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


Greetings All,

In the quest to solve the Robo mystery, Robo2 has been finished.  (is
anything every really 'Finished'? )

In the past, My Robo software, Dale Eason's software version of Robo, and
Mike Peck's "super-secret null finding algorithm." have all
closely agreed with each other. However,  these methods all relied on my
hardware (X-Y dingbat, light source, knife edge, and camera)

In an effort to eliminate the hardware as the culprit, Robo2 has been
completed and tested against Charlie's mirror.  If you just want the short
version of this
post, Robo2 produces the same answer as Robo1, with the exception of the
15th zone. (more on that later)

So, Here's a picture of Robo2 made while I was testing the camera:
http://lerch.no-ip.com/atm/2ndTry/Robo2/Robo2.jpg (62KB)

As you can see, it is fabricated from nice machined parts hand crafted by
Paul McNabb from our ATM lab.  (BTW, the clamp on light is only there to
illuminate the pizza box I used as a test target)  Mechanically Robo2 is
much nicer than Robo1.

The Camera for Robo2 was the next interesting item.  In the back of my mind, all
the lenses inside a typical Zoom lens camera have worried me, towards
eliminating all but one lens, I made the simplest camera possible.  Here's
an image of the completed camera:
http://lerch.no-ip.com/atm/2ndTry/Robo2/Robo2_Camera.jpg (48KB)

The camera started life as a Philips Vesta Pro CCD webcam, with a 640*480
resolution and 5.6 micron pixel size.  After removing the factory lens
(which just un-screws) I created a replacement lens from a single DCX 35mm
focal length
lens.  After mounting, 20mm of aperture is available and the camera now operates
at F/1.75.

Since the lens isn't an acromat and it operates at a fairly fast F/Ratio, I
was worried about how the camera would perform.  To test the camera, I cut
up an old
pizza box, drew a grid on it and placed it the same distance the mirror would be
from the camera (111 3/8")  The grid spacing ended up at 0.707"
squares, so I added some changes to Robo so it would do a series of zones
at 0.707" increments, and draw complete circles for each zone.  Here's
the result: http://lerch.no-ip.com/atm/2ndTry/Robo2/zone_cal.jpg (87kb)

As you can see, the simple lens and camera seems to work fairly well.

Now its time to run a knife edge test.  In this case I tested across Axis A
of Charlie's mirror using the same 15 zone radiuses as I did on the Robo1
tests, here's the comparison: (moving source tester, average of 4 readings
per zone, mean stddev (n-1) Robo1 =0.0014", mean stddev (n-1) Robo2
=0.002")

   Zone Radius                    Robo1       Robo2         Difference
 Zone 1 radius =  1.170      0.000         -0.005         -0.005
 Zone 2 radius =  2.020      0.012           0.017          0.004
 Zone 3 radius =  2.586      0.025           0.016         -0.009
 Zone 4 radius =  3.047      0.032           0.026         -0.006
 Zone 5 radius =  3.446      0.044           0.038         -0.007
 Zone 6 radius =  3.804      0.054           0.055          0.001
 Zone 7 radius =  4.130       0.069          0.066         -0.003
 Zone 8 radius =  4.432       0.080          0.075         -0.004
 Zone 9 radius =  4.715       0.089          0.085         -0.004
 Zone 10 radius = 4.982      0.096          0.097           0.001
 Zone 11 radius = 5.236      0.105          0.110           0.005
 Zone 12 radius = 5.477      0.125          0.123          -0.001
 Zone 13 radius = 5.709      0.139          0.138          -0.002
 Zone 14 radius = 5.931      0.159          0.164            0.005
 Zone 15 radius = 6.145       0.195          0.218           0.023

As I see it, Robo2 shows a deviation from Robo1 for the first 5 zones
(which concurs with the deviation Mike Peck had Vs Robo1) which is a good
thing! Finally that 15th zone is substantially different, which I attribute
to having more pixels to work with.  All the raw data and images for the
Robo2 run can be found here:

http://lerch.no-ip.com/atm/2ndTry/Robo2/

Here's a simple image comparison between Robo1 and Robo2, (masking the
mirror down to 12" diameter, dropping the 15th zone, surface profile
across axis A as seen by FigureXP)

http://lerch.no-ip.com/atm/2ndTry/Robo2/Robo1_V_2.gif (34KB)

At this point, the obvious next step is to star test the optic, and see what we see.

As it stands now, Robo doesn't agree with Couder Mask testing nor
intereferometry done at RoC w/o nulling optics.

I'm pretty confident I've ruled out software, hardware, and the imaging
system as culprits, which only leaves the theory of digital knife edge
testing as a possible flaw.

I guess time will tell!

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

"Anything that can happen, will happen" -Stephen Pollock from:
"Particle Physics for Non-Physicists: A Tour of the Microcosmos"

--- BBBS/NT v4.01 Flag-5