Ed Jones
Carpal Tunnel
Reged: 04/06/04
Loc: Sin-sin-atti
|
Re: AC and other null testing--was: Re: 6"f10 build
[Re: MKV]
#5159048 - 04/06/12 11:43 AM
|
Edit
|
Reply
|
Quote |
Quick Reply
|
|
|
Mladen,
There was an article written years ago in S&T Gleanings. More recently Stellafane has a video by Dick Parker on autocollimation testing here. It's a long video but very good and he describes the test, (correct spelling is Spoelhof BTW). I just added a null lens to the test. With a null lens any mirror can be tested including the ultra-fast f/3s.
The lens spacing is determined by ray tracing. I don't have the original article so I don't know if there is a mathematical model; I think not. There are an infinite number of possible combinations of lenses and spacings. Somewhere I had an Excel file with some setup info. The lens shortens the BFL of the mirror and partly offsets the lengthening due to the light source not at infinity. Tolerances are not nearly as tight as in the Ross since the lens is weaker.
|
MKV
Carpal Tunnel
Reged: 01/20/11
|
Re: AC and other null testing--was: Re: 6"f10 build
[Re: Ed Jones]
#5159056 - 04/06/12 11:48 AM
|
|
Edit
|
Reply
|
Quote |
Quick Reply
|
|
|
Quote:
Mladen,
There was an article written years ago in S&T Gleanings. More recently Stellafane has a video by Dick Parker on autocollimation testing here. It's a long video but very good and he describes the test, (correct spelling is Spoelhof BTW). I just added a null lens to the test. With a null lens any mirror can be tested including the ultra-fast f/3s.
The lens spacing is determined by ray tracing. I don't have the original article so I don't know if there is a mathematical model; I think not. There are an infinite number of possible combinations of lenses and spacings. Somewhere I had an Excel file with some setup info. The lens shortens the BFL of the mirror and partly offsets the lengthening due to the light source not at infinity. Tolerances are not nearly as tight as in the Ross since the lens is weaker.
Thanks for clarying this, Ed. It's quite ingenious. And, sorry for misspelling Spoelhof.
Mladen
|
MKV
Carpal Tunnel
Reged: 01/20/11
|
Re: AC and other null testing--was: Re: 6"f10 build
[Re: DAVIDG]
#5159106 - 04/06/12 12:17 PM Attachment (13 downloads)
|
|
Edit
|
Reply
|
Quote |
Quick Reply
|
|
|
Thanks for the OSLO file issue. As for the wavefront, my wavefrotn analysis is set to minimum RMS. And this is what I get when I change the radius from -206 to -207.
However, if I change the wavefront analysis to best focus I get the results on the bottom. Pretty radical, isn't it?
Any thoughts?
Mladen
|
kfrederick
Carpal Tunnel
Reged: 02/01/08
|
Re: AC and other null testing--was: Re: 6"f10 build
[Re: MKV]
#5159160 - 04/06/12 12:45 PM
|
|
Edit
|
Reply
|
Quote |
Quick Reply
|
|
|
Radius of the[ mirror zone tested] squared devide by radius of the curve= knife edge movement fixed light .Aint no easer than that . Say a 20 inch f5 100/200=.5 Null the center of the mirror move the knife edge back .5 inch null the edge. The ronchi could be thought of a null with curve lines . 
|
Ed Jones
Carpal Tunnel
Reged: 04/06/04
Loc: Sin-sin-atti
|
Re: AC and other null testing--was: Re: 6"f10 build
[Re: MKV]
#5159164 - 04/06/12 12:46 PM
|
|
Edit
|
Reply
|
Quote |
Quick Reply
|
|
|
No I'm the one who misspelled it. 
|
DAVIDG
Post Laureate
Reged: 12/02/04
Loc: Hockessin, De
|
Re: AC and other null testing--was: Re: 6"f10 build
[Re: MKV]
#5159480 - 04/06/12 04:30 PM
|
|
Edit
|
Reply
|
Quote |
Quick Reply
|
|
|
Mladen,
Your not refocusing after you make the changes but when you switch to "Best Focus" for the wave front results that forces the refocusing. Go to the IMG row and click on the box next to THICKNESS and a text box will open and allow you to refocus the system. After refocusing you get very good results.
When you do the test you set the spacing between the lens to mirror then you setup the tester and find the focus point just like testing miror via Foucault.
In your first OLSO setup your keeping the focal plane fixed behind the lens at the original value. If you actual trying doing the test this way, you would find out the knife edge is nowhere near the focus point just like when you doing Foucualt test. What you have calculated is the wavefront error in the defocused image and of course it would have many waves of errors.
The test is actually fairly rebust and RossNull and OSLO will show that you can be off in the spacing between the lens and the mirror by about a total of 10mm (5mm on each side of the perfect spacing) and the wavefront of the will be good to 1/8 wave.
- Dave
|
GlennLeDrew
Postmaster
Reged: 06/18/08
Loc: Ottawa, Ontario, Canada
|
Re: AC and other null testing--was: Re: 6"f10 build
[Re: DAVIDG]
#5159536 - 04/06/12 05:12 PM
|
|
Edit
|
Reply
|
Quote |
Quick Reply
|
|
|
I'm delighted to have skipped over the Foucault test and got directly to using both the Ross null and autocollimation. For the 12" f/2 ellipses I've figured, the process from sphere to finished optic has often been:
- half-diameter test plate with radius equal to the mirror's 70.7% zone (fringe 'counting', where max depression is 22 fringes w.r.t. the neutral zone),
- Ross null, and
- autocollimation in a complete system built to facilitate rapid primary in-and-out.
The thought of using a zonal mask and calculating the shape almost causes me to break out in hives. 
|
MKV
Carpal Tunnel
Reged: 01/20/11
|
Re: AC and other null testing--was: Re: 6"f10 build
[Re: DAVIDG]
#5160371 - 04/07/12 09:10 AM Attachment (18 downloads)
|
|
Edit
|
Reply
|
Quote |
Quick Reply
|
|
|
Dave, looking at your Ceravolo example I am wondering about the Ent beam radius value of 24.569429. The object is a point source (or it should be). As such its entry diameter should be much smaller, something to the order of 0.025 mm instead of 25 mm.
Maybe my schematic drawing can help you visualize what I am talking about.
Hoping to get to the bottom of this,
Mladen
PS Btw, my Google Chrome still opens OSLO attachments on this forum as text files.
|
Dave O
sage
Reged: 12/21/11
Loc: Sri Lanka
|
Re: AC and other null testing--was: Re: 6"f10 build
[Re: MKV]
#5160634 - 04/07/12 11:57 AM
|
|
Edit
|
Reply
|
Quote |
Quick Reply
|
|
|
Quote:
PS Btw, my Google Chrome still opens OSLO attachments on this forum as text files.
Yes, I have the same problem (using Safari). To get around this I right click on the attachment and then select 'Download Linked File' from the list of options. When the file is downloaded I can click on it and open it in OSLO. Dave O
|
DAVIDG
Post Laureate
Reged: 12/02/04
Loc: Hockessin, De
|
Re: AC and other null testing--was: Re: 6"f10 build
[Re: MKV]
#5160756 - 04/07/12 01:10 PM
|
|
Edit
|
Reply
|
Quote |
Quick Reply
|
|
|
Mladen,
The aperture radius of the light source is listed as 10-7 so it is a pin hole source. The Ent Beam value is fine since the light source is not located at infinity but only a few inches away, so it is a divergent source that needs to fully illuminate the diameter of the activate area of the Ross lens used which will fully illuminate the surface of the mirror. If you change the Ent Beam radius to a smaller value and look at the drawing of the optical layout, you'll see that the beam leaving the lens does not fully illuminate the mirror.
All the Best,
- Dave
|
MKV
Carpal Tunnel
Reged: 01/20/11
|
Re: AC and other null testing--was: Re: 6"f10 build
[Re: DAVIDG]
#5161657 - 04/08/12 02:27 AM Attachment (13 downloads)
|
|
Edit
|
Reply
|
Quote |
Quick Reply
|
|
|
Quote:
Mladen,
The aperture radius of the light source is listed as 10-7 so it is a pin hole source. The Ent Beam value is fine since the light source is not located at infinity but only a few inches away, so it is a divergent source that needs to fully illuminate the diameter of the activate area of the Ross lens used which will fully illuminate the surface of the mirror. If you change the Ent Beam radius to a smaller value and look at the drawing of the optical layout, you'll see that the beam leaving the lens does not fully illuminate the mirror.
Thanks a lot, Dave. I understand what you are saying but I don't understand why my setup doesn't work.
I created an additional optic - a "perfect paraboloid" - to form a perfect point source and illuminate the Ross lens as required, i.e. forming the image at the exact object to lens distance, and to the exact working radius on the lens.
In theory this should work. But it doesn't work in OSLO - at least not in this case! I used the same point-source setup in the autocollimation mockup and it wortks like a dream, so I have reason to believe that something is not right with the Ross setup.
Another enigma here is that although the setup shows a wavefront P-V of about lambda/10 the image is never focused!
In this particular case, chromatism seems to be the issue. If you reduce the wavelengths to one, you get sharply focused images.
Regrds,
Mladen
Edited by MKV (04/09/12 09:15 AM)
|
MKV
Carpal Tunnel
Reged: 01/20/11
|
Re: AC and other null testing--was: Re: 6"f10 build
[Re: MKV]
#5161664 - 04/08/12 02:38 AM Attachment (49 downloads)
|
|
Edit
|
Reply
|
Quote |
Quick Reply
|
|
|
And here is my autocollimation setup.
Mladen
|
DAVIDG
Post Laureate
Reged: 12/02/04
Loc: Hockessin, De
|
Re: AC and other null testing--was: Re: 6"f10 build
[Re: MKV]
#5162121 - 04/08/12 12:38 PM
|
|
Edit
|
Reply
|
Quote |
Quick Reply
|
|
|
Mladen,
I took a quick look and noticed that you have the spacing for the parabola that forms the pin hole set at 326mm which is the ROC and not the focal length if the light source feeding the parabola is at infinity. Also when I downloaded your AC setup I noticed that OLSO has the units set to inches and it looks like your working in millimeters.
All the Best,
- Dave
|
Ed Jones
Carpal Tunnel
Reged: 04/06/04
Loc: Sin-sin-atti
|
Re: AC and other null testing--was: Re: 6"f10 build
[Re: DAVIDG]
#5162472 - 04/08/12 05:09 PM
|
|
Edit
|
Reply
|
Quote |
Quick Reply
|
|
|
Could a laser provide enough light to allow for 3 uncoated surfaces? If so you could use an oil flat and test vertically. It might require a creative way of holding the mirror upside down. 
|
Gary Fuchs
Pooh-Bah
Reged: 05/22/06
Loc: Easton, PA, USA
|
Re: AC and other null testing--was: Re: 6"f10 build
[Re: Ed Jones]
#5162601 - 04/08/12 06:54 PM
|
|
Edit
|
Reply
|
Quote |
Quick Reply
|
|
|
Ed,
What kind of oil would you use?
And would that be with viewing through the center?
Gary
|
Ed Jones
Carpal Tunnel
Reged: 04/06/04
Loc: Sin-sin-atti
|
Re: AC and other null testing--was: Re: 6"f10 build
[Re: Gary Fuchs]
#5162703 - 04/08/12 08:32 PM
|
|
Edit
|
Reply
|
Quote |
Quick Reply
|
|
|
Gary,
Mineral oil or cooking oil would work and sure you could make a tray with a center viewing port. Surface tension looks to be less than half that of water. Viewing so low might be hard; may need to use a camera. I can see 3 reflections off glass with a green laser.
|
MKV
Carpal Tunnel
Reged: 01/20/11
|
Re: AC and other null testing--was: Re: 6"f10 build
[Re: DAVIDG]
#5162840 - 04/08/12 10:25 PM
|
|
Edit
|
Reply
|
Quote |
Quick Reply
|
|
|
Dave, I updated the picture for clarity. I think it's pretty clear that the point source is at 163.0301 mm. This is no different than placing the object (point source) at that distance in the OBJ section of the OSLO data entry and setting the distance to 163.301 mm.
The mm units is a default setting so I believe I had it set to mm. At any rate, I re-entered the data to make sure. However, I still get the "blobs" for images that cannot be focused, despite the fact that my wavefront error P-V is at about lambda/10.
I have both an autocollimation setup and a spehere at RoC null setup with auxilliary mirror light source and they both work (they also work by setting the object distance), which makes the Ross setup somewhat enigmatic because it seems to work fine when setting the object distance but not with the auxilliary mirror light source.
Mladen.
|
Dave O
sage
Reged: 12/21/11
Loc: Sri Lanka
|
Re: AC and other null testing--was: Re: 6"f10 build
[Re: MKV]
#5162948 - 04/09/12 12:07 AM Attachment (25 downloads)
|
|
Edit
|
Reply
|
Quote |
Quick Reply
|
|
|
I find this whole discussion to be very interesting -- thank you to all who have contributed.
I like Mladen's method for generating a point source. Using his method, I went back and set up the original test using the data that Dave G. had provided (testing of a 200mm f/4 paraboloid). Using the spacings provided resulted in an error of ~7.5 lamda; however by refocusing (I chose to minimize RMS spot size ...) the focus was moved back ~0.63 mm and the P-V error was ~ 1/85 lambda.
(The OSLO file is attached.)
Mladen's latest example involves a different mirror and set-up parameters; and therefore I am unable to comment.
Dave O
|
MKV
Carpal Tunnel
Reged: 01/20/11
|
Re: AC and other null testing--was: Re: 6"f10 build
[Re: Dave O]
#5163231 - 04/09/12 08:49 AM Attachment (16 downloads)
|
|
Edit
|
Reply
|
Quote |
Quick Reply
|
|
|
Thanks Dave O. The auxilliary mirror method simulates true point source more accurately than the object method, but the latter may be more realistic because we can't generate a true point source. In either case, both methods work (see attached file), or should work. I think I have resolved the problem I had with the two methods not giving me the same results. OSLO is sometimes not the most intuitive program, but hey who's complaining - it's a free gift to the amateur community!
Mladen
|
DAVIDG
Post Laureate
Reged: 12/02/04
Loc: Hockessin, De
|
Re: AC and other null testing--was: Re: 6"f10 build
[Re: MKV]
#5163367 - 04/09/12 10:32 AM
|
|
Edit
|
Reply
|
Quote |
Quick Reply
|
|
|
I've used the Ross Null for the last 10 years and I find it to be an excellent test. My main test is double pass Autcollimation but the Ross has it's place and it is an easy test to do. It works well for other conics like ellipse and hyberbolas. As I have been preaching for years, use a couple of different tests and be sure that they agree. If not find were the problems are.
The RossNull XP programs makes calculating the setup parameters as simple as can be. Also the program does allow other lens types besides plano convex so there are many possible lenses that can be used. Since it also has the export function to OSLO, one can farther explore the setup easily. Too many people get hung up on the quality of the lens that they believe that is needed and the cost. Surplus Shed and others sells precision plano and double convex lens can be used at very reasonable prices so one can purchase a couple of lens and test them. If one purchase a large diameter lens say at 3" in diameter, most of the time only the middle 1/3 of the lens is being used. So the odds are very good that over that area the lens as an excellent figure but again it is not difficult to test it. You can test the back surface against a flat. It doesn't need to perfectly flat just optically smooth and from the number of fringes calculate radius and enter that in the Ross Null XP. For the convex side you can hand hold the lens and look at bright star and examine the image with Ronchi screen. If the lines bow smoothly the lens doesn't have any zones. If you have to get more critical you can make a test plate and test by interference. As for setting the distance, the RossNull Xp will show you the tolerances and many times it +/- 1/4" or more to still maintain 1/8 wave. It is not difficult to set the distance to a much great accuracy by using a wooden stick cut the correct length.
All the Best,
- Dave
|
Previous
Index
Next
Threaded
