Re: [ATM] Ronchi test accuracy

["Jerry" <wa4guu@verizon.net>]

I agree. With the Ronchi test you can see with a glance if the figure is
worth measuring. You can see that in the maskless Foucault test also. But
you can get the bands in a position to clearly show all the changes in slope
at the same time. With Foucault you might have to move the knife to see
different parts of the curve. So I like to figure by mostly looking at the
Ronchi and maybe a quick measure of overall percentage of correction.

I particularly like the Ronchi test when the curve is very close to
finished. Subtle changes in slope that might be hard to detect in Foucault
shadows can be seen a little easier. The better you see it the better you
can fix it.
 
Many mirror makers or would be mirror makers dislike the Foucault test and
"judging" shadows. They are in search of an easy test that is free of
measurement and judgment that will tell them exactly, not just within some
margin of error, but exactly the magnitude of error on the mirror. To that
end some find hope in the Ronchi test. It does away with most of the
measuring and leaves only "comparing" the Ronchi bands to a predicted
appearance. Of course comparing requires a judgment call as to how well they
match and how much error would result from any perceived mismatch. So much
for getting rid of the "judgment". There might be less judgment in the
Foucault/Couder mask test.

I don't particularly like the idea of using the Ronchi alone, although I
think I could make a "pretty good" mirror that way (1/10 wave or so). Of
course it is that I have made mirrors testing by Foucault and Ronchi that I
"think" that I could "judge" the Ronchi bands well enough to get pretty
good.

I would probably find it hard not to use one of the grating lines for a
knife edge to measure correction from center to edge. Using Ronchi and that
little cheat of measuring the correction center to edge with a knife I think
I can safely say I could then get to 1/20 wave on the wavefront on a typical
8 or 10 inch f6 or so. Some others who have used both tests enough to be
comfortable with them feel that way too. 

I think the "newbies" who wants to test only with Ronchi without using and
learning Foucault are asking for trouble. Much can be learned seeing the
Foucault/Couder shadows and zone numbers changing with the changing Ronchi
band shapes. They probably won't do as well as easily as they would using
Foucault and for sure not as good as using both.

Foucault and Ronchi testing are powerful as a pair.

Jerry


-----Original Message-----
From: Guy Brandenburg
 
I don't claim ANY numerical accuracy whatsoever  for the Ronchi test. What I
*DO* claim is that the Ronchi test is a  very fast and convenient method
that you can use to check the overall  figure. If if looks bad under Ronchi,
then there is no need to do any  zonal Couder-Foucault measurements or any
other numerical/quantitative  test whatsoever. A quick Ronchiscan can tell
you what parts of the  mirror need to be worked on, in a jiffy (not really
1/60th of a second,  maybe it will take you 5 to 10 seconds). Now, if your
mirror looks good  under ROnchi, then it's time to begin running numbers
with a couder  mask (or whatever). And I find running the numbers can take
nearly an  hour if you do it right, repeatedly, and on 2 or more diameters.
  
  Ronchigrams save a *tremendous* amount of time, in my opinion.
  Guy



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