Star Testing Complex Optical Systems
by Roland Christen
--------------------------------------------------------------------------------

I should start out saying that I have the highest regard for Suiter
and his book, there are however enough holes in the book vis a vis
refractors and catadioptrics that it cannot be used for all
occasions. If you want to see some wildly or perhaps mildly
asymetrical printouts, contact TEC Telescopes. They can send you some
Zemax results of Mak-Cass designs that may want you to start
nibbling.

Now on to my soapbox: The situation with the star test is more
complicated than manufacturers being afraid to be found out now that
amateurs have an easy test to evaluate the optics. When the optical
system gets more complex than a simple parabolic mirror, then there
are inherent aberrations that affect the star test. An example is the
Maksutov Cassegrain. This system can be manufactured many different
ways, but one popular way is to have all surfaces spherical, which
cuts down the need for hand work. Machines exist now that can lay
down a 1/20 wave or better spherical surface on a piece of glass
without any human intervention. In this pure form, the Mak-Cass has
left over 5th order aberrations and, depending on design, these can
be less than 1/10 wave on the wavefront. By the way, fast Apo
refractors have these same aberrations also. The RMS value will be
better than 1/50 RMS and the Strehl ratio will be exceedingly high.
In other words, the optic will deliver a very high contrast image,
consistent with the high wavefront rating. When tested on the night
sky, the inside and outside diffraction patterns will be quite
different. Any beginner will see it so and may conclude that the
optic is of poor quality.

As an example, I have recently finished exhaustive tests of different
10" F14 Mak-Cass systems, some with these inherent aberrations left
in, some with them meticulously removed. All the optics tested
between 1/10 and 1/12 wave with Strehl ratios of 98% or better. The
system with the uncorrected or pure Maksutov curves had the central
hole of the donut break out 4 times farther on one side of focus than
the other. It would be judged by the Suiter star test method as being
maybe barely 1/4 wave, if that. The hand aspherized version, which
was no better on the interferometer and had similar Strehl ratio,
showed a donut breakout approximately at the same point on either
side of focus - in other words textbook perfect. At focus, both
scopes showed a tight central Airy disc with a faint first
diffraction ring. At focus, it was impossible to tell them apart. The
ratio of brightness of the Airy disc and first diffraction ring was
essentially identical. Both scopes showed the same high contrast on
Jupiter and Saturn. Both scopes could split doubles with equal
precision.

This same thing happens in my short focus Apo triplet lenses. They
also have classical 5th order like Mak-Cass systems. I have seen
literally thousands of star patterns on my test setups. I can tell
you that the shadow breakout difference is almost entirely due to
slight increase in focal length of the very inner zones on these
optics. That is one half of the 5th order defect. This causes no
problem in definition and contrast for 2 reasons. 1, there is very
little energy as you go toward the middle of the optic, and 2, the
depth of focus approaches infinity at the very center. However, the
shadow breakout is linear, not asymtotic like the encircled energy.
This is easily fixed, of course by commercial makers by simply making
a large central obstruction.

The other half of the defect occurs at the outer zone. If this is
left uncorrected, you will see fuzz in the in-focus image which is
highly destructive of contrast. The inside and outside patterns will
look the same, causing you to conclude that the optic is textbook
perfect. It is this outer zone that the competent optician will
concentrate on, since this is the place that contributes most to the
definition, resolution and contrast of the optical system, even
though it has little or no effect on the inside/outside star pattern,
particularly where the shadow breakout occurs.

In view of the above, I have seen amateurs at star parties wrongly
interpret the star test and overestimate the quality of one optic
that was really not that good, and underestimate another that was
really superb.

I think all manufacturers will agree that the star test must be
properly done to place a VALUE on the correction. This is not so easy
with certain types of defects, and so the star test, when improperly
interpreted by untrained individuals, can overestimate and also
underestimate the actual quality. If you ask any reputable maker of
complex optical systems, they will agree that the interferometer will
give the most reliable result. Since Peter Ceravolo will not get on
SAA, I will paraphrase what he has told me time and again : You
cannot rely on anything but the interferometer to give you a
wavefront NUMBER. The interferometer takes no prisoners.

In my experience, when an optic measures 1/10 wave AT THE DESIGN
WAVELENGTH, then the performance will be nothing short of stunning.
Peter's criterion was a bit less at 1/8 wave P-V. Either way, the
customer will have an excellent optic.

By the way, Ceravolo is my mentor on testing of optics. Even though I
have been making optics for a long time, without his guidance in the
testing lab, I never would accomplished what I did. And yes, I used
to use the star test in the distant past when I made longer
refractors. Some of my optics came out good and have surprised me
when retested on the interferometer. But for the fast F6 and F7
triplets I make now, the star test falls short.

The dilemma for manufacturers then is, should we do our best to
produce smooth high contrast optics, or should we please the star
test crowd and do some hand aspherizing to get a more pleasing
out-of-focus star image? I can tell you that it is easy to do some
rough compensation with quick local polishing at several zones to get
more equal inside and outside star patterns, but the result will
almost certainly be a loss of contrast. Add to that a nice big
central obstruction to get rid of the offending inner zones, and
presto! you have a nice "fast food" Mak-Cass that doesn't work any
better than a typical SCT.

In our case, we will do our utmost to produce the closest faximile to
the star patterns in Suiter's book, but they will never be exactly
equal. The overriding concern will be that the optic has a very
smooth and accurate wavefront to produce the highest contrast
possible in the final image, which I assume will be in focus.

Roland Christen ASTRO-PHYSICS
--------------------------------------------------------------------------------
Return to Roland Christen Images and Essays.