Ed Jones
Pooh-Bah
Reged: 04/06/04
Posts: 1414
Loc: Sin-sin-atti
|
|
I was able to model the water test (without using a lens) in Zemax. I measured my viewing distance at 49 inches and knowing the water thickness Zemax computed 8.3 fringes, good agreement with my fringe count. The picture below is the computed interferogram at 49 inches on the left and at 10 feet away on the right for my 8 inch mirror. If you wanted to make a big flat for autocollimation and didn't have a big lens for testing you could use Zemax to compute what the fringe pattern should be.
--------------------
Ed Jones
|
Gordon Rayner
scholastic sledgehammer
Reged: 03/24/07
Posts: 965
|
|
Someone , somewhere has pointed out that autocollimation "flats" do not need to be perfectly flat, just smooth mirrors of very long radius of curvature.
|
Dick Parker
sage
Reged: 08/17/07
Posts: 247
|
|
Yes, it has been documented true that an autocollimation flat does not have to be flat. Malacara, in Optical Shop Testing, appendix 4 gives the following formula d=128*(F/D)^2*e, where d is the depth of the sagetta of a non flat autocollimation flat in fringes and e is the error you will get in your tested mirror in fringes. Essentially, if your flat is not flat, you will, in effect, be testing with a light source closer to or further away than infinity, so you will get an ellipse or a hyperbola. You can be hundreds of fringes off (depending on the focal ratio of the mirror you are testing and the final error in the parabloid you desire)
I suggest you consider that "being smooth" means on hills, holes, zones, edge problems, aspheric surfaces etc. As a practical matter, using collimated light against a good reference flat gives intereference bands that are straight and parallel when the optic being tested exactly matches the reference, in this case a known flat. This becomes a null and errors are easy to detect.
When you are a several fringes off, you have a bull's eye and rings. Although you can count the fringes, that is not all. It is the progression distance of each successive ring from its adjacent one and the shape of the bull's eye that will help determine if your deviation from flat is a true spherical deviation. Also, as you press on one side and the bull's eye moves off center to one side, that circle should remain round. If it changes to an oval or a pear or something else, you have zones or an aspheric deviation from a pure spherical deviation from flat.
I suggest that it is so much easier, in the long run, to make the investment in time and effort to make the slight effort to get collimated light, figure the flat as flat as you can. Don't accept excuses and therefore get results from your autocollimation flat you can be confident in.
My 2 cents
Dick Parker
|
Mark Harry
Post Laureate
Reged: 09/05/05
Posts: 3120
Loc: Northeast
|
|
In the formula above, I agree wholeheartedly with Dick's assessment of flat quality. Make it the best you can. In hypothetical situations, you'll see that a flat that's only 1-2 fringes off will allow you to very closely assess the quantification aspects of overall error -RELIABLY-.
For the guys who want just what may be "good enough", go ahead and use a flat that's just a long sphere. But don't expect everyone whom may be savvy enough to test reliably to believe it!
*************
A question for Ed- does that ZEMAX representation make the flat appear convex in fit?
Mark
--------------------
So many projects, so little time!
|
Coletti
newbie
Reged: 06/12/09
Posts: 4
Loc: Brazil - Curitiba city
|
|
If the intention is to make the collimation lens for Rayleigh test:
What is the accuracy needed in the surface?
Consider a soda-lime glass, 1 "thick, the bending of the lens causes problems in water test?
I need to make an 12" optical flat, so need a collimating lens of D=13" at least, right? The focus of these collimation lens can be short than F/D5?
Polishing a lens so should be very difficult if you need high-precision surface...
--------------------
Best regards,
Sandro Coletti - Brazil
http://sandrocoletti.multiply.com/
|
Ed Jones
Pooh-Bah
Reged: 04/06/04
Posts: 1414
Loc: Sin-sin-atti
|
|
Mark,
Without a lens the fringes appear concave. By making the test flat a very long radius convex sphere I can make the fringes straight. If anyone wants this Zemax file I can share it. I'm convinced that this can be a workable method of making a flat with an ordinary light box if you allow enough viewing distance so as not to have too many fringes and use Zemax to compute the fringe pattern.
--------------------
Ed Jones
|
Dick Parker
sage
Reged: 08/17/07
Posts: 247
|
|
Sandro -
When I made my 12.5 inch autocollimation flat I (temporarily) suspended my 12.5 inch F/5 telescope from the ceiling pointing down as my collimation light source. That gave me full aperture water test with good null fringes at no cost and probably an aftenoons worth of fussing around.
Dick Parker
|
Dick Parker
sage
Reged: 08/17/07
Posts: 247
|
|
If I understand Ed's idea, knowing the water depth and viewing distance, if you produce a null (straight fringes), you will end up with a flat that is only a few rings convex which would be acceptable flatness for an autocollimation flat. By nulling in this condition, you assure that the surface is smooth. This sounds like a very interesting and workable idea.
My earlier discussion was aimed at the frequent comments that an autocollimation flat does not have to be flat, but does have to be smooth. I wanted to address the relative difficulty of assessing smoothness if you are off flat by several rings. (you could use the Ritchey - Common test to sort that out, but that is a bit OT).
Need not overestimate the difficulty of obtaining a collimated light source. All you need is a Newtonian telescope, which everybody has, and you have a Fizeau interferometer of adequate quality at no cost, and only a bit of ingenuity to get it to work.
Dick Parker
|
Ed Jones
Pooh-Bah
Reged: 04/06/04
Posts: 1414
Loc: Sin-sin-atti
|
|
Dick,
It might be more than a few fringes depending on how big the flat is and how close you view it. You can use Zemax to calculate the number of fringes you would see for a given distance away. For example a 30 inch flat with 1 mm of water would show 5.4 fringes viewed at 20 feet but 71 fringes viewed at 6 feet away! A 30 inch primary would be work fine as a collimator but would be hard to come by.
--------------------
Ed Jones
|
Gordon Rayner
scholastic sledgehammer
Reged: 03/24/07
Posts: 965
|
|
Are there now diode lasers stable enough , and with sufficient coherence length, at low cost, to supplant/replace HeNe lasers in replacing the old light sources in the Davidson Fizeau planinterferometers?
I have seen diffusing adhesive tape, or ground plastic or groundglass transmission windows, some slowly rotating, as speckle eliminators for HeNe replacements of the original lightsources.
Would such lasers be applicable to the water test, (with pinholes, and/or beam expanders, beamsplitters, video cameras, or other appropriate items)?
|
Ed Jones
Pooh-Bah
Reged: 04/06/04
Posts: 1414
Loc: Sin-sin-atti
|
|
Many monochromatic light sources will show fringes but a point source would be harder to use than larger broader source IMHO.
Another useful application of this test is testing diagonal mirrors. However the fringes are very difficult to see with an aluminized mirror, 40 to 1 contrast. You can add a strong food coloring solution to absorb the reflection off the aluminum or spray on a weak mixture of cerium oxide which settles out on the mirror and acts like a filter. I look at the reflections of a green laser and try to get both reflections equally bright. You can keep spraying on more cerium to knock down the mirrors reflectivity. If you get both reflections the same the contrast is as good as in air. The cerium won't hurt the aluminum.
--------------------
Ed Jones
|
refractory
Pooh-Bah
Reged: 02/05/05
Posts: 1233
|
|
As some of you already know, I have a 26 inch optical comparator mirror from an old 50 inch Jones and Lamson unit. The surface has a shallow, narrow winding scratch on one side, and a couple of tiny nicks here and there, but otherwise nearly flawless at least to the eye. There is a deep conchoidal chip on the back starting at the edge, where there was a drilled hole that I guess originally held a pin to hold the mirror on its mount.
I did some star tests last year with a Celestron CR6 refractor, and was able to get pinpoint stars at many magnifications at high angles, though above/below focus the images were ellipsoidal- possibly due to the lack of proper support of the mirror?
Anyway, I don't have proper equipment here to do testing- could the water method work, or would I risk damage to the mirror surface? Or would this even work with a mirrored surface?
Thanks,
Jess Tauber
|
Mark Harry
Post Laureate
Reged: 09/05/05
Posts: 3120
Loc: Northeast
|
|
There's an easy way to see the IF pattern when dealing with a coated flat on an interferometer. I used to peel one layer of a Kleenex off, and drape it on the secondary. It's a bit mottled from the random density of fibers, but it makes the pattern easy to see, with no liquids or abrasives on the tested glass. (the same idea as the cerium on the glass)
Mark
--------------------
So many projects, so little time!
|
Coletti
newbie
Reged: 06/12/09
Posts: 4
Loc: Brazil - Curitiba city
|
|
Quote:
Sandro -
When I made my 12.5 inch autocollimation flat I (temporarily) suspended my 12.5 inch F/5 telescope from the ceiling pointing down as my collimation light source. That gave me full aperture water test with good null fringes at no cost and probably an aftenoons worth of fussing around.
Dick Parker
Hello Dick
This is what you did?
http://spie.org/x13539.xml?ArticleID=x13539
In the article, the author says to use an off-axis Parabolic mirror, If you used a common Parabolic mirror, then it is easier. Great tip, thanks.
--------------------
Best regards,
Sandro Coletti - Brazil
http://sandrocoletti.multiply.com/
|
Ed Jones
Pooh-Bah
Reged: 04/06/04
Posts: 1414
Loc: Sin-sin-atti
|
|
A big reflective mirror would be hard to learn on. I'd suggest learning on a small uncoated mirror first. Getting and even density of dye or cerium over a 26 in. mirror might be tricky.
Here's a diagonal I tested yesterday, you can see variations from the cerium nonuniformity. Still works though. I should have used a green filter too.
--------------------
Ed Jones
|
greenglass
professor emeritus
Reged: 01/22/06
Posts: 540
Loc: Ontario, Canada
|
|
How is this for flatness? It's 3.5" of a 4" flat. Ronald
--------------------
10" f/5 dob unf.
7x50 Tasco binos
Edited by greenglass (07/08/09 04:31 PM)
|
Ed Jones
Pooh-Bah
Reged: 04/06/04
Posts: 1414
Loc: Sin-sin-atti
|
|
Its astigmatic. Looks like you clipped it a little too. Did you use the water test? If so each fringe is 3/8 wave. Fringes on a good flat are straight, parallel and spaced the same.
BTY I found a fluorescent light fixture at Wally World for $5 and the F4T5/BL bulb at replacementlightbulbs.com.
--------------------
Ed Jones
Edited by Ed Jones (07/08/09 05:30 PM)
|
greenglass
professor emeritus
Reged: 01/22/06
Posts: 540
Loc: Ontario, Canada
|
|
Yes it's the same water tester, so that's why the fringes fan out, astigmatism, that would explain the straight fringes and curved when they're at another angle across it. Ronald
--------------------
10" f/5 dob unf.
7x50 Tasco binos
Edited by greenglass (07/08/09 07:18 PM)
|
Gordon Rayner
scholastic sledgehammer
Reged: 03/24/07
Posts: 965
|
|
I want to test some aluminized diagonal flat mirrors. I have never used my Davidson Optronics 5 inch planinterferometer. My limited knowledge about Fizeau testing has been dormant for over five years.
It is inconvenient to unpack the Davidson at this time to experiment. I had determined to sell it, and probably shall do so, but this thread has re-interested me .
Please confirm/refute/comment:
Using a laser source increases the coherence length, so fringes can be obtained at much greater spacing, between the reference transmission window with one flat surface and the element being tested(ff. ZYGO (?))?. Diffuse filtering screens can be placed in this conveniently increased gap(?).
Accordingly, to avoid draping light filtering Kleenex on the aluminized mirror(which is probably not a problem),to reduce the reflected light and thus increase fringe contrast, could one use one or more mosquito netting screens above, and not touching, the flat mirror being tested? Any Moire patterns caused by using two or more screens(whose screen patterns would be not parallel, unless great pains were taken to make them not rotated with respect to each other) for this filtration could be ignored (?).
|
Ed Jones
Pooh-Bah
Reged: 04/06/04
Posts: 1414
Loc: Sin-sin-atti
|
|
Gordon,
Yes a HeNe laser has has a long coherence length, many meters. Mosquito netting might work, I was told nylon panty hose was one solution. I tried it many years ago and I believe it worked OK, not as nice as your typical interferogram, I don't remember any Moire problems but there is a lot of scattered light. Other solutions would be to put a partially transmissive coating on your reference flat, say 30% if you check mostly coated mirrors. Another slightly more expensive option is a partially transmittiong pellicle, they are fragile but work great.
--------------------
Ed Jones
|
Previous
Index
Next
Threaded
[Re: 
Edit
Reply
Quote
