ATM Pegasus Answers to Cert Questions

["Anderson, Jay" <andeja@mail.va.gov>]

I (Jay D. Anderson) forwarded some of the questions concerning Pegasus
mirror certifications and mirror testing, etc and be fore warned... John
Hall of Pagasus  put together a good 10 minutes of reading here (if you
read slow like me--ha!).

>There is a conversation going on an ATM list server about the certs you
>give on >your mirrors as some people are interested to know where to buy
> a fine mirror.
>
>>The question came up:
>
>Pegasus doesn't say how they make the test.  It could be caustic, poor
>man's caustic, wire test, foucalt, or who knows what.  Some test methods
>are far more sensitive (when done properly) than foucalt.  Some claim
>the caustic test is sensitive to about 1/100th wave.  John, can you shed
>some light on exactly how the testing is done at Pegasus?
>
ANSWER FROM JOHN HALL OF PEGASUS BELOW
========================================
Thanks for giving me an opportunity to respond.

A little explanation is maybe in order as to which test methods are "far
more sensitive" than Foucault.  I didn't realize that the test I use to
measure within 1/100th wave peak-valley at the wavefront was so
hideously
flawed.  Could this apparent non-sequitur be somewhat in the nature of
the
impossibility of the bumblebee to fly, but the bumblebee didn't know
that
fact of physics?  And flies anyhow?

To answer first question I see, I say exactly how I make the test in the
certifications accompanying each mirror, and thought enough info was
available in the brochure to clearly indicate this.  I use either knife
edge, null test, or interferometer or a combination per the individual
customer.  My error of measurement and error of components used is known
and
allowed for in calculating output numbers.

The knife-edge is used during production because the values are
quantitative, the appearance of the mirror is qualitative for
macro/micro-roughness, and this method has been used successfully by
myself
and many others for many years.  In conjunction with vernier movements
on
the knife and careful application, highly repeatable measurements are
possible within a given confidence level. +/- 1/100 wave peak-valley
repeatability at the wavefront is routinely realized.  If the customer
prefers a null test or interferogram, I will supply same at additional
cost.
In the absence of a specific additional requirement for data, I provide
an
unconditional guarantee of the mirror, in writing, along with a
certification of test data, actual quantitative summary results, and
graphs
of surface profile, error, wavefront profile and peak-valley/rms error,
and
Relative Transverse Aberration (RTA).  (The RTA confirms degree of
conformance to a diffraction-limited tolerance, via comparison of the
theoretical diffraction disk to the calculated diffraction disk radius.
It
is far more meaningful than wavefront p-v or rms, as it directly and
exclusively relates to the capability of the optic to focus, whereas
fractional wave expressions don't.)  I also provide a complete
explanation
of the numbers, how they are derived, and references.  Anyone who has
access
to a vernier knife-edge and some assistance from a knowledgeable amateur
can
independently conduct an evaluation of a mirror I have made to determine
accuracy of numbers supplied.  (It doesn't take as much expertise to do
a
knife-edge test as it does to figure a mirror.)

I don't know about the Caustic test's sensitivity because I do not use
it.
I suspect it is similar to the knife-edge in potential sensitivity,
depending on f-ratio to some extent.

No single method of test is 100% superior to every other method of test.
Each has its advantages/disadvantages.  This is one thing I learned in
many
years' experience in the commercial sector as a Quality Assurance source
representative.  I conducted evaluations of many tests, procedures,
equipment, calibration, etc. and supervised the conduct of hands-on
audits
of NASA and DOD suppliers in various commodities and applications.

You cannot test quality into a product.  Whatever method of test and
measurement is used during production is what determines the accuracy of
the
measured outgoing quality from the process.  What you can measure, you
can
control.  Whatever limitations exist in the test setup, calibration, or
internal components/references will be compounded in the output, and in
the
case of a mirror in a telescope, will be reflected in the modulation
transfer function.

Three things determine quality of a product:  (1) The stringency of the
performance or quality requirement for the product; (2) the ability to
measure the quantitites needed to meet the quality requirement and (3)
control it in manufacturing processes to bring its attributes within the
specification/standard range cited in the quality requirement.  This
latter
factor includes operator skill and experience.

There are three common standards, or quality requirements, for telescope
mirrors:  (1) The Rayleigh criterion, or 1/4 wave peak to valley at the
wavefront; (2) Simultaneously satisfying the Danjon-Couder conditions,
the
first of which is the Rayleigh criterion, the second of which is
conformance
on average to a diffraction-limited tolerance and (3) A
diffraction-limited
tolerance, in which every part of the mirror must direct its light
within
the boundary of the diffraction disk. The Rayleigh criterion, according
to
H. R. Suiter in TM 32, is "noticeably imperfect" and the
diffraction-limited
tolerance is "sensibly perfect."

I use a diffraction-limited tolerance envelope as my standard, which is
the
highest standard in use.  I have the means to measure the conformance of
my
processes and my mirrors to this standard and to control my processes to
produce mirrors conforming to a diffraction-limited tolerance envelope.
I
can and do use all the available tests and methods available to certify
my
optics.  The optics so certified have been independently verified as
conforming as stated.  And my mirrors are known for exceeding the most
rigorous performance standards during conditions of field use.

If you have any questions resulting from these comments, I will
appreciate
receiving them and opportunity to respond further.

Any questions?

Thanks,


John Hall
Pegasus