From: "James Lerch" 
To: "ATM List" 
Cc: 
Reply-To: "James Lerch" 


Greetings All,

I'm afraid to say, this is going to be one of those long winded posts.  If
your not interested in my Robo-Garbage, just hit the delete key and be done
with it
;)

Table of Contents:

Chapter 1, A Change of Plans
Chapter 2, The Mirror
Chapter 3, The Testing Method
Chapter 4, The Test Results
Chapter 5, Interesting Animations
Chapter 6, The UPS tracking number


** Chapter 1, A change of Plans.

In my original "Fresh Start" post, I had  a plan of action that
entailed testing
3 different mirrors with Robo.  Submitting the results to 'independent' 3rd
parties for safe keeping while the mirrors were sent out for testing with
an interferometer.  Once all the results were in, the 3rd parties would
"Reveal" the data, and the whole of the ATM list would know where
Robo stood in the scheme of things.

Well, there's been a change of plans! 

First off, since "I'm putting my money where my mouth (well, keyboard)
is", and
financing this project myself, were going to start with just one mirror. 
If Robo fails, its going to fail just a well on one mirror vs. three.  In
addition,
while having a mirror tested by an interferometer isn't prohibitively expensive,
it isn't cheap either (~$250 after shipping)

Secondly, If Robo fails, I'll at least have ONE mirror with a known surface
to either aid in finding a solution for fixing Robo, or if all else fails,
implementing Jim Burrow's idea of a Robo-Lateral-Wire test (Heck, even if
Robo works, that's still sounds like an interesting project!)

Lastly, "To heck with the independent 3rd party idea!"  I'm going
to divulge every last bit of Robo test data on the latest subject mirror,
and let the whole
ATM list be my independent 3rd party!  Links to the data are found in chapter 5.


** Chapter 2, The Mirror

First I have to extend my deepest appreciation to Charlie from our local
ATM mirror making lab.  Charlie had the intestinal fortitude to hand me his
mirror, let me test the crude out of it with my Robo-Foucault, then pack
and ship it across the US.  In addition, he has the foreknowledge that
every little intimate
detail of his mirror is about to made public!   


So, here's the Mirror's details:

#1 Full thickness cast Pyrex blank
#2 Hand ground, polished, and figured with an 8" tool/lap
#3 Diameter = 12.5"
#4 Focal Ratio = F/4.49

The optical surface does show micro-ripple, sleeks, rolled outer zone, TDE,
and some DUST just for good measure   However, the purpose of this
exercise is to
quantify Robo's abilities to measure an optical surface, towards this goal
Charlie's mirror is an excellent test subject!

It should also be noted, that Charlie does intend on masking the outer
1/4" of radius of this optic.  When viewing the surface error plots,
keep this fact in mind.  If Robo is right, and the optic is masked down to
12", this should make for a well performing optic IMHO.


** The Testing Method.

The optic was tested on Saturday, June 28th at our ATM mirror lab.  We
tested the mirror across three diameters in 60 degree increments.  We
labeled each of the diameters A,B, & C on the edge of the mirror for
future reference.

The pixel scale for the test was ~0.04" per pixel, with the optic occupying ~312
pixels.  A typical average zonal knife edge reading standard deviation (n-1) for
each test was approximately +/- 0.0014".  With most of the deviation
occurring on the inner zones.  I'll leave the exact standard deviation
calculation over the entire set of tests as an exercise for the list ;)

For each diameter tested, we ran a test for 7 zonal radiuses, 4 knife edge
readings per zone.  After the 7 zone test completed, we ran a 15 zone test,
again taking 4 knife edge readings per zone. Care was taken not to disturb
the setup between the 7 and 15 zone tests for each diameter, however it
should be noted that there is no correlation between the 7 and 15 zone
test's knife edge micrometer readings, other than their resulting surface
error rating.

For reference, Robo tests each zone in numerical order, then rewinds and
repeats the test, again going thru each zone in numerical order.  The
process will repeat until a total of 4 zonal readings per zone have been
recorded. (repetitions 0 - 3 in the file documentation)

Also of note is the fact that at the beginning of each test, when Zone 1 is
found for the first time, Robo sets the micrometer reading to zero.  All
subsequent knife edge readings for the remainder of the test will be
relative to
zone 1, repetition 0 (Yes, I did mix base 10 and Binary county methods, what can
I say other than I should probably fix that some day! )

I also have one final comment on the 15 zone tests.  That 15th zone is WAY
OUT there by the edge!  I don't know how much I trust it.  Normally I would
only do 7 or 8 zones on such a mirror.  However at the request of Michael
Peck, I ran the 15 zone tests.

My concern about the 15th zone is its located 3 pixels from the edge of the
optic, and should have a physical radius of 6.145" out of a possible
6.25" available (BTW, the optic doesn't have much of a bevel on it).

However, I should also note that if you review the text data, you'll notice that
Robo pretty much NAILS the same micrometer reading for each repetition of
the 15th zone knife edge reading.  That's fairly amazing IMHO!



** The Test Results

Current Robo test results for this optic are all stored on my web server
for access by the public.  In total, the results occupy 187 Megabytes of
data!

If your the type of person that just prefers to browse around for whatever
you might find interesting, Here's a link for you:

http://lerch.no-ip.com/atm/2ndTry/ (File & Directory  Browsing is
enabled, help yourself)


*Test results in Raw Text format*
Follow the below link to download a zip file containing all 6 test results
(3 diameters, tested over 7 then 15 zones) in plain text, for entry into
whatever data reduction program you prefer.

http://lerch.no-ip.com/atm/2ndTry/Raw_Text_files.zip (2Kb zip file)


*Test results in Sixtests Format*
http://lerch.no-ip.com/atm/2ndTry/Sixtests_files.zip (2Kb zip file)

If you don't already have a copy of Sixtests, you can download it from Jim
Burrow's website here:
http://home.earthlink.net/~burrjaw/public/win6.zip (468kb)


*Test results in Figure45 / FigureXP format*
http://lerch.no-ip.com/atm/2ndTry/Figure_Files.zip (2Kb zip file)

I've also uploaded a new Beta version of FigureXP that has two new
prototype features.

#1 You'll see two green arrows in the bottom right corner.  Clicking on this
will scroll and load the next Figure file in the current directory.  You
might find this useful when you want to quickly compare all six of the
surface profile
graphs against each other. (PS, you might want to re-name the files, so you
get all the 7 zones results first, then the 15 zone results next.  I should
have done this myself, but I didn't want to change the file naming
conventions, plus I got lazy...)

#2 The "Print" button doesn't print anything.   Instead,
it creates a
Picture
of the print job, then pastes it into the Windows Clipboard.  If you want to
'Print' a report, Click 'Print' in FigureXP, load your favorite photo editing
software and select the Edit/Paste function.  This will give you a Picture
of the print job, which you should be able to actually PRINT using your
photo editing software's print function (BTW, printing images in Visual
Basic is definitely a Hit or Miss affair, consequently this is my current
work around!)

Download FigureXP Beta from here:

http://lerch.no-ip.com/atm/FigureXP_Beta.zip (1.58MB Zip)


* FigureXP screen Captures *

If your not interested in running a program to see what Robo and FigureXP
predict the surface of the mirror may resemble, then I've got something for
you as well  The below link is a zip file containing screen
captures from all six
Robo tests.

http://lerch.no-ip.com/atm/2ndTry/FigureXP_screen_cap.zip (1.6MB zip)


* Raw images Downloads *

If your interested in just the raw images that were created each time Robo found
a zone null, You can download those as well!  First go here:

http://lerch.no-ip.com/atm/2ndTry

In this directory listing you will find six folders.  Each folder contains
the Foucault images obtained by Robo when it found each Zone radius null on
each repetition of the test.

The images are all BMP files, and broken into two categories.  Those with
the Prefix "RAW" are captures directly from video stream.  Those
without the "RAW" prefix contain the video overlay data that is
displayed while Robo is running its tests.

Also in each folder, you will find that I've already created a pair of Zip
files, which contain each of the corresponding BMP files in that directory.
Download at will :)


**Chapter 5, Interesting Animations

Robo-Foucault was based on the work of Michael Peck as detailed on his web page here:

http://home.netcom.com/~mpeck1/astro/autof/autof.htm


About half way down on Michael's web page, you'll see he describes a
"Flip and Difference" method of manipulating a Foucault image to
produce useful data. Robo does something similar, but entirely in software.
 (Robo skips the FLIP part, and only works with the Difference part)

I'm not sure why I hadn't tried the below experiment earlier, but in any event I
hope your find the results interesting.

First, let me provide links to the animations I created, below the links
I'll discuss how I created the animations, and what I think they show...

(7 zone Flip & Diff animation, in 2 formats, and 3 sizes)
http://lerch.no-ip.com/atm/2ndTry/Axis_A_Flip_Diff_low.avi (595Kb)
http://lerch.no-ip.com/atm/2ndTry/Axis_A_7_Flip_Diff.avi (1.2MB)
http://lerch.no-ip.com/atm/2ndTry/Axis_A__7_Flip_Diff.gif (2MB)

(15 zone Flip & Diff Animation, in 2 formats, and 3 sizes)
http://lerch.no-ip.com/atm/2ndTry/Axis_C_15_FLip_Diff_low.avi (1.2MB)
http://lerch.no-ip.com/atm/2ndTry/Axis_C_15_FLip_Diff.avi (2.8MB)
http://lerch.no-ip.com/atm/2ndTry/Axis_C_15_FLip_Diff.gif (3.7MB)

The animations were created by taking each of the Overlay BMP images for
the test in question, and apply the following technique:

#1 opening each BMP in Paint Shop Pro
#2 creating a copy of the image
#3 pasting the copy as a new layer
#4 Mirroring the new layer
#5 Selecting the "Difference" function for the new layer
#6 Aligning the new layer with the original image
#7 Merging the two layers into one image
#8 Applying a Gamma adjustment of 2.8
#9 Saving the resulting image

Once I had processed all 28 images for a single 7 zone test, I assembled
them into an animation using Paint Shop Pro's Animation Shop.  I then
repeated this WHOLE process for a 15 zone test (Yup, 60 image
manipulations!)

Anyway, what do they show?   I think they show why measuring an inner zone is
often difficult, the "Ring" isn't well defined (not surprising
considering the radius for zone 1 on the 15 zone test was only 1.17" .
 That's akin to trying to
find a null on a 2 1/3" F/48 optic!!)

The next item of interest is that 'for the most part' Robo does a pretty
good job on making the "Ring" land on the zone radius indicators.
 I can see that on occasion it misses, and I might be able to do something
about this to further increase Robo's consistency, but first things first,
lets see if it even works!



** Chapter 6, The UPS tracking number.

As of this moment, the optic is somewhere North of Jacksonville,FL on its way to
the interferometer.  Scheduled delivery date is currently July 3rd.  So
perhaps in as little as a few weeks, we will KNOW where Robo stands in the
grand scheme of things  :)

If you'd like to share my anxiety, just goto www.ups.com and punch in the
following tracking number 1Z V59 1R3 03 4117 405 5

At the moment, I think I'm going to take an aspirin and go to bed 

Take Care,
James Lerch
http://lerch.no-ip.com/atm (My telescope construction,testing, and coating site)

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