To: Vladimir Galogaza , ATM shore

From: Tim Rickard 
Reply-To: Tim Rickard 



>
>Reading this and subsequent discussion I wander if, taking into account
>statistical nature of surface error  numbers ( Strehl, RMS, PV)
>these are linearly additive for two or more subsequent mirrors or some other
>relationship has to be used to determine  final   error resulting from
>individual errors.
>( James?)

I recall reading that the expected value of the total wavefront error is
given by the square-root of the sum of the squared error components, which
of course must include misalignment, thermal, mirror support, and eyepiece
components. This equation give the geometric mean. Also, multiply to
diagonal error by a coefficient of 1.41 if at a 45 degree angle.

However, for any given scope, the sources of error could reinforce, leading
to larger total error than predicted by this equation, or could cancel out,
leading, in the ideal case, to zero wavefront error. Hence, rotating one
mirror w.r.t. the other might improve performance in some cases. The
equation just gives on the expected error value on average.

I realize that the type of error is also crucial, so this analysis may have
limited value, but it may provide a rough guide for thinking about
cumulative error. Make realistic estimates of all of these error
components, run them through the equation, and become despondent :)

Cheers,
Tim

--- BBBS/NT v4.00 MP