Recognizing Optical Quality--That Old Article (long)
ste...@binghamton.edu
vintage Sky & Telescope article about reflectors of different wavefront
quality. Given that the memories differed--and that I had just been
browsing that old article--I thought I post a summary. I note some
interesting wrinkles in the testing.
The article by Ceravolo, Dickinson, and George appeared in the March 1992
issue, page 253+. The purpose was to see how well observers could
distinguish--in a blind test--among reflectors of known, varying quality.
They used 6" f/8 mirrors.
(Is there any reason to think that the results might differ at lower f
ratios or in larger mirrors? I ask because we should figure out how well
the results generalize.)
First, the primary mirrors were near 1, 1/2-, 1/4-, and 1/10-wave
peak-to-valley (PV). (The mirrors were ground, figured and polished by
Ceravolo and tested using a Zygo interferometer.)
Second, the error that was put in was spherical aberration. Could this
affect the generalizability of the results? Might other types of
error--turned edge, rough surface, etc.--have been more or less apparent?
Third, it should be noted that commercial secondary mirrors of varying
quality were used! Two were 1/20 wave, one was 1/10, and the last was
1/2. They put the three best diagonals with the three best mirrors. (So
much for a standardized test of the primary's quality. Although given the
pairings, and the high quality of the 3 diagonals, it shouldn't matter--or
should it?)
The blind testing was done separately by Terrence Dickinson and Douglas
George, and then at Stellafane.
FINDINGS
DICKINSON
Dickinson did his tests over five separate evenings from his rural home.
He reported on observations of the globular M15, the Orion Nebula, NGC
2158 (cluster beside M35), and Zeta Orionis (2nd, 4th magnitude double
star with separation of 2.4"). An 80mm Celestron refractor matched the
1/2-wave 6" scope on Martian detail but gave a more pleasing image, and
beat it on the double star.
He found that the 1/4 and 1/10-wave mirrors gave generally comparable
views, although on nights of good seeing, he did spot sharper and finer
detail in Jupiter's belts and subtle differences on Mars with the
1/10-wave mirror. Both gave detail on Mars that approached a 5 1/2"
Astro-physics refractor.
"Frankly, I am surprised by the results. . . On those rare nights of
exceptional seeing, maybe the 1/10-wavefront mirror would show its stuff,
but even then I doubt the difference would be noticed by most observers,
except in rigorous side-by-side comparisons on selected test objects"
(255).
He commented that a 12 or 14" reflector need be only 1/4-wave to give peak
performance (256).
GEORGE
George, too, found the 1/4- and 1/10- mirrors were comparable, save under
"careful scrutiny under excellent seeing" (257). He looked at the Orion
Nebula, the globular M79, and Jupiter (including a shadow transit).
He was "amazed" that a well-made 6" approached the performance of a
quality 7-inch StarFire refractor. But he also felt that "many amateurs
would be quite happy" with the performance of the 1/2 wave scope. (Near
the end, he states this as "most amateurs would be satisfied".) Still, he
noted that 1/4-wave is "definitely a minimum standard" for planetary
viewing and that the diffraction limit of 1/4-wave is a "myth" because the
1/10-wave mirror did produce a better view. (In this, he is diverging
from Dickinson, seeming to suggest that we should seek better wavefronts)
He found that the focus "snap test" readily distinguished the 1/4 and 1/2
wave mirrors (meaning that with the better mirror, the image snapped into
a crisp focus rather than being mushy and never quite focusing sharply).
He also noted that the star test--under very good seeing--could separate
the 1/4 and 1/10 mirrors. (Although in the body of his report, he did not
find such clear distinctions using the star test for the 1/4 and 1/10
mirrors.)
STELLAFANE
Alas, the testing done at Stellafane was only of Polaris! Alan MacRobert
conducted the testing and gathered reports from 103 observers. S & T only
used the 3 best scopes--that is, the 1/2, 1/4, and 1/10 wave ones.
No matter what one's level of experience, nearly everybody could pick out
the bad apple in this bunch (1/2-wave). The 1/4 and 1/10 comparison
proved more difficult.
"Nevertheless, about two-thirds did correctly identify the best scope as
giving the best image. This was despite the fact that the atmospheric
seeing was generally average to below average" (256)
But, get this. Many viewers did *star tests* to make their
judgments--rather than relying on the views themselves. (Given the
purpose of the test, this seems to be cheating!) They found it easy--that
way--to distinguish between the two best mirrors. Interestingly, the star
test was effective even though the seeing conditions were only average.
Relying, however, just on the in-focus views to make the discrimination
was much more difficult. Still, more than half picked the right one
(i.e., which was better than random guessing). Some figured out which was
best by seeing how well the 9th magnitude companion showed up.
(This again suggests that the views were generally similar in the 1/4 and
1/10 scopes. A possible wrinkle here, however, is that maybe those who
relied on the view rather than the star test were less experienced. . .)
MacRobert reported that the seeing conditions affected people's
confidence--when the skies steadied, there was more confident selection of
the 1/10 wave mirror as the best.
The conclusion: "when the seeing turns good, many persons can indeed make
out a slight difference in sharpness and clarity between 1/4-wavefront and
1/10-wavefront telescopes in focus."
Note the phrases "good" seeing and "slight difference". But again, this
was only of Polaris.
I highly recommend the article. The details and descriptions are
informative and make for fascinating reading. You can often find old Sky
& Telescope issues at a local university or public library.
tfla...@my-deja.com
ste...@binghamton.edu wrote:
> the error that was put in was spherical aberration. Could this
> affect the generalizability of the results? Might other types of
> error--turned edge, rough surface, etc.--have been more or
> less apparent?
Absolutely! I cannot speak from my own experience, but I have
read in numerous places, and heard directly from people who are
certainly in a position to know, that roughness damages the image
*much* more than a modest spherical aberration, and turned edge
can easily be a disaster. I can't put any quantitative weight
behind this second-hand statement, however.
--
- Tony Flanders
Cambridge, MA
Sent via Deja.com http://www.deja.com/
Before you buy.
Tom Hitchcock
this in '92. I believe Suiter also stated that 1/4 wave was about the cutoff
for most observers in observing, using mirrors with SA. Although, star
testing is a different matter. I wonder how they were determined that the
degree of SA was consistent in all samples? Also, it is mentioned that TDE
is devastating to images because it involves the outer portion of the mirror
and thus a large surface area to wreak havoc on images. In this case you can
have a great specs such as 1/10th wave but a noticeable defect with a
profound/narrow TDE - correct me if I'm wrong. I'm wondering if TDE can be
differentiated from overcorrection at some *point* - maybe it should be
referred to as just a severe form of overcorrection?
In answer to your question - I would say different deformities have
different field testing tolerances. Suiter mentions that the amount of
central obstruction also plays a role on how much SA is noticeable.
Thanks again. I'm going to check out the S&T. Hope you get more responses
from the more knowledgeable.
Tom H.
<ste...@binghamton.edu> wrote in message
news:stedman-1703...@128.226.111.6...
Mike Barrs
> of a vintage Sky & Telescope article about reflectors of different
> wavefront quality. Given that the memories differed--and that I
> had just been browsing that old article--I thought I post a summary.
> I note some interesting wrinkles in the testing.
>
> The article by Ceravolo, Dickinson, and George appeared in the
> March 1992 issue, page 253+. The purpose was to see how well
> observers could distinguish--in a blind test--among reflectors of
> known, varying quality. They used 6" f/8 mirrors.
<snip summary>
Thanks for posting that summary! Some interesting stuff there. Here are a
few random comments, for what it's worth....
First, I think it's significant that Terence Dickenson says "a 12 or 14"
reflector need be only 1/4-wave to give peak performance"... and the crowd
at Stellafane proved him wrong. So much for expert opinion. :-) I do like
Dickenson's books and I respect him as an author, but this shows how
dangerous it can be to make assumptions like that.
I hate to see an interesting test get muddied up with the choices they made
for those secondary mirrors. It would have been nice to see identical 1/20
wave secondary mirrors used on all the scopes. I'm sure this criticism must
have been mentioned before, and the secondary mirror choices probably
weren't enough to invalidate the test. It's just an unfortunate
complication.
I wonder to what degree the results would scale up to larger mirrors, say
18"? On the one hand, there are increased limitations from atmospheric
turbulence with larger mirrors, which you could argue would harm the
performance of the better mirrors. On the other hand, you see so much more
detail with larger mirrors (under decent seeing conditions) that any
increased resolution might help the evaluation process. Maybe those two
things cancel out... or maybe not? My feeling is that it would be easier to
tell the difference between these mirrors if they were larger, but I can't
prove it. Somebody run another test with 18" mirrors, okay? :-)
The choice of spherical aberration might have oversimplified things. They
probably chose that error because it's easy for the mirror maker to "dial
in" different amounts, while controlling other types of error. But it still
fudges the results a little. This is an aberration that isn't so easy to
spot at the higher wave quality levels, unless you're an expert. If they
were comparing equivalent levels of astigmatism or TDE, I wonder if the
results would have been the same? To make any sweeping conclusions about
what the average observer can see in a mirror at different quality levels,
you'd have to run the test with different types of specific errors, along
with the mixes of errors you get in the usual mirror. All the S&T test gives
us is a start on answering these questions. I don't believe it's the final
word on the subject.
As a final comment... your summary seems to knock the star test as a
performance criteria. I don't know if you're adding that yourself, or if it
was written that way in the S&T article. I don't see why you wouldn't use
the star test as the fundamental criteria for judging mirror quality. With
general observing targets (like planets), your eye can be easily fooled by
poor sky conditions. Is that a thin layer of high-altitude haze up there
killing contrast, or is the mirror not showing good contrast because it has
a turned edge? Viewing Jupiter will not alllow you to make that
determination, but a star test will. It seems to me that the Stellafane
attendees knew what they were doing, when they used star tests for rating
the optics.
Mike Barrs
Frank Bov
Thanks! You've done a great service with this piece of research. My library
only went back to about '94 or '95; I would surely have caught something
like this!
Thanks again,
Frank