Re: [ATM] fringe testing three flats

[Mike Lockwood <melockwo@uiuc.edu>]

Thomas,

Thomas Janstrom wrote:
 > I usually aim for about 10-15 fringes.

Try for only a few fringes (~5) if the mirrors are only 4.25" in 
diameter.  If they are clean, this won't be a problem.

> Anyway now that I'm down to an error of less than .5 of a fringe (bulk),
> although there is a low center on one and another has a somewhat nasty
> 1/4W X 1/2" wide turned edge. So what I'd like to know is how do I
> *know* if the surface is concave or convex? 
> I've tried the moving my head towards and away thing as mentioned in ATM
> vol. 1, but there is no discernable movement of the fringes one way or
> the other. So any suggestions? 

Put two of the pieces together and get a bullseye (if you can - if you 
can't see the next paragraph).  When you press on the edge of the pair 
under test, if the fringes move toward your finger there is a net 
convexity between the two pieces.  If the fringes move away, there is 
a net concavity.  Alternately, if you press in the center of the 
disks, if the fringes move inward, there is a net concavity, and if 
they move outward there is net convexity.

If you get only sections of fringes that are reasonably straight (no 
bullseye), apply pressure in the center of the disks with a finger. 
Note the direction that the fringes move as you apply pressure - this 
is the direction of a wider air gap (because you are reducing the air 
gap with the pressure).  Release the pressure, and examine the 
fringes.  Fringes that bend in that direction (towards the larger air 
gap) will indicate a high spot.  (As the fringe crosses a high spot it 
must bend toward a wider air gap in order to maintain its contour of 
equal distance between the glass surface.)

To determine the amount each flat varies from true flatness, you can 
then use the results of three combinations (flat A on B, B on C, and C 
on A) of tests to solve for the error of each of the three surfaces 
(as a function of radius, if you like), as I think the article you 
mentioned in ATM 1 will tell you how to do, and which you probably 
already have done.

If you are using an uncollimated light source, make sure you view the 
fringes from at least ~48" away, to reduce parallax error to 
negligible levels.  If your source is collimated, you don't have to 
worry about it.

Hope that helps, and I hope I didn't accidentally reverse any of 
these.....  ;)

	Mike Lockwood

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