Re: Artificial flat fields

[Tom Droege <tdroege@veriomail.com>]

Rob and all,

This has been discussed.  There is a well known way to do it.  I will try 
to describe it but I may get it wrong.  So experts comment.

Take a large sphere.  Paint the inside of it with super reflective 
paint.  Now poke a hole in it and shine a light in it.  Now go off at right 
angles to the entry light and poke a hole.  This makes a uniform point 
source.  Now go some distance away and put up a screen.  The point source 
will evenly illuminate the screen.  Now point the out of focus telescope at 
the screen.

Way #2 (The Norman Molhant scheme).

Make a long black tube.  Larger in diameter than the lens.  At the end away 
from the lens, put in a black wall with a bunch of tungsten lamps on the 
inside face. (Arne and I had a long discussion about this.  LEDs are a 
problem since they tend to be monochrome sources.  I proposed using a bunch 
of different ones, but that still does not seem too good since it would 
produce a spikey spectrum.  Don't suggest white LEDs until you look at the 
spectrum.  Ugh!  )

The lamps shine on a translucent white screen which is the next element in 
the black tube.  With a suitable space with black walls let the screen 
shine on a "white box".  This is two white screens separated by a white 
wall.  The second element white screen shines on one face of the white 
box.  The box is made of high reflectivity translucent material.  Mine is 
made from good white paper.

Since we shine a bunch of lamps on it, the first white screen is pretty 
uniform.  Now its light falls on the white box.  Light that gets in bounces 
around inside many times.  This tends to make the internal light 
uniform.  The face of the white box away from the lighted screen faces the 
lens.

As I say, we built such a box many moons ago.  I will have to try to use it.

Tom Droege

At 02:47 PM 11/7/01 -0700, you wrote:

>Hey folks,
>
>How much thought has been put into using a large number of unlensed 120
>degree led's, plus a diffuser plate?  I'm getting specific 'cause I just
>talked to our of our EE's who creates our scanner illumination cards (and
>camera systems).  These cards must provide even illumination (for ease of
>software use) over a 5.25" fov, for a line scan camera.
>
>His thoughts run akin to dumping large numbers (at .4" spacing or less) of
>led's on a card, going beyond the TASSIV tube diameter by 1" or so (with
>led's) and putting on a diffuser plate of some sort (plastic light diffuser
>from hardware store as a cheap start).  Depending on the distance to the
>system and the results, depopulation of the card may be possible.
>
>This could be useful since any mounting surface could be used (side wall,
>top, etc.).  The real fun part would be soldering all those led's on...
>
>Later,
>Rob