[ATM] Optical flat part II

[Vladimir Galogaza]

Francis,

>The foucault test and the ronchi test may be performed 
>with an autocollimator. However they become null tests 
>which are much easier to interpret with an autocollimator than without one
>That is why I have become a fan of autocollimation.

One of autocollimation virtues is that the space required for testing 
is halved.
Martti Koskimo once proposed a liquid surface in place of flat mirror
for test in autocollimation. Mirror to be tested is supported above the
pool of oil which serves as autocollimating  flat.
I wander if somebody tried this. Should not be more complicated
than interference test with mirror under the  surface of liquid.

Regards
Vladimir.