Re: [ATM] fringe testing a convex and concave surface together.

[Aplanatic@aol.com]

A fringe analysis program can remove a few fringes of 
curvature for you, obviating the need for collimated light 
and a distant viewing position (or other method to 
reduce the effect of a close-by point source/detector).
 
The amount of perspective-induced curvature in a 
contact Fizeau interferometer is proportional to the 
spacing between the surfaces.  If you are brave enough 
to directly contact the surfaces then there is usually no 
need for a fancy viewing setup.  If you try this, never slide 
the surfaces over each other.  Bad things will eventually 
happen.
 
-- Dave
 
In a message dated 8/4/2005 4:19:12 P.M. Pacific Daylight Time,  
mdholm@telerama.com writes:

Mike  Lockwood wrote:

> Additionally, fringes should be viewed from a  distance of at least 10
> times the diameter of the pieces away from  them in order to get an
> accuracy of 1/10 wave.  Typically I try  to photograph the fringes with
> my digital camera from about twice this  distance away.

Not personal experience here, but reading things in  S&T and the late ATM
Journal.  Isn't the preferred method to use  collimated light?  A fellow
writing in ATMJ (the very last issue, if I  remember correctly) mounted a
Newtonian upside down and used a beam  splitter at the eyepiece end so
the scope acted as both a collimator and  decollimator.  He was also
using the water flat method.  As I  recall the S&T article from even
longer ago, the author suggested using  a Fresnel lens with a simple beam
splitter to collimate.  I don't know  if he viewed at a distance or used
the Fresnel lens as a decollimator as  well.  Evidently, the optical
perfection of the  collimator/decollimator can be pretty awful without
messing up the  test.  The only hitch I remember is that he recommended a
fairly high  f-ratio Fresnel lens.  Most of the ones on the surplus
market are out  of overhead projectors, and have fairly low  f-ratio.




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