Gary Fuchs
Pooh-Bah
Reged: 05/22/06
Loc: Easton, PA, USA
|
AC and other null testing--was: Re: 6"f10 build
#5154834 - 04/03/12 02:42 PM
|
Edit
|
Reply
|
Quote |
Quick Reply
|
|
|
Just thought I'd point out that at least to my knowledge, no one who advocates the AC test--or Ross Null, or other--has said that you shouldn't/can't also use a regular Foucault knife edge or Ronchi screen to look at your mirror.
For example, at the recent Delmarva mirror weekend a Foucault tester was set up and used--occasionally--just a few feet from the AC tester and upstairs from the Ross Null.
If you're making mirrors professionally (time is money?) (or in a situation like the Delmarva weekend where there's just not enough time to look at zones for 16 or so mirrors) it's hard to see how you can justify not using a faster, less error-prone method; and also the most appropriate tests for whatever you're checking for.
Gary
|
Pinbout
Post Laureate
Reged: 02/22/10
Loc: Montclair
|
Re: AC and other null testing--was: Re: 6"f10 build
[Re: Gary Fuchs]
#5154963 - 04/03/12 04:11 PM
|
|
Edit
|
Reply
|
Quote |
Quick Reply
|
|
|
I thought the angle alum. taped together and used as connectors was a very smart for a quick portable setup.
Dick Parker's 12in Autocollimator
|
Mirzam
Post Laureate
Reged: 04/01/08
Loc: Lovettsville, VA
|
AC and other null testing--was: Re: 6"f10 build
[Re: Gary Fuchs]
#5154971 - 04/03/12 04:18 PM
|
|
Edit
|
Reply
|
Quote |
Quick Reply
|
|
|
Boy, I'll second that. My 15.5" mirror literally would have taken days, maybe weeks, of cycling back and forth between Foucault measurements, software analysis, more figuring, more Foucault etc., to accomplish what was done in a single afternoon with Ross Null--even when sharing the tester with about 8 other guys! Of course a big part of the reason this works is Steve Swayze and Dick Parker, who are the fastest guns in the west when it comes to setting up the Ross Null and AC tests.
JimC
|
Pinbout
Post Laureate
Reged: 02/22/10
Loc: Montclair
|
Re: AC and other null testing--was: Re: 6"f10 build
[Re: Mirzam]
#5154990 - 04/03/12 04:30 PM
|
|
Edit
|
Reply
|
Quote |
Quick Reply
|
|
|
Quote:
Of course a big part of the reason this works is Steve Swayze and Dick Parker, who are the fastest guns in the west ...
What about trimming polishers. 
|
MKV
Carpal Tunnel
Reged: 01/20/11
|
Re: AC and other null testing--was: Re: 6"f10 build
[Re: Gary Fuchs]
#5155146 - 04/03/12 06:40 PM
|
|
Edit
|
Reply
|
Quote |
Quick Reply
|
|
|
Gary, you are spot on: no test is off limits, of course. But you hit the nail on its head when you called the AC test "less error-prone method; and also the most appropriate tests for whatever you're checking for".
Indeed, no other test available to amateurs can test finished, complete telescopes with double-pass precision at the focus. The beauty of it is that the criteria are no different than when testing a single mirror bny this method: you always get straight lines and a KE null, or at least that's what you are aiming at getting. No measurments, no zones, no masks, not getting three different readings by three different operators, etc. It's universal.
Mladen
|
kfrederick
Carpal Tunnel
Reged: 02/01/08
|
Re: AC and other null testing--was: Re: 6"f10 build
[Re: MKV]
#5155799 - 04/04/12 08:27 AM
|
|
Edit
|
Reply
|
Quote |
Quick Reply
|
|
|
I use a 60 ronchi over a white light to do most my testing .I white light works best for me . If you use one band it is a knife edge. Astig .zones. edge. smoothness and the correction very easy to see .
|
Gary Fuchs
Pooh-Bah
Reged: 05/22/06
Loc: Easton, PA, USA
|
Re: AC and other null testing--was: Re: 6"f10 build
[Re: MKV]
#5155828 - 04/04/12 09:01 AM
|
|
Edit
|
Reply
|
Quote |
Quick Reply
|
|
|
Thanks Mladen
|
MKV
Carpal Tunnel
Reged: 01/20/11
|
Re: AC and other null testing--was: Re: 6"f10 build
[Re: Gary Fuchs]
#5156043 - 04/04/12 11:39 AM
|
|
Edit
|
Reply
|
Quote |
Quick Reply
|
|
|
I recommend this article by R. F. Royce for anyone interested in learning more how autocollimation (and intefereometry - used mainly by the professional community) relate to the Foucault test, and why ancedotal evidence of knife-edge created "super-optics" is just anecdoctal evidence.
The knife-edge tested "0.99" Strehl mirrors are a toll order because the Strehl is a product of integration, and in order to get that kind of accuracy the KE test would require a mask with zones so small that accurate reading would be impossible.
To clarify this in "common" terms, I believe everyone knows that an accurate a shape can be extrapolated accurately only with many points close to each other, and that using only a few points is a poor approximation of it.
The KE test of a mirror other than a sphere provides only few points (actually slopes) from which the shape is extrapolated mathematically. The problem is that the accuracy and repeatability of KE readings is not very high and often depends on subjective factors.
The only time when the KE test is accurate is when the returning wavefront is spherical (and the slope at all points is the same), a condition known as the optical null. This condition is encountered most often by amateurs in three instances: when testing a sphere at the center of curvature, when autocllimating a paraboloidal mirror at its focus, and when testing complete optical systems (telescopes, cameras, etc.) at their focus. Rarely, a forth case is testing ellipsoids at conjugate foci.
Mladen
Edited by MKV (04/05/12 12:29 AM)
|
Benach
Pooh-Bah
Reged: 01/24/08
|
Re: AC and other null testing--was: Re: 6"f10 build
[Re: MKV]
#5156449 - 04/04/12 05:06 PM
|
|
Edit
|
Reply
|
Quote |
Quick Reply
|
|
|
Mladen: you're forgetting other null tests such as Waineo, Dall null lens etc. There are zillions of null test setups that are sometimes encountered by the amateur that is willing to deviate from KE.
|
MKV
Carpal Tunnel
Reged: 01/20/11
|
Re: AC and other null testing--was: Re: 6"f10 build
[Re: Benach]
#5156959 - 04/05/12 12:06 AM
|
|
Edit
|
Reply
|
Quote |
Quick Reply
|
|
|
Quote:
Mladen: you're forgetting other null tests such as Waineo, Dall null lens etc. There are zillions of null test setups that are sometimes encountered by the amateur that is willing to deviate from KE.
Well, I didn't forget them, Benach, I didn't mention them because they are not universal tests. Why make a system that tests only a specific mirror instead of one that will test a whole range of mirrors and entire optical systems, be they Newtonians, Cassegrains, refractors, catadioptrics, etc?
The Waineo test consists of a mirror being tested and a spherical mirror at least 1/2 the diameter of the mirror being tested, that has to be ground, perforated, polished, figured and aluminized specifically for the mirror being tested! All that so that you can null test your paraboloid...
The Dall/Ross null test is no better in that respect: a different lens matched for every new mirror, unless you plan on mass-producing the same mirror over and over. Unlike the Waineo method, the specific lenses can be purchased as off the shelf items and then the mirror you plan to make can be matched to the lenses at hand instead of the other way around. But that means you can make mirorrs only to match your lens collection. Just so that you can perform a different null test!
For these reasons, I completely fail to see any advantage of the Dall/Ross/Waineo tests over an autocllimation flat which can be used on any paraboloidal mirror of any focal length. You amke the AC once and then you can test hundreds of different mirrors with it!
That being said, the autocllimation has its limits. It cannot test conics that are not paraboloids. Spheres are null tested at their centers of curvature, and ellipsoids, prolate or oblate, at their conjugate foci using the ordinary Foucualt setup.
For hyperboloids things are a little more complicated. As far as I was able to find, the Waineo test is inetended for paraboloids and is not used for hyperboloids. If they are, I'd appreciate a link. The best method for hyperboloids is the Paltzek-Gaviola test which is not a null test, but a Caustic focus integration method that permits highly accurate and repeatable assessment of the surface tested to about 1/100 wave.
Mladen
|
Benach
Pooh-Bah
Reged: 01/24/08
|
Re: AC and other null testing--was: Re: 6"f10 build
[Re: MKV]
#5157090 - 04/05/12 04:16 AM
|
|
Edit
|
Reply
|
Quote |
Quick Reply
|
|
|
Okay Mladen, I will give the link in a few days. But you're mistaken about the Waineo (know that one the best). No, you don't make one mirror to test another one. You can test multiple shapes and sizes with one mirror.
You have basically two DOFs: The distance between the mirror and the pinhole and the distance between the mirrors. That being said, you can nearly always create a null with such a setup independant of the conicity and relatively not irrelevant of the focal ratio.
The only problem is that you have to know the distances fairly accurately and your Waineo must be really smooth.
A friend of mine made a Waineo and he was able to correct a a paraboloid up to 1/20 lambda PV and a Dall Kirkham's primary with 1/8 PV and conicity within 0.1 if I'm not mistaken. Will ask for the pics in a few days.
|
kfrederick
Carpal Tunnel
Reged: 02/01/08
|
Re: AC and other null testing--was: Re: 6"f10 build
[Re: Benach]
#5157185 - 04/05/12 07:43 AM
|
|
Edit
|
Reply
|
Quote |
Quick Reply
|
|
|
I think of a couder mask with a knife edge as a null test Of sorts .Each set of holes nulls just as a sphere at RC. As for knike readings being much diferent from 3 diferent people who know what they are doing .FALSE They will de the very close .For most ATM we do not have a 20 inch flat .or a 500$ rossnull lens so we have use what we have .
|
Benach
Pooh-Bah
Reged: 01/24/08
|
Re: AC and other null testing--was: Re: 6"f10 build
[Re: kfrederick]
#5157242 - 04/05/12 08:42 AM
|
|
Edit
|
Reply
|
Quote |
Quick Reply
|
|
|
Kevin: then it is not a null. A null acts as a sphere for all zones with the same position of the KE.
As flats or reference optics are a problem to some amateurs, I recommend you to google for flat mirror making. It is not as difficult as it sounds.
|
Gary Fuchs
Pooh-Bah
Reged: 05/22/06
Loc: Easton, PA, USA
|
Re: AC and other null testing--was: Re: 6"f10 build
[Re: kfrederick]
#5157276 - 04/05/12 09:08 AM
|
|
Edit
|
Reply
|
Quote |
Quick Reply
|
|
|
Mladen wrote:
Quote:
The Dall/Ross null test is no better in that respect: a different lens matched for every new mirror
Not sure if I'm understanding correctly but as far as I know the Ross Null test doesn't need a lens matched to the mirror.
Kevin wrote:
Quote:
For most ATM we do not have a 20 inch flat .or a 500$ rossnull lens so we have use what we have .
"Most" ATM's aren't making 20" or larger mirrors like you either! But I agree that AC for large mirrors is likely to be prohibitively expensive for most. However, for relatively large mirrors--which are going to cost a relatively large amount just for glass and coating (including shipping) that $500 for the Ross lens may not be so outrageous when you consider how much time it's likely to save you over knife edge testing. And if you're in business making mirrors--unless you have everything under such good control that you only need to test once or twice--then the investment for either a flat or Ross lens is probably well worth it.
Also, I don't think you need to spend that much for the lens. I'm pretty sure you can use much less expensive glass as long as it's a good lens and you know the glass type, and radius to a sufficiently precise degree.
In any case, there's no question in my mind that ATM's should know how to use the standard Foucault knife edge test and I don't think any of the AC advocates suggest otherwise. In a situation with a lot of people and limited time--like for instance the Delmarva weekend, or Dick Parker's workshops, there might not be time for each participant to learn KE, but they can do that on their own.
Gary
Edited by Gary Fuchs (04/05/12 09:42 AM)
|
DAVIDG
Post Laureate
Reged: 12/02/04
Loc: Hockessin, De
|
Re: AC and other null testing--was: Re: 6"f10 build
[Re: Gary Fuchs]
#5157477 - 04/05/12 11:28 AM
|
|
Edit
|
Reply
|
Quote |
Quick Reply
|
|
|
No matter what optical test you do, what is critical is understanding the type, source and magnitude of the errors that are in the result. Assuming you have no errors will lead to inaccurate results and inflated wave ratings. Also one has to understand the difference between precision in a measurement system vs accuracy since many times these are confused. Just because you get the same results doesn't mean you have the correct results.
All the time I see people report very high wave rating from knife edge measurements but I never see anyone report what the errors are in those measurement to justify the results. As a research engineer, when I report a result I have to also report the uncertainity in that measurement or it has no meaning.
In any measurement process one needs to analysis the equipment and the procedure used to do the measurement to determine the errors and then cross check the results with other test methods to determine if they agree.
One of the advantages of a Null test method is that your seeing the whole surface and then figure it until it looks like a perfect sphere. That process removes a fair amount of uncertainity in the results vs other methods BUT the results need to be double checked against other methods to be sure that the results are valid.
When it comes to double pass autocollimation the process has reduced errors since there is nothing to measure and no critical distance to set. You figure the mirror until it nulls and any errors you see on the surface are the same as you see as when figuring a sphere i.e. a hole is hole and a hill is hill and you fix them in a similar manner.
- Dave
|
rwiederrich
Goldfinger
Reged: 11/17/05
Loc: Always Dark skies of Belfair W...
|
Re: AC and other null testing--was: Re: 6"f10 build
[Re: DAVIDG]
#5157529 - 04/05/12 12:04 PM
|
|
Edit
|
Reply
|
Quote |
Quick Reply
|
|
|
Quote:
No matter what optical test you do, what is critical is understanding the type, source and magnitude of the errors that are in the result. Assuming you have no errors will lead to inaccurate results and inflated wave ratings. Also one has to understand the difference between precision in a measurement system vs accuracy since many times these are confused. Just because you get the same results doesn't mean you have the correct results.
All the time I see people report very high wave rating from knife edge measurements but I never see anyone report what the errors are in those measurement to justify the results. As a research engineer, when I report a result I have to also report the uncertainity in that measurement or it has no meaning.
In any measurement process one needs to analysis the equipment and the procedure used to do the measurement to determine the errors and then cross check the results with other test methods to determine if they agree.
One of the advantages of a Null test method is that your seeing the whole surface and then figure it until it looks like a perfect sphere. That process removes a fair amount of uncertainity in the results vs other methods BUT the results need to be double checked against other methods to be sure that the results are valid.
When it comes to double pass autocollimation the process has reduced errors since there is nothing to measure and no critical distance to set. You figure the mirror until it nulls and any errors you see on the surface are the same as you see as when figuring a sphere i.e. a hole is hole and a hill is hill and you fix them in a similar manner.
- Dave
Dave your comment, *a hole is a hole and a hill is a hill* as regards to the AC...when viewed, outward bent ronchi lines indicate a hiil and inward indicate a hole? Since there is no shadow as in Foucault. 
Rob
|
MKV
Carpal Tunnel
Reged: 01/20/11
|
Re: AC and other null testing--was: Re: 6"f10 build
[Re: Benach]
#5157544 - 04/05/12 12:17 PM
|
|
Edit
|
Reply
|
Quote |
Quick Reply
|
|
|
Benach, I found the links, so never mind. That being said, the Waineo method requires very precise alignments and multiple distance settings in order for it to work.
There is also an "ideal setup" for the sphere diameter and focal ratio for each mirror, so while theoretically one Waineo sphere can be used on more than one mirror it would require rearranging the separations accurately for each!
In reality, to accommodate different mirror sizes and focal ratios, one would need to make a bunch of Waineo spehres, and that means ground, cored, polished, figured and aluminized. And with all that you still couldn't test finished telescopes of any design!
How practical is that? Especially when we consider that all you need is one flat to accommodate every optic up to its diameter, and any finished optical configuration of up to the same diameter without measurements and tredious alignments?!
So, why go for such a method when AC, and non-paraboloidal null tests with KE require no measurements whatsoever???
Thinking as a mirror maker, if I had to rley on measurements, I would sooner have a Caustic tester professionally machined than wonder into "Wainoeland".
The Waineo test would be an exercise in futility if it were to be used in a mirror making class where you had a bunch of different mirrors to test, not to talk about testing finished telescopes as well...
Mladen
|
DAVIDG
Post Laureate
Reged: 12/02/04
Loc: Hockessin, De
|
Re: AC and other null testing--was: Re: 6"f10 build
[Re: rwiederrich]
#5157560 - 04/05/12 12:31 PM
|
|
Edit
|
Reply
|
Quote |
Quick Reply
|
|
|
Rob,
If your using a Ronchi screen with Double pass autocollimation, then you interpret the bending of lines just like you do when your testing a spherical mirror. If you use a knife edge again it is just like testing a spherical mirror. If you see a zone, a hole looks like a hole in a spherical mirror and hill looks like like a hill in a sphere. The only difference is that your seeing the defect in reference to a parabola instead of sphere. So any place that it nulls like a sphere, you have the correct parabolic figure.
One other point, if your testing a lens, then everything opposite. For example if it looks like a hole it is a hill.
- Dave
|
MKV
Carpal Tunnel
Reged: 01/20/11
|
Re: AC and other null testing--was: Re: 6"f10 build
[Re: Gary Fuchs]
#5157561 - 04/05/12 12:31 PM
|
|
Edit
|
Reply
|
Quote |
Quick Reply
|
|
|
Quote:
Not sure if I'm understanding correctly but as far as I know the Ross Null test doesn't need a lens matched to the mirror.
The lens focal length is a factor that detemrines the distance in the Dall and Ross tests, and is not entirely arbitrary. Besides, the Dall test requires a substantial angular separation from the lens and the image - which can lead to astigmatism in faster mirrors, and the Ross requires a really high quality lens because of the double-pass nature of the test.
The biggest problem, is establishing accurate and precise spacing between the lens and the mirror to within a mm, with a marign of error much smaller than that.
None of this is needed or required for the autocllimation test.
Mladen
|
MKV
Carpal Tunnel
Reged: 01/20/11
|
Re: AC and other null testing--was: Re: 6"f10 build
[Re: DAVIDG]
#5157578 - 04/05/12 12:39 PM
|
|
Edit
|
Reply
|
Quote |
Quick Reply
|
|
|
Quote:
The only difference is that your seeing the defect in reference to a parabola instead of sphere.
Not really, Dave. You are looking at a wavefront which can be spherical or aspheric. Spherical wavefront is spherical wavefront - regardless if it comes from a paraboloid under AC tests, an ellipsoid at a conjugate focus, a sphere tested at the center of curvature, or a Cassegrain at the focus.
A "perfect" imaging system forms a perfeclty spherical wavefront whose radius of curvature is at the focus. We never see the surface. We simply imply what the surface must be based on the deformation seen in the wavefront and the theoretical basis of the test.
Mladen
|
Previous
Index
Next
Threaded
