Re: [ATM] figuring 10 inch mirror

[Mark Holm <mdholm@telerama.com>]

Wayne Hilliard wrote:

I have posted Sixtests results in the following three files:

http://www.atmlist.net/contrib/mdholm-at-telerama-dot-com/W_Hillard_7-22-05_Sixtests_Para.jpg

http://www.atmlist.net/contrib/mdholm-at-telerama-dot-com/W_Hillard_7-22-05_Sixtests_Conic.jpg

http://www.atmlist.net/contrib/mdholm-at-telerama-dot-com/W_Hillard_7-22-05_Sixtests_Para_rms.jpg


The first file shows your Foucault results reduced to the best fit
parabola.  The second file shows them reduced to the best fit conic.
The third file shows the best fit parabola again, but with RMS weighting
applied to the residual errors.

The best fit conic shows that your overall shape is very nearly a
parabola, b = -0.993.  This is so close to -1 that the residual is
probably beneath your measurement noise.

The best fit parabola shows that your Strehl ratio is 0.873.

The RMS weighted parabola chart shows that the turned edge is by far the
most significant remaining contributor to optical imperfection.

I will note that there is a lot of uncertainty in any turned edge
analysis by Foucault.  You only have real data on the inside of the
edge.  Sixtests has done about as well as any numeric method can at
estimating the edge, but it is an estimate.  since the slope at the edge
is fairly steep, the uncertainty is pretty high.

I think the qualitative criteria for turned edge are more likely to be
trustworthy in many cases.

I really can't claim to be expert at figuring, especially fixng turned
edges.  I'll defer to others on the list for the best advice under the
circumstances.

If I were doing it myself, with no outside advice, I would try a full
size lap, well pressed, doing a mix of chordal strokes TOT and normal W
strokes TOT alternating with MOT.  The idea is to get more work on the
outer zones in order to get at the turned edge.  It seems paradoxical
that one needs more work on the outer zones to fix a turned edge.  You
might think it was too much work there that made it in the first place,
but lots of observations point to just the opposite.  If the correction
(b) moves more negative (overcorrected) I would use more TOT.  If it
goes more positive (undercorrected) I would use more MOT.  This method
is likely to be slow, but may let you keep the figure fairly near the
parabola as you reduce the turned edge.

A recent bit of interferometric work posted on this list indicates that
hand warmth might be more a factor in edge turning than most think.  Try
not to curl your hands over the back edge of the mirror if you can help
it.  If you have to, only curl them over as little as you can get away
with.  If you have handled your mirror by the edge while putting it woen
on the lap, let it equilibrate at least 20 minutes before you start
working.  You might consider cotton flocked diswashing gloves.  They cut
down heat transfer quite noticibly.

Others on the list have their own favorite methods for fixing turned
edge.  They are mostly much more experienced at it than I and probably a
lot more successful, so I suggest taking their advice over mine if there
is a conflict.

-- 
Mark Holm
mdholm@telerama.com

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