RTV Considered Harmful (was [ATM] Re: Second Thoughts onMirrorSupport)

[jlerch1 at tampabay.rr.com (James Lerch)]

----- Original Message ----- 
From: "Mel Bartels" <mbartels@bbastrodesigns.com>


> > Can you elaborate? How thin and how large? How was the wavefront measured?
>
> 16 inch, 20 inch, and 24 inch mirrors from 1 1/4 inch center thickness to 2
> inches center thickness.  Weight applied via sticky pad to locations on the
> back of the mirror while near zenith, with star test to measure results.

Perhaps its time I throw another recent example into the mix, a plate glass 22"
3/4" edge thick F/5.6 RTV'd to an 18 point whiffle tree plate glass cell, as
seen in this image:

http://lerch.no-ip.com/atm/Projects/Rons_Mirror_Cell.jpg

Our current supposition as why this failed so miserably is precisely related to
the image below, right hand side:

http://www.surfnetusa.com/sho/TeleGraphics/Support.jpg

I'll quickly elaborate on the method used to evaluate the "Failed so miserable"
statement.  With the optic, whiffle tree, and wood cell propped up on our test
stand (supported only by the wood ring), Interferometry was impossible at best.
The gross amount of error present on the surface (irregardless of orientation)
made creating a reasonable fringe image nearly impractical.  In an attempt to
reduce this fringe data, it looked like nearly 19 waves of primary astigmatism,
mixed in with many other odd errors.  Rotation of the unit (Optic, Cell,
Support) didn't change the orientation of the astigmatism.  However due to the
nature of the incredibly poor fringe images to work with, quantitive analysis
was impractical and skeptical at best.

Putting the same unit (Optic and Cell) in the OTA, and attempting to focus stars
demonstrated astigmatic images like none ever seen!  (inside focus looked like a
line with a bulge in the middle, as focus was approached, the line got shorter
and the bulge bigger, once past 'focus' the line re-formed at 90 degrees to its
original location.  In a nutshell, it had really gross astigmatism.

Fortunately all is not loss, being the 'never give up' type of guys Ron and I
are, we cut the mirror loose from the cell, replaced the RTV with Felt pads, and
implemented a sling.  This helped enormously!

However, one benefit of such an incredibly thin optic is it easily demonstrates
the interactions between the cell and the optic.

With the sling implemented, we observed something important about slings.  (NOTE
the following is supposition, with limited facts other than observations to
support it!)

The sling we used contacted the edge of the mirror at 3 or more points.  Classic
kinematics tells us that 3 points define a plane.  If the plane defined by the
sling, isn't coplanar (or at least parallel) to the plane established by the
whiffle tree support, Really interesting things happen!!!!  (like induces
astigmatism, twisting of the optic, and jus hard to describe stuff!  (BTW, we
did all this testing with the interferometer at about 10 degrees elevation)

One interesting observation was the appearance that the sling didn't cause any
problems from trying "Crush" the optic by squeezing its edge under compression.

So anyway, after much fiddling with sling position, and one other "off the wall"
idea, we ended up attaching springs to the back of the optic (4 at 90 degrees
locations around the outer perimeter, and one at the center.)  By judicious use
of tensioning the springs, we could get the optic in some semblance of a
parabloid (as in a strehl of .5 area on interferometry)

To date, the optic is coated, and as far as light buckets go, it works
surprising well, but you can forget about 600x power on the planets :0

Current work is being done on improving the sling implementation.  Currently its
our belief that a 'perfect' sling would have 4 degrees of freedom, only
constraining the mirror with respect to keeping it centered on the cell.  Of
course fabricating such a sling is left as a future exercise!

BTW, the scope will be available for viewing at our local Orange Blossom Star
party in Florida this month, if anyone is in the area.

Since this is starting to appear as a novel, I do have one possible 'Idea" to
'fix' (or perhaps reduce) the original posters concern as seen in this image:
http://www.surfnetusa.com/sho/TeleGraphics/Support.jpg

On the left side of the wiffle tree support, where the eye-bolt swivel is
located.  If one were to extend this bolt somewhat further, then add three
counterweights of the appropriate weight and location.  It may be possible to
eliminate any need for edge support.  Of course figuring out how much each of
the three counter weights should weigh, and their positions is left as a future
project of exploration (BTW, is what I just described something similar to the
'astatic' mirror cells I've seen in the past?)

Well I guess this ends my foray into 'controversial' land, while hoping I
haven't made an utter fool of myself yet again :0

Take Care,
James Lerch