Re: [ATM] Transverse Error

[Mike Lockwood <melockwo@uiuc.edu>]

Hi,

Alexis Cousein wrote:
> I'm going to be a nit and object to the term "being direct
> outside". It does give you a *qualitative* appreciation for how
> little that error will cause energy to be put in diffraction rings
> rather than the central disc, but the error itself doesn't in
> itself direct real physical energy along the ray.
> And if you want to do a *quantitative* analysis, then you need to
> use a physical model that's based on wavefront modeling and not
> ray-casting: light doesn't behave like ray bundles.

In Foucault testing we mostly assume light is behaving like a ray 
bundle, and it works sufficiently well to make superb mirrors. 
FigureXP is used for reducing Foucault data, so I explained transverse 
error in that context.

FigureXP also does a more advanced simulation of a star image at, 
inside of, and outside of focus if one wishes to see the effects that 
errors have, as simulated on the wavefront.

> In the times of slow PCs (the time Rutten and Van Venrooij's book
> was written), you couldn't do fast FFTs on PCs, so the geometric
> approach was useful to determine when good enough was absolutely
> certainly good enough (with some added data points from models that
> did model wave diffraction, like the size of the central portion of
> the Airy pattern for a given circular aperture).
> That doesn't mean ray casting models a physical reality, though.

Rays are still good enough, IMO.  I get enough FFTs at my day job, and 
if I keep transverse errors very low my eyes and my clients never 
complain.

	Mike Lockwood

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