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

Spider and Secondary Diffraction: what to do, what to avoid

  • Please log in to reply
370 replies to this topic

#351 RAC

RAC

    Viking 1

  • -----
  • Posts: 753
  • Joined: 30 Sep 2011
  • Loc: New Zealand

Posted 26 September 2015 - 03:50 PM

I made a vulcano mask and tested it two nights ago with my wire spider 20" f3.8. Seeing was very good and I was looking at Antares with it's little blue companion. The mask did add a little more diffraction around Antares but it was hard to pick. I like the view a little better without it but it looks like an ok solution for no spikes.

looked horrible either side of focus though haha.
  • SouthernOracle likes this

#352 mark cowan

mark cowan

    Vendor (Veritas Optics)

  • *****
  • Vendors
  • Posts: 9406
  • Joined: 03 Jun 2005
  • Loc: salem, OR

Posted 28 September 2015 - 06:02 PM

Been thinking more about how to get a clean micron sized source for the test and something Ed Jones mentioned here: http://www.cloudynig...i/#entry6802963

 

A 9micron source may do the trick and it's not as complex as the alternative.  Ordered a 10m piece of fused silica optical fiber.



#353 kfrederick

kfrederick

    Soyuz

  • *****
  • Posts: 3996
  • Joined: 01 Feb 2008

Posted 04 October 2015 - 03:39 PM

http://astro-1.org/astro1-mission    looks like these guys do not wont the obstructions     Funny there pic has spikes



#354 mark cowan

mark cowan

    Vendor (Veritas Optics)

  • *****
  • Vendors
  • Posts: 9406
  • Joined: 03 Jun 2005
  • Loc: salem, OR

Posted 04 October 2015 - 04:10 PM

And for a good reason:

 

Such clean stellar images make it possible to block the light from a bright parent star using a coronagraphic instrument to search for orbiting planets and other surrounding debris. This will also help to observe subtle distortions in distant galaxy shapes caused by the gravity of intervening dark matter, which can help to calibrate data from other telescopes surveying significant fractions of the sky.

  • kfrederick likes this

#355 kfrederick

kfrederick

    Soyuz

  • *****
  • Posts: 3996
  • Joined: 01 Feb 2008

Posted 07 October 2015 - 07:47 AM

 

And for a good reason:

 

 

 

Such clean stellar images make it possible to block the light from a bright parent star using a coronagraphic instrument to search for orbiting planets and other surrounding debris. This will also help to observe subtle distortions in distant galaxy shapes caused by the gravity of intervening dark matter, which can help to calibrate data from other telescopes surveying significant fractions of the sky.

  One 8.4 meter side mirror for the GMT is finished and is offaxis  .By the time they get all  done they know very well how to make them and maybe we will see some 8.4 meter offaxis .



#356 mark cowan

mark cowan

    Vendor (Veritas Optics)

  • *****
  • Vendors
  • Posts: 9406
  • Joined: 03 Jun 2005
  • Loc: salem, OR

Posted 07 October 2015 - 11:26 AM

1.8 meter is puny compared to those. They already know how to make them. You should consider making your 75" offaxis. The computer controlled stressed polishing lap isn't exactly a piece of cake but it's doable, if you have the interferometer to measure.

http://www.gmto.org/.../GMT_5494-8.pdf

http://www.loft.opti...y_Published.pdf
  • ctcables likes this

#357 Pinbout

Pinbout

    ISS

  • *****
  • Posts: 22499
  • Joined: 22 Feb 2010
  • Loc: Montclair, NJ

Posted 07 October 2015 - 11:32 AM

 

You should consider making your 75" ...

 

.006" difference from 0 conic to -1 conic. :crazy:


Edited by Pinbout, 07 October 2015 - 11:33 AM.


#358 kfrederick

kfrederick

    Soyuz

  • *****
  • Posts: 3996
  • Joined: 01 Feb 2008

Posted 07 October 2015 - 11:56 AM

1.8 meter is puny compared to those. They already know how to make them. You should consider making your 75" offaxis. The computer controlled stressed polishing lap isn't exactly a piece of cake but it's doable, if you have the interferometer to measure.

http://www.gmto.org/.../GMT_5494-8.pdf

http://www.loft.opti...y_Published.pdf

 Be fun to dream . And nice to know people can do that .



#359 mark cowan

mark cowan

    Vendor (Veritas Optics)

  • *****
  • Vendors
  • Posts: 9406
  • Joined: 03 Jun 2005
  • Loc: salem, OR

Posted 07 October 2015 - 08:53 PM

You could do it.  I'm expecting to have to build a stressed lap polisher for certain sizes, speeds, and conics, where it really fights back.  The principle isn't that hard, but I'm sure the implementation is a ****.



#360 Nathan Carlie

Nathan Carlie

    Lift Off

  • -----
  • Posts: 12
  • Joined: 20 Dec 2012

Posted 23 December 2015 - 12:09 PM

HI All, 

   I know this post has dried up recently, but I had a question. Hope it's not too much of a necro. 

 

I really like the idea of using this, what I think of as a spider-mask, to moth reduce diffraction spikes and hide the focuser draw tube so that you can reduce the secondary size. 

 

I'm looking to build a faster (f/4) 12 inch newtonian as an experimental platfor, and I would like to keep the secondary to ~20%. This would tend to put the draw tube in the light path if one wanted a decent sized fully-illuminated field (which I do). 

 

The "volcano mask" block a aweful lot of light, and then there were the concerns over the asymmetrical appearance of stars, so why not change it up? 

 

Pardon my poor drawing, I only have powerpoint... 

 

Here, I'm using 2 "volcanos" at 90 degrees apart, and changed them to cover 45 degrees per arc segment, for a total of 45x4=180 degrees. I can still move the draw tube a good way in before it can be seen by the primary and it blocks a lot less area.  Since they are at 90 degrees, the starts should appear symmetrical once again. 

 

I tried ot run this on maskulator, but I can't get it to work... it just does nothing. Can one of you guys run it and see what you get? 

 

Thanks!

 

Any suggestions?

Attached Thumbnails

  • Modified spider mask.png


#361 wh48gs

wh48gs

    Surveyor 1

  • *****
  • Posts: 1840
  • Joined: 02 Mar 2007

Posted 24 December 2015 - 08:36 AM

Here's what OSLO shows. I didn't try to replicate the exact shape, but this is close enough to illustrate the effect.

 

stl.PNG

 

Vla


  • Nils Olof Carlin, Dave O and Earthbound1 like this

#362 Nathan Carlie

Nathan Carlie

    Lift Off

  • -----
  • Posts: 12
  • Joined: 20 Dec 2012

Posted 24 December 2015 - 12:33 PM

But what about a second one at 90 degrees? 

 

Also, isn't the boost to MTF at high spacial frequencies interesting?  It looks like the opposite of apodization. Not sure if that is a good thing or a bad thing. 


Edited by Nathan Carlie, 24 December 2015 - 12:40 PM.


#363 Nils Olof Carlin

Nils Olof Carlin

    Vanguard

  • *****
  • In Memoriam
  • Posts: 2227
  • Joined: 26 Jul 2004

Posted 25 December 2015 - 06:17 AM

Vla,

post #361 is the one I have waited for all this time :grin:

It has been established over and again that you can suppress far-field spikes (and related far-field diffraction phenomena) with odd-shaped secondary supports, but the near-field effects have not been well modeled. In this post, you see that such supports have (as I suspected) considerable, and usually  asymmetric, effects in the near field. Much as a much enlarged secondary obstruction would have - so, while you avoid spikes, detail contrast (such as planetary) will indeed suffer.

Not only that - also, it is clear from the point-spread functions that what appears as improved resolution to the right in the MTF graphs is not for real. The MTF graph represents how a sinusoidal line pattern, as seen by the optics in one direction, across the lines, would appear when the spatial frequency is varied. The enhanced inner ring (as seen in the horizontal direction in the 2 first images, in all directions in the 3rd) from neighboring lines will constructively interfere with the peak from the middle line, causing a "resonance" at high frequencies. This does not mean improved resolution in a random 2-dimensional pattern of details.

Don't try to use the MTF curve to guess what the optics will do to the image - use the PSF instead (point spread function). At least with telescope optics... 

 

Nils Olof

<flame-resistant hard hat on>



#364 wh48gs

wh48gs

    Surveyor 1

  • *****
  • Posts: 1840
  • Joined: 02 Mar 2007

Posted 25 December 2015 - 10:17 AM

But what about a second one at 90 degrees? 

 

Also, isn't the boost to MTF at high spacial frequencies interesting?  It looks like the opposite of apodization. Not sure if that is a good thing or a bad thing. 

 

Here's the double one, and the one with a wider base. One thing I didn't pay attention to the first time is the different cutoff OSLO gives for this form of obstruction. I went with the graphs normalized to the cutoff frequency, but OSLO gives lower cutoff for this form of obstruction: for the narrower stalk (single or double) it gives 208 lines per mm, and for the wider one 195lpm, with the central obstruction, while for the clear or central obstruction alone gives 221lpm cutoff (it's 6-inch f/8.15).
Considering that the size of central maxima doesn't change, that doesn't add up (similarly, a 33% obstruction PSF is practcally identical to the PSF of a 10% larger, linearly, aperture, with 1/4 wave of spherical aberration, yet the standard MTF formalism states that the latter has 10% higher cutoff). I just can't help but conclude that the standard MTF formalism using pupil autocorrelation for deriving MTF is not appropriate for obstructed apertures. In other words, I believe that the actual contrast transfer is closer to that presented in the first post. But this is what OSLO shows.

stl3.PNG

Vla


Edited by wh48gs, 25 December 2015 - 10:19 AM.


#365 wh48gs

wh48gs

    Surveyor 1

  • *****
  • Posts: 1840
  • Joined: 02 Mar 2007

Posted 25 December 2015 - 10:23 AM

Nils,

That's pretty much my thoughts when I looked into this thread some time ago. Any additional obstructed area moves more energy out of the central maxima. It may be partly compensated by its reduced size, but the overall effect is always negative.

Vla



#366 mark cowan

mark cowan

    Vendor (Veritas Optics)

  • *****
  • Vendors
  • Posts: 9406
  • Joined: 03 Jun 2005
  • Loc: salem, OR

Posted 25 December 2015 - 01:41 PM

Thanks Vla, this handily demolishes the idea for me - the tradeoff is both heavily nonsymmetrical and detrimental to the simple obstructed Airy disc.

 

The optimal secondary obstructions for large fast are not terrible thanks to geometry, so that's where I'll stay. :waytogo:



#367 Oberon

Oberon

    Mercury-Atlas

  • -----
  • topic starter
  • Posts: 2859
  • Joined: 24 Feb 2013
  • Loc: Hunter Valley NSW Australia

Posted 25 December 2015 - 04:34 PM

Vla,

post #361 is the one I have waited for all this time :grin:

It has been established over and again that you can suppress far-field spikes (and related far-field diffraction phenomena) with odd-shaped secondary supports, but the near-field effects have not been well modeled. In this post, you see that such supports have (as I suspected) considerable, and usually  asymmetric, effects in the near field. Much as a much enlarged secondary obstruction would have - so, while you avoid spikes, detail contrast (such as planetary) will indeed suffer.

Not only that - also, it is clear from the point-spread functions that what appears as improved resolution to the right in the MTF graphs is not for real. The MTF graph represents how a sinusoidal line pattern, as seen by the optics in one direction, across the lines, would appear when the spatial frequency is varied. The enhanced inner ring (as seen in the horizontal direction in the 2 first images, in all directions in the 3rd) from neighboring lines will constructively interfere with the peak from the middle line, causing a "resonance" at high frequencies. This does not mean improved resolution in a random 2-dimensional pattern of details.

Don't try to use the MTF curve to guess what the optics will do to the image - use the PSF instead (point spread function). At least with telescope optics... 

 

Nils Olof

<flame-resistant hard hat on>

 

Nils, this post

http://www.cloudynig...void/?p=6786135

2nd row on left

eliminates the asymmetric effects in the near field



#368 Mike I. Jones

Mike I. Jones

    Skylab

  • *****
  • Posts: 4374
  • Joined: 02 Jul 2006
  • Loc: Fort Worth TX

Posted 25 December 2015 - 05:12 PM

Just a comment: The use of the terms "near field" and "far field" in this thread in relation to MTF aren't correct. The MTF curve has two axes, the horizontal axis always being spatial or angular frequency, and the vertical being amplitude or phase. "Near field" and "far field" qualitatively describe the electromagnetic field distribution as a function of distance from a scattering aperture or object. They are not applicable to MTF curves. I think "near field" in this thread is being mistakenly used to mean low spatial frequency, and "far field" is instead higher spatial frequencies.
Mike



#369 Nils Olof Carlin

Nils Olof Carlin

    Vanguard

  • *****
  • In Memoriam
  • Posts: 2227
  • Joined: 26 Jul 2004

Posted 26 December 2015 - 03:53 AM

Just a comment: The use of the terms "near field" and "far field" in this thread in relation to MTF aren't correct. The MTF curve has two axes, the horizontal axis always being spatial or angular frequency, and the vertical being amplitude or phase. "Near field" and "far field" qualitatively describe the electromagnetic field distribution as a function of distance from a scattering aperture or object. They are not applicable to MTF curves. I think "near field" in this thread is being mistakenly used to mean low spatial frequency, and "far field" is instead higher spatial frequencies.
Mike

Mike,

I did not relate near/far field to MTF at all - the less references to MTF, the better in my mind. What I had in mind was "near field" as very approximately the first several diffraction rings around the Airy disk, "far field" what is outside this range (such as diffraction spikes from straight vanes, and the spread-out diffraction from curved ones). Thus, at least for me, "near/far field" is entirely in the spatial domain, not frequency domain.

After all, the image as recorded and perceived is in the spatial domain (think Point Spread Function) - moving to frequency domain (MTF) involves Fourier transformation. This is (to me at least, and I would not think I am alone) a very non-intuitive procedure, even though some general rules can be simple and clear enough. I think the parallel thread in the Reflectors forum 

http://www.cloudynig...-3#entry6962538

illustrates the difficulties in mentally trying to go from MTF to image. I may be in a small minority, but I think the PSF is much easier to use in predicting the effects of various optical constraints as it is, rather than transformed to MTF.

 

Nils Olof


Edited by Nils Olof Carlin, 26 December 2015 - 04:03 AM.


#370 Nils Olof Carlin

Nils Olof Carlin

    Vanguard

  • *****
  • In Memoriam
  • Posts: 2227
  • Joined: 26 Jul 2004

Posted 26 December 2015 - 04:25 AM

 

 

 

Nils, this post

http://www.cloudynig...void/?p=6786135

2nd row on left

eliminates the asymmetric effects in the near field

 

 

It seems to do so quite well. The elliptic central obstruction seems to make the total obstruction about equal to a circular one the size of the major axis.

There are images of what I have called "near field" (see #369), but rather "over-exposed", and not easy to compare (I confess I haven't attempted to look up all the ~360 older posts). There are some better ones in this respect, also by Maskulator, but not with direct comparisons to plain circular central obstructions. I find Vla's recent OSLO simulations much more helpful in this respect.

Nils Olof


Edited by Nils Olof Carlin, 26 December 2015 - 06:15 AM.


#371 mark cowan

mark cowan

    Vendor (Veritas Optics)

  • *****
  • Vendors
  • Posts: 9406
  • Joined: 03 Jun 2005
  • Loc: salem, OR

Posted 26 December 2015 - 02:07 PM

The OSLO simulations are revealing finer detail and appear free of spurious structure so I'm trusting them more.  What Vla has shown convinces me to avoid extremes with any of the masks.  I can see it being useful when it comes to allowing the use of a smaller secondary with the same overall diffractive effects simply because that  smaller secondary costs less and is easier to support.  But I'm not sure where that tradeoff lies at present and may end up just field testing it eventually.


  • brucesdad13 likes 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






Cloudy Nights LLC
Cloudy Nights Sponsor: Astronomics