Re: [ATM] [atm_free] RE: data and musings on thin mirror

[Stephen Koehler <s.c.koehler@gmail.com>]

Vladimir,

> Yes, but you still need to know what is the grating location for the
> tested pattern. For the line number count, any effect of change
> in the grating location can be cancelled out by an appropriate change
> in the amount of aberration vs. perfect parabola, and vice versa.
> In other words, one practically identical count can be result of many
> different location/aberration combinations. Starting out with the test
> pattern alone, it is impossible to compare it to a reference pattern
> without knowing the grating location (assuming the aberration is unknown).

I could be wrong, but I think my analysis is correct.  This is an
interesting disagreement, because I think it is one that we can prove
one way of the other by example, theory, or both.  It's possible that
we are not yet talking about the same thing, so  we should nail this
down, first.

Here's a summary of my position.  I contend that for a relatively
smooth mirror figure (focus position doesn't change direction as you
go from the inner zone to the outer zone) if you are outside the focus
for both outer zone and center, that there is a fixed line count
difference between the actual line count across the mirror and the
line count gotten by extrapolating the inner line spacing to the full
width of the mirror.  Moreover, this line count difference can be used
to approximate the focal shift of the mirror from center to outer zone
if you now the spacing of the grating, the focal ratio of the optic,
and whether the test is fixed or moving source.

I have, so far, demonstrated one example where this worked, at least
for five positions of the grating.  If your contention is right, we
should be able to find an example where the line counts do not result
in the proper focal shift.  Assuming you want to pursue this, do you
have a suggestion for what to try?

--
Steve Koehler
_______________________________________________
ATM mailing list http://www.atmlist.net/