Question about distance for a star test with an artificial star.

[Caput Gorgonis]

Hello everyone! I hope someone can help me out with this one.

I want to star test my 16" f4,5 mirror (1800mm focal lenght). As I do not want to waste a nice, steady (and rare!) night testing optics, I have purchased an artificial star from Hubble optics (info here: http://www.hubbleoptics.com/artificial-stars.html) which basically is a LED lamp with 5 precision pinholes: 50, 100, 150, 200 and 250 microns).

I am following the instructions of "the book" (i.e. Suiter) and I have a question with regards to distances. First I will present my numbers, in case I am getting my math wrong:

1. I get my telescope to pinhole distance for my mirror (I use equation 5.5. to get a multiplier of 59). 59 times * 1,8 m (my focal length) = 106,2m.
2. As I want to test a 1/8th mirror, I will double that distance (as per page 92): 106,2 * 2 = 212,4 m which I will round up up to 220 metres.
3. My artificial source must have (as per table 5.3b) a diameter between 0,29 and 0,45 mm. I will take 0,30mm.

So as per my calculations, I will need to put my 0,30mm artifical star at least 220 meters away. I could select the 250 microns (0,25mm) pinhole from the artifical star and I think I would be OK to go. But my question is:

Can I select a smaller pinhole and get the telescope closer? My intuition tells me so, but I would like to have some feedback from someone with experience in this field. If I can reduce the distance: what is the relationship? If I take a pinhole 5 times smaller, can I get the telescope 5 times closer? Or do artifical stars follow the inverse-square law?

Thanks a lot for your help. Further advice to carry the star test will be appreciatted as well. 

Caput

[Sporadic Dobstronomer]

10 hours ago, Caput Gorgonis said:

Can I select a smaller pinhole and get the telescope closer?

 

No you can't!

The distance needs to be enough that all the possible paths from the pinhole to the image via the mirror are the same length within your wavefront tolerance so making the pinhole smaller does not help.

The importance of the pinhole size is that it needs to be smaller than the resolving power of your telescope at that distance.  At 220m I calculate 0.37mm is small enough.

BTW, I can't make sense of Suiter's figures.  My calculation is that you need a distance of 470m to get 1/4 wavefront accuracy -so just in case I am right you need to err on the long side...

 

[Peter Drew]

It's hard enough for an experienced optical worker to determine whether a large mirror is 1/8 wave or not. Do you have concerns that your mirror falls short of this quality?

If out of focus stars look round, inside and outside extrafocal rings look similar and regular then factors outside your control will govern the performance most of the time.  Ideally, star testing for optical quality should be conducted at as near infinity as possible to ensure parallel light rays, professionals do this on an optical bench that can simulate an infinite light source. You will have noticed that several of the images in Suiter, despite the demonstrated errors, still seem to provide a decent focused star. I certainly wouldn't trust my assessment based on that.  

[Caput Gorgonis]

Sporadic, Peter, first of all thank you for your answers. I hope you do not mind some new questions as start test is, at least for me, still a dark art.

11 hours ago, Sporadic Dobstronomer said:

No you can't!

The distance needs to be enough that all the possible paths from the pinhole to the image via the mirror are the same length within your wavefront tolerance so making the pinhole smaller does not help.

The importance of the pinhole size is that it needs to be smaller than the resolving power of your telescope at that distance.  At 220m I calculate 0.37mm is small enough.

BTW, I can't make sense of Suiter's figures.  My calculation is that you need a distance of 470m to get 1/4 wavefront accuracy -so just in case I am right you need to err on the long side...

 

Ok, so what I gather is that once the pinhole is small enough, (i.e. smaller than the resolution of your telescope) there is no advantage in going smaller. And the distance is set by the wavefront tolerance, so you cannot get closer. Thank you for the insight. 

When you say that you cannot make sense of the Suiter's figures... you mean you do not agree with Suiter's formulas or that my calculations are wrong? (I have checked again and I got the same results... )

9 hours ago, Peter Drew said:

It's hard enough for an experienced optical worker to determine whether a large mirror is 1/8 wave or not. Do you have concerns that your mirror falls short of this quality?

If out of focus stars look round, inside and outside extrafocal rings look similar and regular then factors outside your control will govern the performance most of the time.  Ideally, star testing for optical quality should be conducted at as near infinity as possible to ensure parallel light rays, professionals do this on an optical bench that can simulate an infinite light source. You will have noticed that several of the images in Suiter, despite the demonstrated errors, still seem to provide a decent focused star. I certainly wouldn't trust my assessment based on that.  

Well, my intent is to check the mirror and in the process learn a bit about optics and how to star test telescopes. My mirror is a Nichol Optics which is Null tested but comes with no Zygo report so I would like a least to have a more educated opinion about the mirror than the views I get when observing (they are good but until now always seeing limited)

I agree with your second paragraph. Some focused (and defocused) star images look basically the same. Suiter says time and time again that it is not that difficult, but I am sure it is a challenge. And of course I will not make an assesment based on my star test, but at least I will get some hints.

Now I will have to prepare the test. At least we have Google Earth and GPS to plan ahead where and how to get this done! 

Thanks again!

Caput

[Sporadic Dobstronomer]

1 hour ago, Caput Gorgonis said:

 

...you mean you do not agree with Suiter's formulas or that my calculations are wrong? (I have checked again and I got the same results... )

 

I do mean that I don't agree with Suiter's formulas.  I'm reluctant to declare emphatically "I''m right, he's wrong" given how much he knows, but I can't find an error in my calculation and he does use the terms "empirical" and "1991".

My calculation is done by comparing the paths at the centre and edge and subtracting one from the other.  A simple calculation that is possible because the spreadsheet I use has about 15 figures accuracy in calculations.  It would not have been a good way to do it with older computers or calculators that had 8 figure accuracy; 470m is 855,000,000 waves or about 7,000,000,000 1/8 waves!

[cs1cjc]

I think you are doing  useful thing in attempting to learn more about optics by star testing.  I have found it useful, in Newtonians particularly for:

* for identifying and trying to tackle tube currents when all I thought I had initially was poor seeing...

* looking at how bad coma really is off-axis on a F/4.5 mirror (whether an observer can spot it or not) and seeing the improvement offered by a coma corrector.

I have never seriously tried to estimate the error of an optic using a star test.  It looks quite complicated and subjective, but I would not discourage you.  It would surely be worthwhile to understand better how your instrument performs against others and how that correlates with star test results.  I think that recording the test as a drawing or image would be very helpful for building up experience over time.  Wide experience is I am sure important and I do not have it.

There is an issue with contrast on the Hubble optics artificial star.  I put a long hood of black foam secured by elastic bands around mine, which helps, as does making observations under a cloudy sky or at dusk.

The distances in Suiter are certainly minima.  As I understand it, they only hold for simple optics when primary spherical aberration (4th order terms only, in equation 10.1 on page 186 in the 2nd edition of Star Testing) dominates design imperfections.  Simple optics appear to include long focal length achromats and newtonians but no modern optics such as fast newtonians, apochromats and complex designs such as optimised/corrected/modified Dall-Kerkham. I would believe specific calculations rather than Suiter's approximation, even if done on a spreadsheet (a technique I have used myself for estimating coma).

 

[cs1cjc]

In passing Tommy Nawratil of teleskop-austria.com records his (highly defocused) Vega star tests here: Vega_star_test_images