ATM Re: LED Fou Cult Tester: still not right

[Bratislav Curcic <epabcc@epa.ericsson.se>]

Hej Nils,

< lots of effort in cutting the LED close to the chip deleted here >

one simple way to get very small pinhole is to use fibre (glass or
plastic, doesn't matter) and MELT the end of the fibre in hot flame,
trying to form nice clean sphere at the end. The other side, in contact
with LED is of course polished flat. The light cone formed this way
is more than enough for most applications (Maksutov corrector fans 
and other loonies making f/1 optics need to be more imaginative!), and
virtual pinhole is VERY small.
As for minimum required slit/pinhole, just remember that knife effectively
cuts parts of the incoming light, reducing the slit/pinhole to next to 
nothing just before you cut the light completely. This is how slitless
tester works, after all. So, yes, it doesn't matter at all how wide your 
slit is. But small pinholes are desirable for eyepiece tests of residual
astigmatism, and they need to be smaller than 1.22*lambda*F_eff . 
Also very nice for studying diffraction effects. In fact, I have never 
seen so beautiful and steady Airy disk and rings as while testing an f/10
sphere. A very neat trick is to use thin Al foil and hold it in the
flame until just starts to burn/melt. Tens of perfectly round and small
pinholes develop, forming a beautiful "star cluster" that can be
studied at leasure. (until you parabolize beast :-)

Bratislav