From: "Dwight K. Elvey" 
To: atm{at}shore.net
Reply-To: "Dwight K. Elvey" 


>From: Arthur 
>
>
>I'm putting together the testing devices for my Foucault tester. I have the
>sliding base, which follows the normal design for this device. I would like
>to make a multi-tester,i.e., slitless, ronchi, and possibly the wire test. I
>have looked at several web pages with testers, which use mini mag-lite bulbs
>and various colored LED's. All seemed good,however, one developer said that
>some colors of LED's have an affect on reading certain zones when measuring
>(?). What size of LED (mcd) should I use. I have a orange LED that is 1200
>mcd. if I can use this, what amount of resistance is need with 3 volt.

Hi
 The LED's towards the blue end produce sharper shadows but I know
several that use red LED's on their testers. Orange should work OK. The
main issue is to have a uniform light source over the angle that the source
edge of slit forms as a cone of light, across the face of the mirror. This
is just a simple f/ratio thing. Many cut the end from the LED and then
place frosted glass or magic mending tape ( 3M name ) to get a uniform
source. The source needs to be close enough to the source edge to get
uniform light but not so close that surface detail is also focused by the
resolving edge. You only want the source slit or edge to be the defining
image and no interfered with by source issues.
 You need to know the running voltage and current of your LED to
determine the size of the resistor. It is just Ohms law:

 3v

 ( 3 - Vf ) / MaxCur = R

 You'll get better consistency light if you use 3 cells with
4.5V instead of just two. This give a little more drop across the resistor.
This tends to regulate better. It also allows my method of making a
variable source easier. Rather than putting a variable resistor in series
with the current limit resistor, I like to add another fixed resistor and
variable resistor in parallel to the LED. This does draw a little more
current but I get better light control. One needs to experiment with the
particular LED, that they have, to select the best combination.

>
What type of light source should I use that will satisfy the requirements of
>each test method? I know every little about electronics, how do I figure the
>amount of resistance to use a LED  powered with a 3v battery. (can someone
>send me a diagram of what they use?)  I want to be able to test several
>blanks sizes, does this mean that I will need a brighter light source for
>the larger glass?

 The best source for the Ronchi is something with a large surface
area. I use a jumbo green LED from RS. Edge and wire test don't require as
large a source. For most typical mirrors, the F/ratio is high enough the
there is no need for different intensities ( unless you expect to do f/2's
and also f/30's at with the same setup ).
 The Ronchi should technically have the glass side towards you for
a minimum effect of the glass. The gratings from Edmond have good quality
glass and I've not seen enough difference that I feel the need to check
each time I set it up.

>I have a ronchi screen that I bought some years ago, and can't remember who I
bought it from.(could have been WillBell, or Edmonds?of) It's is made of
glass (almost looks tinted) proximately 1.5 x 1.5x .25 thick. Will the
thickness of this ronchi screen and then changing to knife edge have an
effect on the test?

 Yes and no. It does cause the distance from the tester edge to the mirror
to be slightly different. As for measuring, we are looking at delta value
and not absolutes. The thickens has no effect. Dwight

>
>

--- BBBS/NT v4.01 Flag-4