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

["vladimir sacek" <vla@toast.net>]

Nils,


> one last attempt from me to clarify the issues I see here:
>
> The distance grating - center of mirror must be known (to some
> moderate precision - millimeters, not microns). Also, of course the
> Ronchi line spacing must be known (to one or at most a few microns).
> The location of the COC (of a best fit sphere, or the paraxial COC of a
> best fit parabola or other conic of your choice, etc) is the result of
> data reduction after the measurements. Thus, when you suggest as an
> alternative "you need to know the COC (grating) location", it seems
> that you are not aware of this distinction.

Seems that we are still talking past each other. You are talking about
"date reduction", but you can't have any direct data from the pattern alone,
unless you know its precise location. Also, best fit conic is the next step,
after you determine surface profile approximation. It is not a part of
determining the profile itself, and doesn't help clarity to mix it in.

>
> Again: your previous statement was:
> "Gaviola test is different, because it's based on direct measuerement
> of zonal foci, from which the local (zonal) radius is derived. In other
> words, it gives you two reference points - zonal ceneter and focus -
> determining the radius. "
>
> I claim there is no difference. In all cases, you decide the normal to
> the zone in question, in order to integrate the zones into an estimate
> of the profile (or deviation from the ideal conic desired).
> The position of the zonal center is one point to decide the location
> of the line, so one other point is needed.
>
> In Foucault, a point on the axis is chosen. In Gaviola, another point
> more or less close to the zonal COC is chosen. In the quantitative
> Ronchi method discussed here, the point is the position of the relevant
> line corresponding to the zone seen at the mirror.  The choices of
> points are different, the evaluation is not.

I wasn't talking Gaviola vs. Foucault. The thread subject was Ronchi test,
so when you entered Gaviola, I was talking about difference Gaviola vs.
Ronchi (which should be clear from the context). Again, the difference
of both, Gaviola and Foucault vs. Ronchi is that the former two use direct
measuerement, while the latter can only be used in that manner if the
grating location is exactly known.

Vlad

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