Re: [ATM] ronchi accuracy

["Mel Bartels" <mbartels@bbastrodesigns.com>]

> Suiter sez, p. 297, the "geometric Ronchi test" can't distinguish between
> a
> good mirror and a mirror with "unacceptable correction errors".  The main
> problem is diffraction messes up what you see.  I wrote the DIFFRACT
> simulation program in an attempt to use wave optics to model the
> diffraction patterns and turn Ronchi into the "poor-man's interferometer"
> -
> it didn't work.
> 
> Suiter also says that the Ronchi test sensitivity can be improved by
> moving
> the Ronchi screen longitudinally, measuring the movement and comparing
> with
> theoretical patterns.  I think Mel Bartels uses this method (but I'm not
> sure).  Suiter says, p. 296, "Unfortunately, the test is almost never
> undertaken in this manner, probably because it so closely resembles the
> Foucault test that the user was trying to avoid in the first



I've been using the Ronchi test for 25 years.  I make better mirrors using
the Ronchi test compared to using other tests.

Mirror testing serves one main purpose: producing a good mirror.  Testing
also serves secondary purposes of learning what figuring strokes do, and of
learning about optics.

Before we proceed with tests, we need to describe good mirrors.  I'll
confine my thinking to a variety of Newtonians and the occasional
Cassegrain.  

1) A good mirror is properly corrected, that is, it's fully parabolized such
that spherical aberration is absent at the focal plane.  
2) A good mirror has no zones.  
3) A good mirror is smooth, that is, there are no flat, deep, high, or low
portions.  
4) A good mirror is fully polished out.  
5) Finally, a good mirror has no 'primary' or third order astigmatism, where
a vertical slice of the mirror has a different best focus than a horizontal
slice of the mirror.

Not only do I need to test for all the above, but I also need a test that is
reliable, consistent, accurate, inexpensive, easy and fast to use.
Particularly in mirror making classes, the bottleneck is testing, when
students line up, waiting for their mirror to be tested and analyzed.

There's an amazing variety of tests to select from.  Some are easier to use,
some are more predictable, some are more accurate, and some are more
expensive.  Many tests use math.  Most Americans are mathematically
illiterate.  To recommend a math intensive test is to tell many that mirror
making is beyond them.

I'll only discuss the Ronchi so as to stay on the thread's topic.  It's
non-mathematical, and it's quick, cheap, and easy to use.  It does a super
job testing for zones (#2 above), and testing for smoothness (#3 above).

Very few tests adequately test 'stig (#5 above), so the Ronchi test is at no
particular disadvantage.  The long time knock is testing for overall
correction (#1 above), which is Suiter's reference.

The Ronchi can be used to test for 'in the ballpark' correction (#1 above).
I do this by comparing Ronchigrams of mirrors that are 1/4 wavefront under
and overcorrected to the ideal shape.  On large fast mirrors, the
differences are slight.  On small slow mirrors, the differences are obvious.
The Ronchi grating does need to be moved precisely up and down the optical
axis of the mirror, typically to quarter of a millimeter (hundredth of an
inch).  So some sort of sliding stage mechanism needs to be built.

For precision correction, the Ronchi suffers.  But so do most other tests,
specifically Foucault in the hands of the average mirror maker.  It's my
experience based on teaching mirror making classes over the years that the
starting amateur attains an overall correction using the Foucault that is no
better than the starting amateur obtains using the quantitative Ronchi test.
Additionally, the slowness of the Foucault means one or two tests an evening
whereas the Ronchi equipped amateur can run through a half dozen figuring
spells in the same time.  And then there's still the issue of 'stig (#5
above).

Consequently, I use and recommend to students in my mirror making classes to
conclude with the star test.  It tests (in many cases, overly sensitive) for
spherical aberration (#1 above), and it tests for 'stig (#5).  It exposes
problems with mirror mounts.  Mirror makers also get a sense of how the
in/out focus star test relates to the figure of their mirrors.  And that's
very worthwhile beyond simple mirror making.

Mel Bartels


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