Jump to content

  •  

CNers have asked about a donation box for Cloudy Nights over the years, so here you go. Donation is not required by any means, so please enjoy your stay.

Photo

Classic Telescope Testing

beginner classic collimation optics refractor
  • Please log in to reply
113 replies to this topic

#51 DAVIDG

DAVIDG

    Cosmos

  • *****
  • Posts: 8518
  • Joined: 02 Dec 2004
  • Loc: Hockessin, De

Posted 26 February 2018 - 11:27 PM

A Ronchi screen make it easy to understand for the both the expert and beginner to what is going on ie straight bands is good, not straight not so good.  The double pass test is the use of the optical flat but it doesn't have to use a Ronchi screen to examine the returned image.  You can use a knife edge.

   I don't think that the Ronchi test was invented when the Clark's first started to use the test. What double pass is doing is making an artificial star.  So if the light source is  a pin hole and you use an eyepiece your going to see an image of that artificial star so you can do a star test. 

   I think the drawing  in the article is not what they were doing and not just using an eyepiece but also a knife edge so they were looking for a null. If they did not see one they were seeing the zone errors and then could polish the lens to remove them. 

 

                 - Dave 


Edited by DAVIDG, 27 February 2018 - 09:51 AM.


#52 deSitter

deSitter

    Still in Old School

  • *****
  • Posts: 12129
  • Joined: 09 Dec 2004

Posted 26 February 2018 - 11:51 PM

It's a much different sort of test.

 

https://en.wikipedia...avefront_sensor

 

oops

 

grrr

 

http://www.spot-opti...ques_shtest.pdf

 

-drl


Edited by deSitter, 26 February 2018 - 11:55 PM.


#53 starman876

starman876

    Hubble

  • *****
  • Posts: 18520
  • Joined: 28 Apr 2008
  • Loc: VA

Posted 26 February 2018 - 11:54 PM

It's a much different sort of test.

 

https://en.wikipedia...avefront_sensor

 

oops

 

-drl

oops?



#54 starman876

starman876

    Hubble

  • *****
  • Posts: 18520
  • Joined: 28 Apr 2008
  • Loc: VA

Posted 26 February 2018 - 11:56 PM

A Ronchi screen make it easy to understand for the both the expert and beginner to what is going on ie straight bands is good, not straight not so good.  The double pass test is the use of the optical flat but it doesn't have to use a Ronchi screen to examine the returned image.  You can use a knife edge a 

   I don't think that the Ronchi test was invented when the Clark's first started to use the test. What double pass is doing is making an artificial star.  So if the light source is  a pin hole and you use an eyepiece your going to see an image of that artificial star so you can do a star test. 

   I think the drawing  in the article is not what they were doing and not just using an eyepiece but also a knife edge so they were looking for a null. If they did not see one they were seeing the zone errors and then could polish the lens to remove them. 

 

                 - Dave 

I was wondering the same thing if the drawing was actually describing what Clark was doing.



#55 deSitter

deSitter

    Still in Old School

  • *****
  • Posts: 12129
  • Joined: 09 Dec 2004

Posted 27 February 2018 - 12:18 AM

 

It's a much different sort of test.

 

https://en.wikipedia...avefront_sensor

 

oops

 

-drl

oops?

 

Stupid URL mistakes - that PDF is good for a quick description. This is a hard test to understand.

 

-drl



#56 AllanDystrup

AllanDystrup

    Apollo

  • *****
  • topic starter
  • Posts: 1354
  • Joined: 27 Sep 2012
  • Loc: Denmark

Posted 27 February 2018 - 04:32 AM

 


DPAC with flat and Knife Edge (Null):
G.W.Ritchey, Yerkes Observatory 1903: On methods of testing optical mirrors during construction (p.63ff)

 

Ronchi Grating
V. Ronchi 1923: "Le frange di combinazioni nello studio delle superficie e dei sistemi ottici". Rivista d'Ottica e Meccanica di precisione;   [Combination fringes in the study of surfaces and optical systems], [Journal of Optics and Precision Mechanics], vol. 2, pages 9-35.

Malacara: Optical shop testing: Chap 9: Ronchi Test (p.317ff)
 

Allan


Edited by AllanDystrup, 27 February 2018 - 04:49 AM.


#57 starman876

starman876

    Hubble

  • *****
  • Posts: 18520
  • Joined: 28 Apr 2008
  • Loc: VA

Posted 27 February 2018 - 06:38 AM

now back to the original question. Anyone have a clue how they devised these tests.  Like what thinking process went into developing these tests?



#58 DAVIDG

DAVIDG

    Cosmos

  • *****
  • Posts: 8518
  • Joined: 02 Dec 2004
  • Loc: Hockessin, De

Posted 27 February 2018 - 09:57 AM

 The bottom line is that double pass autocollimation is a  very sensitive test with very few sources of error. In combination with Ronchi screen it is easy for beginner to understand what is going on.  DPAC is the real "magic" in why Clark lens were better then rest .  They had much better test method which allowed them to better figure their optics. 

 

           - Dave 



#59 deSitter

deSitter

    Still in Old School

  • *****
  • Posts: 12129
  • Joined: 09 Dec 2004

Posted 27 February 2018 - 10:18 AM

now back to the original question. Anyone have a clue how they devised these tests.  Like what thinking process went into developing these tests?

Do you mean the Hartmann mask test? Optics is physics, those guys are pretty smart! It's a basic thing in applied math, fitting observed data to develop a model of the stimulus. It's similar in principle to determining the orbit of a comet from a few observations of its RA and DEC.

 

We have Ernst Abbe to thank for all this, I believe he was the first guy to say "Enough with trial and error, let's do some thinking and calculating and save ourselves some trouble!" Abbe had a close association with (or should I say inspired) Fritz Zernicke, whose methods allowed one to separate out the various orders of aberrations in an optic by comparing an ideal pattern to an observed pattern. If one had to characterize all this, it was the transition of applied optics from the realm of geometrical ray tracing to wavefront analysis based on the theory of diffraction.

 

-drl


Edited by deSitter, 27 February 2018 - 10:54 AM.


#60 DAVIDG

DAVIDG

    Cosmos

  • *****
  • Posts: 8518
  • Joined: 02 Dec 2004
  • Loc: Hockessin, De

Posted 27 February 2018 - 12:04 PM

 When it comes to double pass it  comes down to  understanding how a telescope works. It takes parallel light  and brings it to focus.  Stars and other astronomical bodies are so far away that the angle of the light coming from them is parallel. So you want to some how  to make parallel light in the shop.  Understanding that if you place a light source at the focus of  a telescope so it works backward, the light coming out of it is parallel. So now you have a source of parallel light.   To test a telescope you need  another telescope that is producing the parallel light that is as large or larger to fully illuminate the optics your testing. So in the case of the Clarks and the fact that they are making  really big lenses, they would need  another lens of that size or larger. Also because big lens have long focal lengths you would need a long  room to set them up back to back. So it wouldn't take much thought to say " lets use a mirror instead ! That is a  great idea ! "  Now you only need 1/2 the space, you only need to figure the one surface to be flat instead of 4 surfaces in an addition lens and the flat mirror doesn't need to be super flat just optically smooth.  Also the errors in the flat don't directly add to the errors in the lens your testing  while the error in additional lens do. 

 Having worked in  a research lab for 34 years I see this type of thought process happen all the time. 

 

                - Dave 


Edited by DAVIDG, 27 February 2018 - 12:05 PM.

  • Richard Whalen and Piggyback like this

#61 davidc135

davidc135

    Viking 1

  • -----
  • Posts: 621
  • Joined: 28 May 2014
  • Loc: Wales, UK

Posted 27 February 2018 - 01:53 PM

 When it comes to double pass it  comes down to  understanding how a telescope works. It takes parallel light  and brings it to focus.  Stars and other astronomical bodies are so far away that the angle of the light coming from them is parallel. So you want to some how  to make parallel light in the shop.  Understanding that if you place a light source at the focus of  a telescope so it works backward, the light coming out of it is parallel. So now you have a source of parallel light.   To test a telescope you need  another telescope that is producing the parallel light that is as large or larger to fully illuminate the optics your testing. So in the case of the Clarks and the fact that they are making  really big lenses, they would need  another lens of that size or larger. Also because big lens have long focal lengths you would need a long  room to set them up back to back. So it wouldn't take much thought to say " lets use a mirror instead ! That is a  great idea ! "  Now you only need 1/2 the space, you only need to figure the one surface to be flat instead of 4 surfaces in an addition lens and the flat mirror doesn't need to be super flat just optically smooth.  Also the errors in the flat don't directly add to the errors in the lens your testing  while the error in additional lens do. 

 Having worked in  a research lab for 34 years I see this type of thought process happen all the time. 

 

                - Dave 

How do you mean 'the errors in the flat don't directly add to the errors in the lens you're testing...' I think they do... if they are errors of smoothness, that is.

 

David



#62 DAVIDG

DAVIDG

    Cosmos

  • *****
  • Posts: 8518
  • Joined: 02 Dec 2004
  • Loc: Hockessin, De

Posted 27 February 2018 - 02:43 PM

 If you have a copy of "Optical Shop Testing" by Malacara  he has the math is his  book that explains why the errors in a smooth optical flat do not directly add to the results in double pass. Basically when you have a smooth flat that is many waves from flat what you have is spherical mirror which a huge radius of curvature. That surface does not distort the wavefront by the amount of waves from being non flat. It just moves were the object is located ie were the infinity position that the telescope sees. So it is the same principle as if you test a telescope  with a light source that is a few miles away. At some distance the object is far enough away that the spherical aberration added for  not having the object at infinity is small enough as not to be a problem. If a flat that is many waves from flat but smooth it is the same but in that case the "object" is nearly at infinity so the error is very small. 

   It has been a misunderstanding in the ATM community for years that if you wanted to test a telescope to 1/10 wave you needed at least 1/10 wave flat. Again that is not the case and  a flat that  is non flat to some thing like 50 wave but smooth, adds  around 1/50 wave error to the test. You can't measure that size of error  very easily so  in all practical terms no error from the flat.  So if you find a 1/2 wave flat that is smooth it is more than good enough to test and make excellent optics.

 

            - Dave 


Edited by DAVIDG, 27 February 2018 - 03:22 PM.

  • Richard Whalen, rcwolpert, AllanDystrup and 1 other like this

#63 davidc135

davidc135

    Viking 1

  • -----
  • Posts: 621
  • Joined: 28 May 2014
  • Loc: Wales, UK

Posted 27 February 2018 - 03:27 PM

Certainly, I agree. I should have said 'errors of lack of smoothness'.. or asphericity.

 

David



#64 DAVIDG

DAVIDG

    Cosmos

  • *****
  • Posts: 8518
  • Joined: 02 Dec 2004
  • Loc: Hockessin, De

Posted 27 February 2018 - 04:10 PM

Certainly, I agree. I should have said 'errors of lack of smoothness'.. or asphericity.

 

David

 Of course you don't want  zones and turned edge in the flat because they will add to the error in the test results. Like I said thou a 1/2 wave flat which you can test pretty easily  for these problems and if it smooth  it adds basically no error to the results.

    Flats of this quality show up on Ebay or other places from time to time or can found in places like the  plano side of lens and filters for example or surplus optical equipment.  Scratches and small edge do no harm so you can find some real flat, optical flat for cheap because they are bit beat up. I got a 8" one off of Ebay a few years ago for $35 that had a few minor scratches and the aluminium coating was bad. It is flat to at least 1/10 wave across it's surface and I'm sure new cost at least $1000.  

 

                  - Dave 


  • tim53 and Piggyback like this

#65 Joe Cepleur

Joe Cepleur

    Gemini

  • *****
  • Posts: 3246
  • Joined: 18 Mar 2010
  • Loc: Dark North Woods

Posted 27 February 2018 - 06:26 PM

So, the flat need not be perfectly flat, because its gross error will be barely spherical, meaning both "the same along all radii" and "too small to matter;" but, the surface must be perfectly smooth (as smooth as humanly possible), to prevent the light from scattering out of parallel at the micro scale, thus not returning parallel rays to the telescope?

And, a pan of oil may return less light (because it is less reflective than a mirror), yet it passes both tests, and the lower reflectivity is easily compensated for with a brighter LED!

#66 tim53

tim53

    Voyager 1

  • *****
  • Posts: 13832
  • Joined: 17 Dec 2004
  • Loc: Highland Park, CA

Posted 27 February 2018 - 07:59 PM

 

Certainly, I agree. I should have said 'errors of lack of smoothness'.. or asphericity.

 

David

 Of course you don't want  zones and turned edge in the flat because they will add to the error in the test results. Like I said thou a 1/2 wave flat which you can test pretty easily  for these problems and if it smooth  it adds basically no error to the results.

    Flats of this quality show up on Ebay or other places from time to time or can found in places like the  plano side of lens and filters for example or surplus optical equipment.  Scratches and small edge do no harm so you can find some real flat, optical flat for cheap because they are bit beat up. I got a 8" one off of Ebay a few years ago for $35 that had a few minor scratches and the aluminium coating was bad. It is flat to at least 1/10 wave across it's surface and I'm sure new cost at least $1000.  

 

                  - Dave 

 

Johann gave me an 8" flat a few years ago.  It's got "challenged" coatings, but I don't find that to be a problem for most tests I've done.  I've missed out on a couple reasonably priced 12" flats.  On here.  I've never seen any decent ones for cheap on ebay in the few years I've been looking.

 

I still plan to make a carefully constructed stand for optics out of tubes for testing with the oil pan flat.  But I'm getting in the final stretches getting the house built at Cosmic Acres, so telescopes have taken a back seat to a lot of things.  Then, I need to build an observatory out there.  

 

-Tim.


  • bremms likes this

#67 tim53

tim53

    Voyager 1

  • *****
  • Posts: 13832
  • Joined: 17 Dec 2004
  • Loc: Highland Park, CA

Posted 27 February 2018 - 08:01 PM

Meant to say for testing optics not in OTAs.  I don't plan to make the stand or the optics out of tubes!



#68 AllanDystrup

AllanDystrup

    Apollo

  • *****
  • topic starter
  • Posts: 1354
  • Joined: 27 Sep 2012
  • Loc: Denmark

Posted 28 February 2018 - 03:24 AM

.

     This is an interesting objective: my classic NS 61/900mm achromatic Fraunhofer doublet (2.4” Polarex).  In DPAC it shows quite a well figured lens with straight lines over most of the surface indicating a wave-front error in the order of 1/10λ in green light (max 1/8λ polychromatic for this lens design, as explained by Dave).

    

     It does however show a turned up edge in the outer ~5mm of the objective, which degrades its performance. I intend to mask down the aperture to 55mm and redo the DPAC to see if this will eliminate the effect of the TUE.

    

NS-61-900.png
*click*

    

Allan


Edited by AllanDystrup, 28 February 2018 - 03:37 AM.

  • rolo and Piggyback like this

#69 DAVIDG

DAVIDG

    Cosmos

  • *****
  • Posts: 8518
  • Joined: 02 Dec 2004
  • Loc: Hockessin, De

Posted 28 February 2018 - 09:40 AM

So, the flat need not be perfectly flat, because its gross error will be barely spherical, meaning both "the same along all radii" and "too small to matter;" but, the surface must be perfectly smooth (as smooth as humanly possible), to prevent the light from scattering out of parallel at the micro scale, thus not returning parallel rays to the telescope?

And, a pan of oil may return less light (because it is less reflective than a mirror), yet it passes both tests, and the lower reflectivity is easily compensated for with a brighter LED!

 When I say smooth, I mean lack of zones. You can have a  flat that is measures out to  a total of 1/4 wave  but can be 1/10 wave over most of the surface and have  a turned down edge and that can make it a 1/4 wave. Zones are sharp changes in the figure  like turned edge, holes, hills,  ring zones. Smooth means it has spherical shape be it concave or convex with no sharp changes to the slope  of the figure. 

 

                   - Dave 



#70 bremms

bremms

    Fly Me to the Moon

  • *****
  • Posts: 5135
  • Joined: 31 Aug 2012
  • Loc: SC

Posted 28 February 2018 - 10:28 AM

I was getting strange results with a 6" optical comparator flat. It had a big zone ( more than 1/2 wave) from 50-80% and it was a little rough.  Picked up a 1/20wave 6" flat. The errors in the the other flat were very obvious under monochromatic light and during back to back DPAC tests. Tested most of my smaller optics.. now the super flat sits.  We should have a flat loaner program for non optical fabricators. You don't need them that often.



#71 DMala

DMala

    Apollo

  • *****
  • Posts: 1411
  • Joined: 02 Jun 2015

Posted 28 February 2018 - 10:40 AM

Naive question: when you guys do your DPAC tests with your home-made equipment (actually, the following would apply to commercially-made systems too...), do you first have a method to validate the accuracy of the results for each test? In other words, do you test your test system each time you use it? 



#72 AllanDystrup

AllanDystrup

    Apollo

  • *****
  • topic starter
  • Posts: 1354
  • Joined: 27 Sep 2012
  • Loc: Denmark

Posted 28 February 2018 - 11:11 AM

.
I made a mask for stopping down the NS 2.4" objective by 5mm:

NS-60-900 Mask.jpg


The DPAC result seems a little better, but still not perfect. I think I'll leave it at that for now.

NS 55-900.png


Allan

Edited by AllanDystrup, 28 February 2018 - 01:05 PM.

  • tim53, rolo, Bomber Bob and 1 other like this

#73 DAVIDG

DAVIDG

    Cosmos

  • *****
  • Posts: 8518
  • Joined: 02 Dec 2004
  • Loc: Hockessin, De

Posted 28 February 2018 - 11:16 AM

Naive question: when you guys do your DPAC tests with your home-made equipment (actually, the following would apply to commercially-made systems too...), do you first have a method to validate the accuracy of the results for each test? In other words, do you test your test system each time you use it? 

 The beauty of double pass is that there are very few sources of error. First there is nothing to measure, since it is null test. When it comes to alignment is it easy to tell if your aligned or not so that is not a problem. There is no critical distance to set. The flat just needs to farther away then the focal length of the mirror, lens or complete telescope so three inches in front of the scope or 3 feet it doesn't matter. As I said if the flat is optically  smooth it adds such a small error to the test that is doesn't matter.   It is double pass test so the errors show by 2x so it very sensitive. 

All of this results in a very high confidence to the results and I know of no way to get a false positive ie having the test show you have good optics when they are not.  as you can see from the results posted here it is easy yet very powerful test that shows you in a few minutes the quality of your optics. 

 

                       - Dave 


  • rcwolpert, AllanDystrup and Piggyback like this

#74 bremms

bremms

    Fly Me to the Moon

  • *****
  • Posts: 5135
  • Joined: 31 Aug 2012
  • Loc: SC

Posted 28 February 2018 - 11:38 AM

In a former life, a friend had access to some serious testing equipment at APL. The set up could test 0.90 Meter optics to the 1/100 wave category.  He tested a couple of our scopes as training demo. It was interesting. My friend Matt's C8 tested about 1/3-1/4 PV wavefront. It was 1/3.7 IIRC. My D&G 5" F12 tested 1/7.4 wave front in green. An 8" F5.6 (Coulter maybe) Mirror was about 1/4 wave as well. Matt's C8 was one of the best I've used. That's a good number for an SCT. We couldn't get access and he isn't allowed to take photos. The numbers the interferometer gave are right in line with star tests and my Ronchi star test of that particular lens.



#75 DAVIDG

DAVIDG

    Cosmos

  • *****
  • Posts: 8518
  • Joined: 02 Dec 2004
  • Loc: Hockessin, De

Posted 28 February 2018 - 04:30 PM

 Allan,

   Did you try loosening the three phillips screws with the red paint to see if the lens is being pinched and that is what is causing the turned edge ? On my Unitron 114 I removed the lens from the cell and added three taps of tape at 120° center to rear surface and three taps of tape at 120° center to the back of the retainer ring. Then made sure the lens when reinstalled in the cell had air spacers  over these three taps of tape. So now I had even pressure on the spacers and that improved my test results.

 

              - Dave 


  • deSitter and AllanDystrup like this


CNers have asked about a donation box for Cloudy Nights over the years, so here you go. Donation is not required by any means, so please enjoy your stay.


Recent Topics





Also tagged with one or more of these keywords: beginner, classic, collimation, optics, refractor



Cloudy Nights LLC
Cloudy Nights Sponsor: Astronomics