From: "Herb Kasler" 
To: "Jerry B. Hillman" , 
Reply-To: "Herb Kasler" 



Jerry,

If you see obvious zonal defects on the eyepiece Ronchi test, I think you
should be a bit concerned. I think I was told by one of the experts at
Chabot that a Ronchi star test can see zonal defects down to about 1/4
wave. I know that my lines on the eyepiece Ronchi were straight as a die,
even when there were only 3 of them on the mirror (the extreme edge is
masked off), and I'm sure I don't have a perfect parabola.

I don't want to seem incredulous or in any way disparaging of your
achievement, but I would have to say that if you really hit it anywhere
near as close as your numbers say you did on your first round of
measurements with a 16.5" f/4.5 mirror, it would almost be a miracle.
To give you an idea of what I'm saying, Paul at the Chabot workshop
measured 5 zones on my 16" f/3.9 mirror and pronounced it inside 1/4
wave. I was trotted out in front of the group for a round of applause and
everything. As near as I can tell, it was in fact still almost a wave
undercorrected at that time. And nobody who looked at it could tell in the
slightest that the Ronchi lines were anything short of perfect (except the
outermost 1/8").  Like you, I started my measuring with a 9-zone
Couder Mask ala Texereau, but it soon became clear to me that it was not
accurate enough. I ended up doing wire tests with an 11-zone pinstick that
had four arms and 8 teeth for each zone. That thing was a real pain in the
behind to make, because the tolerance for the position of the teeth is
something like 0.2mm. Having eight of them helps, because then your errors
in fabrication will tend to cancel out if you make your measurement at the
"best fit". It was clear to me, after I became proficient using
it, that I was finally getting meaningful numbers. Couder masks on mirrors
that big and short are not really that useful, because the shadows are (or
should be) so sharp. On longer mirrors, where the tolerances are not so
tight, you can just kind of judge shades of grey, but that really doesn't
work well at f/4.5.

At any rate, after I started using that 11-zone pinstick, I did 5 more
retouching steps on my mirror before I decided it was done. four times, the
mirror moved in the direction I had intended for it to. One time it seemed
not to change at all, I think because my strokes were a bit too long. The
last step was with an annular star lap, and it lasted maybe 2.5 minutes. It
did exactly what I intended for it to do, reducing the 30-70% zone by maybe
35 nanometers at the peak. The average of four measurement sessions after
that made my mirror 1/9 wave PV on the wavefront, strehl 0.967. The
individual sessions weren't that good, but they were all less than 1/5 wave
PV and over 0.85 Strehl. Given that we have an atmosphere on this planet
and I don't live on top of Annapurna, that's good enough. Looking back over
my files, I count 12 long measuring sessions with the 11-zone setup, and 8
more with other types of masks/pinsticks. The upshot of all of this is that
it's very easy to get fooled by only one series of measurements, even if
you spent the whole day on it.. You may wish to proceed solely with the
star test, but my understanding is that with a 16.5" scope, you'll
need to use at least 600X mag in order to see the diffraction rings well.
limiting mag for a 16" aperture is 800X, and you should be in that
ballpark to do a star test. At 550X I was able to see something disclike in
dim star images on my scope on one exceptionally still night, but the discs
were still really tiny at that mag. To give you an idea, the diameter of
the Airy disk for a 16 scope is 0.7 arcsec. At 495X magnification, that's
like an object 0.8 mm across held at arm's length from your face. 
Tiiineeee. Seeing is a real problem for big apertures. You may wait a long
time in between opportunities to get meaningful tests. Collimation is also
super critical at short F ratios.

So I would reccomend you make yourself a pinstick with 10 or 11 zones, and
use Sixtests to analyze your measurements. It is really a much more
powerful tool than Tex. You can find an excellent discussion (by David
Harbour) of the Focault test and the advantages of using a pinstick here:
http://www.fortunecity.com/marina/manatee/1879/foucault/harbour1.htm

As for Sixtests, its quite simple. Rather than calculating where to put
your mask openings, you just arbitrarily decide where to place the radii of
your teeth and then enter them into the table. Just make them closer
together near the edge than in the middle. My radii for example were 60,
85, 105, 125, 140, 155, 165, 175, 185, 193, and 198 mm. As for converting
to metric, well, all you have to do is multiply by 25.4. Having gotten as
far as you have, I'm confident you can do that. The rest is simple. You put
the measured distance from the middle of your mirror to the slit under
source distance, and the distance from your KE to the mirror when your
micrometer reads zero under Xbias. Then you enter all your radii and
micrometer readings, in millimeters,  and click "surface".  The
best thing about Sixtests is that by using the Target feature, you can
change the radius of your target parabola in ways that allow you to choose
multiple strategies for your next figuring step. For example, by shortening
the target radius, you can often see ways to approach your parabola by
digging out the center rather than messing with the edge.



Good luck,


Herb.

--- BBBS/NT v4.01 Flag-4