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

ROSS NULL TEST 330MM F3 MIRROR- HELP

  • Please log in to reply
384 replies to this topic

#1 macleod

macleod

    Viking 1

  • -----
  • topic starter
  • Posts: 907
  • Joined: 21 Sep 2014
  • Loc: Napier, New Zealand, 4110

Posted 05 August 2020 - 12:35 AM

i set up the ross null test today with difficulty as my mirror lens and ronchi screen are on different tables, and alignment was difficult, even with a laser. However I seem to have it working in a fashion, but probably not the correct one unfortunately. I have watched all the videos, read all the help including Ceravalo etc.I have set the distances mirror to lens and lens to screen as accurately as I can. What I see are over 12lines, which I can straighten by moving the lens in all directions until they are straight, absolutely. BUT i cannot get down to only 2 or 3 lines for accuracy - when I move forward I get more lines, and moving back from this position the thing breaks into a ronchi curve or weird diamond shape which is geometrically pretty interesting but obviously wrong.  It appears as if when I withdraw a little I.m close to what looks like roc.As I said I am getting a return image through the ross lens because tilting it alters the lines.However it has occurred to me that I may not be seeing the return beam correctly. I do notice several ghost images / reflections too. I,ll have another go tomorrow, but any suggestions would be appreciated. This should be a simple  foolproof test by the looks of it, but the originator has not met this fool !! Macleod.



#2 starspangled

starspangled

    Explorer 1

  • -----
  • Posts: 88
  • Joined: 12 Jun 2007

Posted 05 August 2020 - 12:41 AM

Hi , If you are seeing curved ronchi lines at two or three lines then you have spherical aberration that probably needs correction .You probably , for the sake of easier alignment , want to use an led shining through a ronchi screen which you look through slightly above .



#3 davidc135

davidc135

    Apollo

  • -----
  • Posts: 1,392
  • Joined: 28 May 2014
  • Loc: Wales, UK

Posted 05 August 2020 - 02:27 AM

i set up the ross null test today with difficulty as my mirror lens and ronchi screen are on different tables, and alignment was difficult, even with a laser. However I seem to have it working in a fashion, but probably not the correct one unfortunately. I have watched all the videos, read all the help including Ceravalo etc.I have set the distances mirror to lens and lens to screen as accurately as I can. What I see are over 12lines, which I can straighten by moving the lens in all directions until they are straight, absolutely. BUT i cannot get down to only 2 or 3 lines for accuracy - when I move forward I get more lines, and moving back from this position the thing breaks into a ronchi curve or weird diamond shape which is geometrically pretty interesting but obviously wrong.  It appears as if when I withdraw a little I.m close to what looks like roc.As I said I am getting a return image through the ross lens because tilting it alters the lines.However it has occurred to me that I may not be seeing the return beam correctly. I do notice several ghost images / reflections too. I,ll have another go tomorrow, but any suggestions would be appreciated. This should be a simple  foolproof test by the looks of it, but the originator has not met this fool !! Macleod.

Having done your researches I'm sure that you are aware of these points but;

 

If you are inside focus approaching the RoC and if the lines look like (()) it sounds like the mirror is over-corrected. Have I got that right? Can a photo be taken at this point? It may not be wrong.

 

I expect that if the lens can be tilt adjusted, whilst preserving the right separations, a symmetrical pattern free of coma and astigmatism can be formed.

 

When you say that with 12 lines showing these can be straightened by moving the lens in all directions did this involve altering the required spacings? 

 

David


Edited by davidc135, 05 August 2020 - 02:35 AM.

  • MKV likes this

#4 macleod

macleod

    Viking 1

  • -----
  • topic starter
  • Posts: 907
  • Joined: 21 Sep 2014
  • Loc: Napier, New Zealand, 4110

Posted 05 August 2020 - 02:41 AM

Starspangled - I,m NOT seeing curved lines, just 12 very straight ones. My problem is how to show only 2 or 3 ??

David135 -  NO I do not alter spacing - just the tilt of lens on its gimble, and rotating slightly horizontally.I have looked at Pinbouts Youtube videos of the Ross Test, and I can reproduce all the faults he shows. BUT when I get the lens lined up, the lines are dead straight. BUT according to Pinbout you should only test with 2 or 3 lines for accuracy.Moving the ronchi screen unit a little brings the image to look like a conventional ronchi test.That,s what I,m curious about.

PS:- Starspangled , I am using an LED ( 10,000 mcd + diffuser ) and a 133lpi ronchi screen.My object to lens distance is 198mm and lens to mirror is 1557mm they DO NOT alter.



#5 starspangled

starspangled

    Explorer 1

  • -----
  • Posts: 88
  • Joined: 12 Jun 2007

Posted 05 August 2020 - 03:27 AM

macleod...I just suggested you had spherical aberration because you said as you moved the ronchi screen it looked like a conventional ronchi , I assumed you meant when you moved to less lines they started to bend .


  • MKV likes this

#6 MKV

MKV

    Cosmos

  • *****
  • Posts: 9,278
  • Joined: 20 Jan 2011
  • Loc: Southeast, USA

Posted 05 August 2020 - 04:58 AM

BUT when I get the lens lined up, the lines are dead straight. BUT according to Pinbout you should only test with 2 or 3 lines for accuracy.Moving the ronchi screen unit a little brings the image to look like a conventional ronchi test.That,s what I,m curious about.

Pinbout is right. What do you mean by "looks like a coinventional" Ronchi test?



#7 macleod

macleod

    Viking 1

  • -----
  • topic starter
  • Posts: 907
  • Joined: 21 Sep 2014
  • Loc: Napier, New Zealand, 4110

Posted 05 August 2020 - 05:55 AM

MKV - perhaps I am moving the screen too far back - because the ross test breaks down into a  ronchi test with the same patterns one sees doing that test. The point where I get the straight ( ?null ) lines is a VERY proscibed position. I have been trying to move the screen to get fewer lines , but unsuccessfully so far. That,s what I am on about. How do I get fewer lines showing ? Macleod



#8 Pinbout

Pinbout

    ISS

  • *****
  • Posts: 25,280
  • Joined: 22 Feb 2010

Posted 05 August 2020 - 06:27 AM

first set it up without the lens. so you can move the screen in / out RoC...  then set the lens infront of the screen buy 150mm and move the lens/screen up buy whatever the program tell you to.

 

that's how swayze did it, that's how I do it, that's how jesus would do it. lol.gif lol.gif lol.gif  [trying to remember the line from Iron Man]

 

steve

https://www.youtube....W8rBwA&index=14

 

https://www.youtube....s2L6aP7&index=3

 

moi

https://www.youtube....W8rBwA&index=13


Edited by Pinbout, 05 August 2020 - 07:33 AM.


#9 gr5org

gr5org

    thegr5store

  • -----
  • Vendors
  • Posts: 228
  • Joined: 06 Sep 2018
  • Loc: Boston

Posted 05 August 2020 - 06:43 AM

A 330mm F/3 can be tested with a Bath Interferometer pretty easily.  No null needed.

 

My understanding of any Ross null is that the position where you place the null (distance from mirror) has to be extremely accurate.  Depending on the mirror maybe an error of 0.1mm (in distance between mirror and null) might add unacceptable error (where you think the mirror is perfect but it's not).  Maybe for your mirror you have much more tolerance (say 10mm).  You have to do ray tracing to determine just how much error in position is allowed as every situation is different.  And if you put the null in the wrong spot it can add a lot of SA.  And if you are polishing you may be making your mirror worse instead of better.  For these 2 reasons (easy to make your mirror worse, need to do ray tracing which I've never done) it just seems so much easier to use a Bath or Foucault.  The Foucault of course would need a micrometer and you'd have to test many zones.



#10 Pinbout

Pinbout

    ISS

  • *****
  • Posts: 25,280
  • Joined: 22 Feb 2010

Posted 05 August 2020 - 06:55 AM

 

My understanding of any Ross null,,,

guress you never used it to help make a mirror.

 

the Ross null is the easiest test to get "an overall look at the mirror"... at a glance,  which is a huge help while figuring.

 

shoot your bath thru it and get straight fringes...



#11 Pinbout

Pinbout

    ISS

  • *****
  • Posts: 25,280
  • Joined: 22 Feb 2010

Posted 05 August 2020 - 07:23 AM

1st set up the grating at RoC to get about 4lines outside of RoC

 

then place the lens 420mm infront of the grating.

 

slide the grating up to about 200mm behind the lens - this distance changes cause you want to slide it back and forth to get inside/outside views of the grating lines.

 

adjust height, side to side, tip and tilt of the lens to get things symmetrical.

 

then measure the mirror to lens distance to 1560mm

 

[your lens may change these measurements grin.gif ]

 

13f3 100 400 lens.JPG


Edited by Pinbout, 05 August 2020 - 08:03 AM.


#12 BGRE

BGRE

    Gemini

  • *****
  • Posts: 3,132
  • Joined: 21 Mar 2016
  • Loc: New Zealand

Posted 05 August 2020 - 07:24 AM

Using a cheap lens with a potential irregular error of up to 1/4 wave on each surface is an exercise in futility since the lens in a Ross null is double passed the resultant wavefront error can be as large as 1/2 wave.

Hoping that the null lens is better than that without an independent verification of its errors is an exercise in self deception.

Even if the null lens has adequate quality, only setting the distance from the null lens to the test surface without checking that the axial position of the test head at the focus is also correct doesnt constrain the test surface shape to the desired shape when a null is achieved.

The videos lack a clear explanation of how to setup a Ross null.

#13 Pinbout

Pinbout

    ISS

  • *****
  • Posts: 25,280
  • Joined: 22 Feb 2010

Posted 05 August 2020 - 07:32 AM

 

Using a cheap lens with a potential irregular error of up to 1/4 wave on each surface is an exercise in futility since the lens in a Ross null is double passed the resultant wavefront error can be as large as 1/2 wave.

 

1st Ceravola uses a 1/10 lens - gets 1/60 wave results when tested with his IF to qualify the results.

 

2nd with a 100mm lens he's only using 70.8mm of it. 

 

 

nly setting the distance from the null lens to the test surface without...

wrong... even ceravola moves his ke in the test to see if he's under or over... the "head" is never stationary, let alone measuring to a specified distance.



#14 MKV

MKV

    Cosmos

  • *****
  • Posts: 9,278
  • Joined: 20 Jan 2011
  • Loc: Southeast, USA

Posted 05 August 2020 - 07:45 AM

Using a cheap lens with a potential irregular error of up to 1/4 wave on each surface is an exercise in futility since the lens in a Ross null is double passed the resultant wavefront error can be as large as 1/2 wave.

A refractive surface of index n ≈ 1.5 rated at 1/4 wave will produce a wavefront error of ≈ 1/8 wave pv.

 

The source to lens distance should be fixed on a sled and the Ross lens to test mirror should be set as called for, and aligned. Only the testing device (k-e or Ronchi screen) should move independely.

 

 

1st Ceravola uses a 1/10 lens - gets 1/60 wave results when tested with his IF to qualify the results.

I don't think he tested the mirror with his IF in a Ross null configuration. That wold make no sense. Ceravolo's IF will test a mirror without any additional optics. It already has a high-precision reference element in its architecture. 



#15 Pinbout

Pinbout

    ISS

  • *****
  • Posts: 25,280
  • Joined: 22 Feb 2010

Posted 05 August 2020 - 08:52 AM

 

I don't think he tested the mirror with his IF in a Ross null configuration. That wold make no sense. Ceravolo's IF will test a mirror without any additional optics. It already has a high-precision reference element in its architecture.

it was a separate test with his if config. to qualify with numbers since the ross doesn't give numbers... and when you pay a high price people like numbers... it makes them feel warm and fuzzy so they can sleep well at night. lol.gif


Edited by Pinbout, 05 August 2020 - 08:52 AM.


#16 Pinbout

Pinbout

    ISS

  • *****
  • Posts: 25,280
  • Joined: 22 Feb 2010

Posted 05 August 2020 - 08:57 AM

 

A refractive surface of index n ≈ 1.5 rated at 1/4 wave will produce a wavefront error of ≈ 1/8 wave pv.

and  you use less of the lens so the portion being used will produce even better wavefront since typically the outer portions of the lens deviate more...


Edited by Pinbout, 05 August 2020 - 08:57 AM.


#17 DAVIDG

DAVIDG

    Cosmos

  • *****
  • Posts: 9,920
  • Joined: 02 Dec 2004
  • Loc: Hockessin, De

Posted 05 August 2020 - 09:48 AM

 Here is how  I set up the Ross Null.  I set up the mirror and tester just like doing standard Ronchi test BUT I place the tester   farther back then the lens to mirror distance by a couple of inches. Since the tester is not at the radius of curvature of the mirror, you won't see the mirror filled  light like but you will see the light from the tester in the mirror. You move the tester and/or the mirror so the light is in the centered in the  mirror. Now I measure the distance from the center of the mirror to were the Ross Null lens is suppose to be placed plus any offset for how the lens is mounted. My lens stand has a flat front surface. At that position I place a heavy book. so the front of test stand  can rest against  the edge of the book and hold the distance from the lens to the mirror.  With  the lens NOT in place I look thru the Ronchi screen and you'll see just the center of the mirror illuminated again this is because your not at the radius of curvature of the mirror Now I slide the lens in place. Looking at the outline of lens and the the outline of the mirror in the distance  I center the out line of the lens around the out line of the mirror  At this point  when look you thru the Ronchi grating  and through the lens at the mirror you should  see the return image and the Ronchi pattern. You may have to adjust the tester slightly up or down to it centered and move the tester  closer or farther away from the lens to get the Ronchi pattern down to a few lines. The pattern may not be symmetrical top to bottom or right to left. You adjust the tip and tilt of the lens to make the pattern symmetrical so you may need to  twist the lens and/or tip forward or back  but you do not change the spacing between the lens and mirror. 

  At this point you should now see a typical symmetrical  Ronchi pattern of a couple of Ronchi bands. The mirror should look perfectly centered in the lens if you have everything  aligned correctly.  You now do the test just like  a standard Ronchi test moving the Ronchi screen closer or farther away and going from inside focus to focus and outside of focus. 

 

 

                     - Dave 


Edited by DAVIDG, 05 August 2020 - 08:30 PM.

  • LarsMalmgren, GaryF and gr5org like this

#18 gr5org

gr5org

    thegr5store

  • -----
  • Vendors
  • Posts: 228
  • Joined: 06 Sep 2018
  • Loc: Boston

Posted 05 August 2020 - 09:59 AM

@pinbout - Yes I get that the Ross null would make bath lines straight.  Which isn't necessary but would be nice.  Anwyay...

 

That software says the "object to lens distance" should be 199.545mm.

 

What if it's 199.7mm?  How much spherical aberration will that add?  What if it's 202mm?  Is that enough to add serious SA?  Probably?  I'm told you must do ray tracing analysis for this particular setup (330mm F/3 mirror with 403mm FL ross) to find out how precise this position needs to be.

 

I get the impression that for a 13 inch F/6 mirror the positioning of the null lens can be off by several mm (maybe even 10mm?) and still have excellent SA.  But for a 13 inch F/3 (that's a much more serious/difficult mirror) the positioning will have to be MUCH more accurate.  I'm not sure how much more accurate.  I think it's important to do ray tracing to find out.

 

@BGRE also makes the point that the Ross null needs to be quite good quality. 

 

If you make dozens of mirrors and have done star tests on them then this can validate that your process (and null) has been working.  For those mirrors.  But if you have been having excellent results for a 13 inch F/5 and then try to make an F/3 you might not get such good results.



#19 Pinbout

Pinbout

    ISS

  • *****
  • Posts: 25,280
  • Joined: 22 Feb 2010

Posted 05 August 2020 - 10:30 AM

 

 

That software says the "object to lens distance" should be 199.545mm.

 

What if it's 199.7mm?

object to lens distance is a ref.  you always move the ronchi or ke to see how the correction is going when its close.

 

its the mirror to lens distance that's important.

 

biggest thing is getting your lines straight. that's most of the battle. everything else is fine tuning correction.

 

 

But if you have been having excellent results for a 13 inch F/5 and then try to make an F/3 you might not get such good results.

start a different thread.


Edited by Pinbout, 05 August 2020 - 10:31 AM.


#20 DAVIDG

DAVIDG

    Cosmos

  • *****
  • Posts: 9,920
  • Joined: 02 Dec 2004
  • Loc: Hockessin, De

Posted 05 August 2020 - 11:07 AM

 Pinbout is correct that  the "object to lens" distance is reference in that  if the mirror was perfectly figured that would be the exact focal point of the system. It is the same if you were figuring a spherical mirror. If the mirror was a perfect sphere there be only one exact focal point but if it has zones the zones would have different focal points. 

   What is critical to understand is you set the lens to mirrors spacing and that stays fixed during the test and  you just move the light/source/ronchi grating just like your doing a standard ronchi test. You figure the mirror so it shows dead straight Ronchi bands with 4 or less showing on both in the inside and outside of focus. Again just like testing and figure a spherical mirror. 

   When it comes to setting up the test and aligning the lens,  you first need to align the tester to the mirror so the tester in on the optical axis of the mirror. It is same as doing a standard Ronchi test. The only difference is that the tester will not be at the radius of curvature of the mirror so the mirror will not be fully illuminated. Once you get the tester aligned you can then align the lens. If you try to do both at once you'll have problems figuring out which one or both are out of alignment. 

     Everyone gets hung up that you need to have  super accurate lens or the test is worthless. What you need to undesrtand is  results gives you  information . What you need to do is double and  triple check the results against other test methods and they ALL need to agree. If they don't agree you need to figure out why. You just don't take the best one. If not this is why over 35+ years of making and testing optics I have seen many many optics that were stated to be "1/10" or better yet when correctly tested turn out to worse then 1/4 wave.  You need to make sure you found and corrected for the errors in the test results no matter what test method you use.

 

                                 - Dave 



#21 gr5org

gr5org

    thegr5store

  • -----
  • Vendors
  • Posts: 228
  • Joined: 06 Sep 2018
  • Loc: Boston

Posted 05 August 2020 - 11:19 AM

I'm sorry to repeat myself so much.  But this distance, object to lens (mirror to lens), is what I was talking about.  setting it to 199.7mm isn't good enough if you don't know the tolerance.  And the tolerance for an F/3 is probably very tight.  You can polish your mirror until you get perfect ronchigram but if you have the object to mirror distance at the wrong distance (again the accuracy for some mirrors might need to only be within 10mm but others 0.1mm) you are going to have unacceptable spherical aberration on your final mirror which you will discover only with a star test or Bath test (you won't likely notice it with foucault test on an F/3 mirror unless you are doing a LOT of zones and careful measurements).



#22 DAVIDG

DAVIDG

    Cosmos

  • *****
  • Posts: 9,920
  • Joined: 02 Dec 2004
  • Loc: Hockessin, De

Posted 05 August 2020 - 11:33 AM

 The 199.7mm is  object to lens distance  which is the light/source Ronchi  grating distance to the lens.  It is that same as when you test  a spherical mirror with a radius of curvature of 2000mm. The object to lens distance is the same as the ROC is standard Ronchi test. It tells were the exact focus will be when the mirror is a perfect sphere but  when doing the Ronchi test you don't set the grating at the exact focal plane. You move inside and outside of focus to check the correction. It is the same in the Ross. The lens to mirror spacing is what is critical since that set the amount of spherical aberration to cancel that of the mirror and give you a null setup. When you run the Ross Null program you can tell it what tolerance you wish for the mirror ( 1/4 , 1/8 or 1/10 wave) and it will tell you what the plus/minus values needs to be in the distance of the lens to mirror spacing.  In the data posted in message #11 from the Ross Null program you will see a box in the upper right  " Include Lens to mirror error" when checked it will give you the error tolerance in the measurements.

 

                    - Dave 


Edited by DAVIDG, 05 August 2020 - 11:35 AM.


#23 Pinbout

Pinbout

    ISS

  • *****
  • Posts: 25,280
  • Joined: 22 Feb 2010

Posted 05 August 2020 - 02:25 PM

its the lens to mirror    that changes the reading in correction

 

not the ke/ronchi to lens.



#24 macleod

macleod

    Viking 1

  • -----
  • topic starter
  • Posts: 907
  • Joined: 21 Sep 2014
  • Loc: Napier, New Zealand, 4110

Posted 05 August 2020 - 05:20 PM

A big thank you to ALL the posters, I appreciate your help and increasing my understanding of these tests.

DAVIDG - Thank you especially, your posts have cleared up all my issues !! Your work teaching at Delmarva ( ?) has obviously told you what amateurs find puzzling in these tests. Your post in  #17 says exactly what I have been seeing , and actually been trying to do.  Danny, Pinbout has also been most generous in lending me a lens, and helped me along the way, and I,m most grateful for all his help. However David your clear concise information has cleared up my questions , particularly the relationship between a pure Ronchi test on its own and the insertion of the lens for Ross Null - as well as why I was seeing "Ronchi Patterns " at times ! It had not twigged with me fully that this was basically a modified ronchi test, I had it in mind that it was a totally seperate  test ! I think its this query of mine in earlier  post that made MKV ask in #6  "what do you mean "it looks like a ronchi test". Sorry to cause confusion , but I thought it was a totally seperate entity , not a modified ronchi. 

So now on to practical set up - I am using  my Glatter laser in eyepiece holder to align optics on axis, so it passes thru center of lens to center of mirror,but my  biggest problem is getting returned beam where I want it ! So to quote the Salvation Army - onward christian soldiers, fight the good fight , and lets beat the ross null ! Macleod.



#25 macleod

macleod

    Viking 1

  • -----
  • topic starter
  • Posts: 907
  • Joined: 21 Sep 2014
  • Loc: Napier, New Zealand, 4110

Posted 05 August 2020 - 05:32 PM

In case anyone is interested, the figures are :- Mirror 330mm , Focal length1976mm to nearest whole number. Lens :- 100mm dia , Radius 206.7mm , thickness 10mm , R index 1.55 . This lot seems to give me Object to lens 197.8 ( 198 to me ) x lens to mirror 1557mm to nearest whole number. I can,t measure realistically to decimals of a mm ! I have carefully cut and sanded a length of dowel to mirror/ lens distance as close as I can.Macleod.




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






Cloudy Nights LLC
Cloudy Nights Sponsor: Astronomics