a 
27 point cell for an 81 post mirror: one support point for each three 
posts.  The layout I have in mind can be found at the following URL:

http://www.cs.berkeley.edu/~jonah/18plus/images/p81va.html

(Click on any image to upload the .gr file)

If you wanted to use a conventional wiffle tree support, I'm not sure 
I'd go for less than 27 points.

On the other hand, the previously mentioned RTV + steel + carpet + 
plywood support might work too!  It would certainly be easy enough to 
try and test -- a lot easier that making the mirror in the first place 
and definitely easier than making an 81 point flotation cell.  Though if 
it comes to that, laser cut steel is probably the way to go to get the 
right tolerances at a moderate price.

Jeff Anderson-Lee
Sacaramento, CA