[ATM] PLOP and the mirror support

["lou krajci" <loukrajci@comcast.net>]

>From: "Jerry Reddell" <jerry.p.reddell63@sbcglobal.net>

>I have been playing with it to determine how best to modify a commercial 10
>inch mirror cell (all cast aluminum) to improve its support.. While
>experimenting with different support rings, number of supports, and radius
>of support rings I found several surprises. The most notable being the
>dramatic negative affect on the error by adding supports in the area where
>the most droop ( i want to say sag) occurs for a near ideal support. The
>intuitive thought is that the added support would further reduce the error,
>but instead the error increase by a factor of 10 or more.....

One approach with PLOP is to use the optimize feature on only one ring at a
time.  Start with a simple support arrangement (one or two rings of
supports)...optimize it...then 'lock' the position of these rings and add a
third ring at the radius you estimate is good, but leave this ring free to
change radius so PLOP can optimize it...and let PLOP do it's work to find
the optimum solution.  Repeat as needed as you add more rings.

Or another approach is to pick from one of the many available online
examples...which are often good starting points.
See:  http://www.cs.berkeley.edu/~jonah/18plus/ and scroll to bottom of page
to choose from 18, 21, 27...36...54...81 support points, etc.

How thick is your 10 inch mirror?  Mine is 1.3 inches thick...and rests on a
four point support.  PLOP and Richard Schwartz (the voice of ATM authority)
say four points are enough...and actual use/star testing confirm that.

>The analysis obviously uses 'point' support for calculations. The size of
>these PLOP support points is very small compared to the size of the actual
>built support area...Since the as build support area is in affect many PLOP
>support points arranged in a group, and considering the above mentioned
>unexpected increase in error by adding support points in obvious locations,
>the question comes to mind that the as built cell and supports may
adversely
>perform quite differently by increasing the error above what one expects
>using PLOP...

>Does anyone know of a way to get PLOP to simulate the actual as built
>support area?

See:
http://overton.tamu.edu/aset/krajci/
specifically
http://overton.tamu.edu/aset/krajci/cell-wide-pad/cell-wide-pad.htm

My finding is that wide pads don't help much in most amateur cases.  Better
to add more support points.

>Also when one builds a floatation type cell, is the mirror siliconed to
each
>of  the support points or is the mirror allowed to 'slip and slide' a bit
>with the restraints being elsewhere.?

Either method will work.  So far I have preferred silicone glue methods.

Notice that you say 'slip and slide' as a possibility in your support
scheme.  Slipping implies friction/stiction...which is bad for a mirror
support scheme.  Try and eliminate stiction as much as possible in your
design and construction.

I hope this helps.

Tom Krajci
Albuquerque, New Mexico
---
Outgoing mail is certified Virus Free.
Checked by AVG anti-virus system (http://www.grisoft.com).
Version: 6.0.788 / Virus Database: 533 - Release Date: 11/1/2004

_______________________________________________
ATM mailing list http://www.atmlist.net/