RE: [ATM] (Long) How good is good enough? (was how muchcelldeformation is too much?)

["Scott Milligan" <starzkey@charter.net>]

Thanks a bunch Mike.  As usual, your posts are concise yet full of insight.
I was not aware of the two papers that you cited, but am sure that I will
enjoy reading them. It was especially revealing that one could calculate the
overall probability of achieving a moment of seeing at a given rms wavefront
error as a function of telescope aperture and the r0 parameter; it brought
to mind an empirical observation made by Ron Dantowitz in a talk he gave at
NEAF 2003, in which he explained how the number of video-rate frames needed
to capture a moment of near diffraction-limited seeing increased in a highly
non-linear fashion as the telescope aperture increases.  Perhaps if he is
lurking on this list he could comment further regarding whether there is
good correlation between your probability calculation and his experience in
the field.

My own experience in connecting bench test metrics to what is seen in the
field is not universal, in that it is mostly restricted to apertures < 13
inches, and the admittedly compromised atmospheric seeing of New England.
However, within these limits, I do think that I have gained a pretty good
feel for what an optic should look like on the bench in order to take full
advantage of those few nights from which memories are made.  

Since my ATM interest lead to an early decision to pursue precision optics
as a career, most of my bench testing has been done interferometrically, and
so the notes which follow don't contain much information that was gained via
the knife edge.

My first scope was a 10" F/5.6 newtonian, which tested on the bench at about
1/10 wave P-V (@ 633 nm), around 0.025 waves RMS.  These are surface, not
wavefront numbers.  The surface was smooth out to the edge, with no zones
greater than about 1/16 wave in amplitude. Doing the math, I calculate a
Strehl ratio of 0.87 @ 550 nm.  The majority of the residual error could be
modeled as trefoil, the mirror was a full thickness Pyrex substrate, and the
edge support in the test stand was simple on two pegs spaced 90 degrees
apart.  I did not make any attempt to separate errors due to self weight
deflection from the test data.

With this mirror installed, and a secondary giving a 25% obscuration (big,
tall focuser), this telescope was a favorite back in the mid 1980's, for it
consistently yielded high contrast images and detail on planetary objects
that few other scopes present could improve upon.  One summer night in
particular was memorable, when the atmosphere seemed to "get out of the way"
and a few of us viewed Saturn at 700 x without image breakdown. (Maybe this
is no big deal if you live in the keys or some places out west, but here in
New England I have not seen another night to match this one in over 20 years
of trying!)  Oh, I should probably add that although star tests of this
instrument were sensibly perfect, one could detect a hint of a turned up
edge on steady nights at magnifications > 200-300 X.  

A second small Newtonian was made on the knife edge, and subsequently
checked interferometrically. Details are: F/4.34, full thickness pyrex,
overall figure again smooth, but a turned edge extended in wards about 1/8
inch.  Overall figure accuracy on the interferometer (masking off the edge)
was between 1/10 and 1/12 wave P-V, and I do not remember the RMS value, but
since it was a smooth surface, I expect that it was in the range 0.025 -
0.030 waves or thereabouts.  An interesting detail about this mirror is that
the first coating attempt failed, and required stripping which was done
using hot sodium hydroxide solution (I now recommend not using this method,
but that is a subject for a different post).  After recoat, testing at ROC
using the knife edge revealed that when the knife was positioned to
completely cut off the image, scattered light still managed to find a path
past the knife from the entire surface of the mirror, whereas prior to
stripping, I did not recall seeing this phenomenon.  In the telescope, using
a curved vane spider and a secondary obstruction of 25%, with the turned
edge mostly masked off (magic marker), images were sharp at all
magnifications up ot 250x, yet suffered from a noticeable drop in contrast
when compared against memories of the 10" F/5.6 described above.

A 13" F/10 Schupmann-type instrument was constructed and installed at
Stellafane in 1985, following the design worked out by Jim Daley and Bert
Willard, both long time members of the Springfield Telescope Makers.  A
detailed description of the optical fabrication effort may be found on the
www.stellafane.com website.  Here, it will suffice to say that the overall
wavefront error as reveled by an interferometric test was again on the order
of 1/4 - 1/5 wave P-V, with no aberration terms subtracted from the data
other than piston or tilt.  This instrument has yielded some of the finest,
high contrast views at high magnification that this writer has yet seen of
Jupiter, Saturn, and Mars, and despite the fact that the design itself is
quite poor off-axis (the MTF degrades more rapidly with field angle than an
equivalent focal length paraboloid), the views of deep sky objects are
memorable because the contrast between the object and background is
noticeably better than what I have seen in most Dobs that I have looked
through.

Well, that is enough from me.  Perhaps we will here from more folks with
larger instruments and better seeing conditions, now that you have broken
the ice.

Cheers,

Scott Milligan





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