Re: [ATM] Crazy simple mirror cell idea

["Nils Olof Carlin" <nilsolof.carlin@telia.com>]

Robert,

> I could be wrong, but these numbers appear overrated by a factor of 2.

no, the mistake is mine, and many thanks for pointing it out - I was working 
from memory, and likely, as you say, doing the conversion from surface to 
wavefront errors twice, and not even checking before posting :(  So: 20 nm 
on the surface, and 40 nm (or 1/14 waves of 560 nm light) is what puts a 
mirror at the limit for Maréchal's criterion (or, in the special case of 
low-order spherical aberration, the "classic" Rayleigh's criterion) by 
lowering the Strehl ratio by 0.2.

Here, I hope, is a corrected version of the passage:

A simple memorizable approximation: a 20 nm RMS surface deformation (or 400
square nanometers MS if you like ;-) corresponds to twice that on the
wavefront, and assuming 560 nm green light, this deformation would lower the
Strehl ratio by 0.2 (from 1.0 to 0.8 for a perfect mirror, or from 0.9 to
0.7 for a less perfect one). 10 nm will lower by (5/10)^2*0.2 or by 0.05
which is not readily noticeable - 5 nm by 0.0125.

It may make sense to aim for (or just below) this when designing a cell for 
a quality mirror - if nothing else, it will give you some safety margin. A 
PLOP test run of the original 14" mirror, under reasonable assumptions, 
would give IIRC about 6.6 nm RMS error, and lower the Strehl ratio by 0.02 
or so.

I must agree that this would be hard or impossible to detect in practice.

Nils Olof

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