Re: [ATM] Computing surface from Ronchi image?

["Nils Olof Carlin" <nilsolof.carlin@telia.com>]

Dale,

as I understand it (no guarantee for correctness), using a 2-dimensional 
grid of known dimensions in a Ronchioid test (as in John Sherman's images) 
will give a correspondence between apparent spot positions on the mirror 
(measured with a "pinstick" or better on an image) and the grid spot's 
position, thereby defining the normal to the mirror surface at that position 
(by defining the position on the mirror whose perpendicular passes throught 
the spot).

Using a linear, common Ronchi grid, you can compute a profile along the 
diameter (perpendicular to the lines) or along any line parallel to it, but 
not a truly 2-dimensional "profile" including astigmatism. However, the 
required maths would be much the same, and I guess a 2-d extension of Jim 
Burrows' algorithms in SIXTESTS would do the job:
http://home.earthlink.net/~burrjaw/atm/atm_math.lwp/atm_math.htm
(not that I'm anywhere near understanding it, but I expect this is the 
stuff).
I have a suspicion that the basic problem of evaluation is related to that 
of the Schack-Hartman test.

As in ordinary Ronchi, there is a limit to the combination of grid spacing 
and spacing of the point/line images on the mirror - closer lines will cause 
wider diffraction, and somewhere the points/lines will blend and be 
unreadable. Thus, the finer grid, the fewer samples on the mirror (this can 
be demonstrated by Jim's DIFFRACT).

Like in SIXTESTS, the "autofocusing" feature can estimate the ROC of "best 
focus".

Nils Olof

>>>I wonder if it would be possible to apply the same to the geometric 
>>>equations for Ronchi simulation.  I would think the variables are screen 
>>>lpi, mirror ROC and diameter and screen offset.  Knowing those and the 
>>>traced ronchi fringes perhaps it is possible.


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