501 replies to this topic

[cytan299]

Okay, it's clear to me now that "collimation" her just means misalignment, still don't understand why/how that became the case, but so be it.

 

Yes, there's nothing you can do about trefoil since that is not an alignment error but a mounting error.  Also you won't measure it with a 120 deg symmetry anyway.  I have no experience with SCT's, but yes, if you can only adjust the tilt of one mirror then you can't correct both astig and coma.  Astigmatism comes from attempting to correct for coma with the wrong motion, for example trying to tilt coma out when the coma was due to a decenter error.  I guess my point was that you can see both coma and astigmatism with this type of mask whether or not you can correct them both.

 

The big reason I brought this up, btw, is because I am polishing up a paper for publication about using the Bahtinov mask to measure collimation (in the parallel rays sense) so I've been looking around to see what else might have been looked at with the B-mask and collimation (again, the parallel rays sense).

Hi wpullen,

 

   I assume that when you say " you can see astigmatism" is when you look at the effect of the diffraction pattern from the Tri-Bahtinov inside and outside focus. I can't see why you can see an asymmetry in the Tri-Bahtinov spikes when you are focused which is where it is normally used.

 

   Do you mind posting a link to your pre-print when you are done? Hiding among us here are mathematicians, physicists, optics engineers and advanced amateurs who would appreciate (or criticize ) your insights. 

 

   

cytan

Edited by cytan299, 10 July 2020 - 07:32 AM.

[Oregon-raybender]

I cut black cast acrylic 1/8 or 3mm should do. I have a Boss 1420 / 65w  for model and sign making.

 

Starry Nights

[wpullen]

Have you looked at Carey masks? I did, and I'm hooked. Much easier to interpret than Bahtinov.

Yes, I looked into the Carey mask.  That's a cool extension as well.  Ultimately this really just provides a better visual indication, when coupled with software processing the intersections, there might be some, I don't think there will be that much of an advantage.  One advantage certainly is the balanced light level in each of the four diffraction spikes compared with the Bahtinov which, by default, has more in the central spike (unless you block more light from that portion of the mask of course).  This would make it a little bit easier in software to post-process I'd image.

 

Ultimately it's the same thing though, it's only an indicator of a defocus error.  It may be important to note that, for either of these masks, spherical aberration will shift the diffraction spikes in very much the same way (until the spherical gets too big of course).

 

This is essentially my point about the importance of collimating a telescope (in the parallel rays sense).  Imagine you put a perfect point source at the focus of your telescope and watch the rays propagate through you scope.  If the primary and secondary mirrors are aligned axially then you get a perfectly flat wavefront (collimated beam) out of your scope.  If the tilt and decenter between the two are good, but there's a spacing error between the two, then your output beam is either converging (focusing) or diverging.  You can make the beam collimated again (parallel rays) by changing your point source location, but there will be spherical aberration into the wavefront.  This is equivalent to looking at a star, adjusting focus until the Bahtinov, or Carey mask shows "perfect focus", but if the primary and secondary spacings are off, you still have spherical aberration in your image.

 

In practice maybe commercial telescopes are good enough that due to the seeing limit and such it's not a big enough factor to matter (remember I'm very much a novice when it comes to actually using commercial telescopes).

[wpullen]

Hi wpullen,

 

   I assume that when you say " you can see astigmatism" is when you look at the effect of the diffraction pattern from the Tri-Bahtinov inside and outside focus. I can't see why you can see an asymmetry in the Tri-Bahtinov spikes when you are focused which is where it is normally used.

 

   Do you mind posting a link to your pre-print when you are done? Hiding among us here are mathematicians, physicists, optics engineers and advanced amateurs who would appreciate (or criticize ) your insights. 

 

   

cytan

 

Describing this in the same way you did in your background explanation on Github considering an optical path length error:

 

A defocus error is just a radius of curvature error, or difference from the reference wavefront in the exit pupil (optical path difference).  So the single Bahtinov mask is showing what that RoC error is as an average over the full pupil.

 

What your tri-Bahtinov mask is doing is splitting the pupil into three off-axis sub-apertures, separated symmetrically by 120 degrees, so it is measuring the RoC, or defocus error, over each of those three subapertures.

 

Astigmatism is an RoC error in two orthogonal directions.

 

Imagine you had added another Bahtinov mask to your tri-mask making it a quad-Bahtinov mask, one every 90 degrees.  If this mask was lined up with the astigmatism in your system at your camera was at the median focus, the top-bottom diffraction pattern would show a defocus error in one direction while the left-right pattern would show the same amount of defocus error in the opposite direction.  If the clocking was wrong you can still back out the amount of astigmatism, it just won't be equal and opposite in the two orthogonal directions.

 

Back to the tri-Bahtinov mask, you still have information about the two orthogonal axes so you can still see astigmatism error, it just won't be as obvious to deduce by eye what that should look like.  Especially when you through coma into the mix which is only single-fold symmetry.

 

So let's say one of your tri-mask sections is aligned with the direction of astigmatism (no other aberrations) and you corrected that patter for defocus.  What you just did is moved to the sagittal (or tangential) focus and the other two patters are halfway in between the same, defocus corrected axis and the orthogonal, out of focus axis, so that will make the other two diffraction patterns shift, but not as bad as the full astigmatism error. 

 

If you had trefoil and no other aberrations, taking a measurement with the tri-mask in one orientation, then rotating the mask 60 degrees and taking another measurement, you'd be able to back out the amount of trefoil (or equivalently make a hex-Bahtinov mask).

 

The problem with spherical aberration is that it is radially (azimuthally) symmetric so you'd need a separate Bahtinov mask at different radial positions.

 

Meh, I just went on longer than your question was really asking... oops...

 

 

I will definitely post the paper here when it's finished!   And I'll definitely take any criticism, corrections or suggestions for improvement or followup!

[wpullen]

I cut black cast acrylic 1/8 or 3mm should do. I have a Boss 1420 / 65w  for model and sign making.

 

Starry Nights

Well that's a nice tool to have!  I just use my wife's Cricut machine and make masks out of cardstock...

[Psittacula]

Hi wpullen,

 

Previously, in another thread, an issue was raised regarding the sensitivity of tri-bahtinov masks. So I've been looking at ways to improve sensitivity to misalignment and focusing on cytan's original patterns.

 

In the process of trial and error, I confirmed the following by experiments. That is, simply rotating the orginal bahtinov mask 120 degrees around the center of the aperture is not sensitive to misalignment.

 

On the other hand, it was this pattern explained on this page that confirmed the higher sensitivity. I've included a little speculation in the Consideration chapter about why this pattern improves sensitivity.

 

This finding may help us to better understand the tri-bahtinov mask.

 

Satoru

Edited by Psittacula, 11 July 2020 - 12:00 PM.

[wpullen]

Hi wpullen,

 

Previously, in another thread, an issue was raised regarding the sensitivity of tri-bahtinov masks. So I've been looking at ways to improve sensitivity to misalignment and focusing on cytan's original patterns.

 

In the process of trial and error, I confirmed the following by experiments. That is, simply rotating the orginal bahtinov mask 120 degrees around the center of the aperture is not sensitive to misalignment.

 

On the other hand, it was this pattern explained on this page that confirmed the higher sensitivity. I've included a little speculation in the Consideration chapter about why this pattern improves sensitivity.

 

This finding may help us to better understand the tri-bahtinov mask.

 

Satoru

 

That's a correct statement about the original Bahtinov mask since you are sampling the entire pupil.  Think about a Shack-Hartmann wavefront sensor with a single aperture the size of the pupil (basically a single lens at a pupil plane, the size of the pupil).  In this case you only get the lowest-order slope information (i.e. tip/tilt).  The original Bahtinov mask is analogous to the single-aperture SH-sensor, only its lowest (and only) mode is defocus.  Upping the number of Bahtinov mask patterns over more sub-apertures of the pupil has a similar effect to increasing the number of SH subapertures.  You're sampling the pupil more so you get higher wavefront frequency information.  There's a limit to how far you can push this due to the size of the pattern required (how small you can cut the pattern before you need to go to photolithography, which gets way too expensive for such a mask.  If you're going to spend that kind of money you might as well design a CGH specific to your telescope that gives you everything you need.).

 

Note, by way of example, that if your two-mirror telescope's primary and secondary both have power and their spacing is wrong, you get spherical aberration, which, on small scales, has the same effect on the Bahtinov mask diffraction pattern as a defocus error, the two are indistinguishable with a Bahtinov mask.

 

(I just remembered I can attach images so I'll add something in here later...)

 

Also your "this page" link seems to be broken for me...

[Psittacula]

 

Also your "this page" link seems to be broken for me...

From time to time I've been told that pages are inaccessible, so I created a mirror on githubPages.
See the following page of the report on sensitivity improvement.

• https://satakagi.git...ensitivity.html

 

By the way, I also prepared mirrors of mask generator webApps.

• Original: https://satakagi.git...i-Bahtinov.html
• Improved Sensitivity Version: https://satakagi.git..._symmetric.html

 

Satoru

Edited by Psittacula, 15 July 2020 - 11:24 AM.

[meegja]

I had a Tri-Bahtinov made for my scope. I used the original version. I really don't know why I preferred that, was more or less a gut feeling

And it is so much easier to use then the normal way. Reading spacing between spikes is so much more intuitive then reading circles. Took me a little while to learn and read the pattern but after that it is so easy.

I made a huge improvement now with collimating but the final touch is too tiny to read. Have to wait for real clear and still skies ... or software that can read it 

 

Menno

[Ben Diss]

I've a head-scratcher I'm hoping the collective tri-Bahtinov mask people can help with. I've been working to improve my Edge11 images and I tried something interesting this evening before the clouds rolled in. I got the alignment in the center relatively close, and then I moved the bright star I was using to the edges to observe the pattern the spikes would make. I have no idea how to interpret this observation, so I'll post these pictures here for others to comment on.

 

 

--

--

--

--

Edited by Ben Diss, 16 July 2020 - 11:07 AM.

[meegja]

I've a head-scratcher I'm hoping the collective tri-Bahtinov mask people can help with. I've been working to improve my Edge11 images and I tried something interesting this evening before the clouds rolled in. I got the alignment in the center relatively close, and then I moved the bright star I was using to the edges to observe the pattern the spikes would make. I have no idea how to interpret this observation, so I'll post these pictures here for others to comment on.

First: could you use JPG images? You use PNG images of 74 Mb each and those are on a very slow server with a download of 74 kb/s

Second: I am no expert but it could be caused by deformed star images near the edges? Do you normally use some kind of correction?

[tmiddendorf]

Has anyone experience comparing using masks whether Bahtinov, 3-hole, or a combo against software solutions such as FocusMax, Maxim D/L, or CCDware for either focus and/or collimation? If so what is your analysis and opinion of either method/solution?

[Ben Diss]

First: could you use JPG images? You use PNG images of 74 Mb each and those are on a very slow server with a download of 74 kb/s

Second: I am no expert but it could be caused by deformed star images near the edges? Do you normally use some kind of correction?

I'm working to improve the star shapes in the corners, so I think it indicates something. I just don't know what. I suspect it's field curvature resulting in out of focus stars at the edge.

 

Oh, I changed the pics. Sorry about that.

[meegja]

I'm working to improve the star shapes in the corners, so I think it indicates something. I just don't know what. I suspect it's field curvature resulting in out of focus stars at the edge.

 

Oh, I changed the pics. Sorry about that.

Did you have anything in between your scope and the camera when using the Tri-mask?

Edited by meegja, 16 July 2020 - 11:28 AM.

[Ben Diss]

Did you have anything in between your scope and the camera when using the Tri-mask?

The image train is Celestron .7 reducer, Celestron OAG, Baader filter drawer, camera. Maybe I should remove the reducer and diagnose without it.

[meegja]

The image train is Celestron .7 reducer, Celestron OAG, Baader filter drawer, camera. Maybe I should remove the reducer and diagnose without it.

Yes. The disadvantage of the Tri-mask is that is shows every misalignment in the whole setup. So indeed first use it with the camera as close as possible to the scope and then check the collimation on a star in the middle. If the scope is collimated, then add something to your train and then you can check if something in the train is causing a misalignment..

In your case I also suspect that the reducer is causing the thing you are seeing now.

[Ben Diss]

Yes. The disadvantage of the Tri-mask is that is shows every misalignment in the whole setup. So indeed first use it with the camera as close as possible to the scope and then check the collimation on a star in the middle. If the scope is collimated, then add something to your train and then you can check if something in the train is causing a misalignment..

In your case I also suspect that the reducer is causing the thing you are seeing now.

 Well, I have no problem collimating the center, what I'm challenged with is the corners. I'm curious what the pattern may be telling us, particularly since the top and bottom are so different.

[meegja]

 Well, I have no problem collimating the center, what I'm challenged with is the corners. I'm curious what the pattern may be telling us, particularly since the top and bottom are so different.

Could be that the image-train is slightly out of the center or slightly crooked. But why you are missing spikes then ... that I don't know.

[Jinux]

Interesting to see that this is EXACTLY happening to me when I use Focal Reducer's'/Field Flattener's'.

I put 's' because I stacked conventional 0.63x SCT FR/FF with 0.8x FR/FF from refractors. I reported the result on EAA forum but it seems there isn't much of interest while I think result is quite decent though with two inexpensive FR/FF (with bunch of rings) achieving up to F/3.7(!) from F/10 SCT with 25mm bright image circle and ~30mm maximum coverage enough to cover APS-C without much of distortion at edge.

 

https://www.cloudyni...on-sct-063x-08/

 

Anyway, I 'think' I understand the issue but want to hear from other's opinion. I'm not in expert of optics by any means.

It seems light rays are limited by baffle tube (or rings, or reducer frame, I'm looking) in the corners when you use FR/FF. If you look at the images of all four corners, locations of missing spike are different on each corner while there is no missing middle spike. You can duplicate this issue by putting hand (or some covering) on one (or two) outer regions of slits when star is at center because each region of the mask patterns is corresponding to each spike in whole pattern.

To confirm this, you can defocus the star "a lot" enough to make star image show secondary mirror's black dot (shadow) without mask and move stars to corners.

You'll see secondary shadow seems tilted to one side, or in other way to put is one side of outer main mirror disc is not visible, making secondary shadow is tilted.

And location of 'shadowed main mirror disc' will be different per corners. As I explained above, because each mask pattern is corresponding to each spike of diffraction pattern, once some part of main mirror light doesn't reach to focal point, corresponding spikes will not show up in image. This is 'MY' interpretation of the symptom.

The effect goes small when you're at focus but outer edge stars are not exactly round and shows similar to obstructed main mirror disc shape.

 

Since EDGE optics are in the end field flatter embedded and you put 0.7x FR afterwards, I guess you're experiencing the same effect I'm seeing with conventional SCT with heavy FR/FF (or just FR/FF, I haven't checked this thoroughly yet).

 

-Jinux

[cytan299]

Interesting to see that this is EXACTLY happening to me when I use Focal Reducer's'/Field Flattener's'.

I put 's' because I stacked conventional 0.63x SCT FR/FF with 0.8x FR/FF from refractors. I reported the result on EAA forum but it seems there isn't much of interest while I think result is quite decent though with two inexpensive FR/FF (with bunch of rings) achieving up to F/3.7(!) from F/10 SCT with 25mm bright image circle and ~30mm maximum coverage enough to cover APS-C without much of distortion at edge.

 

https://www.cloudyni...on-sct-063x-08/

 

Anyway, I 'think' I understand the issue but want to hear from other's opinion. I'm not in expert of optics by any means.

It seems light rays are limited by baffle tube (or rings, or reducer frame, I'm looking) in the corners when you use FR/FF. If you look at the images of all four corners, locations of missing spike are different on each corner while there is no missing middle spike. You can duplicate this issue by putting hand (or some covering) on one (or two) outer regions of slits when star is at center because each region of the mask patterns is corresponding to each spike in whole pattern.

To confirm this, you can defocus the star "a lot" enough to make star image show secondary mirror's black dot (shadow) without mask and move stars to corners.

You'll see secondary shadow seems tilted to one side, or in other way to put is one side of outer main mirror disc is not visible, making secondary shadow is tilted.

And location of 'shadowed main mirror disc' will be different per corners. As I explained above, because each mask pattern is corresponding to each spike of diffraction pattern, once some part of main mirror light doesn't reach to focal point, corresponding spikes will not show up in image. This is 'MY' interpretation of the symptom.

The effect goes small when you're at focus but outer edge stars are not exactly round and shows similar to obstructed main mirror disc shape.

 

Since EDGE optics are in the end field flatter embedded and you put 0.7x FR afterwards, I guess you're experiencing the same effect I'm seeing with conventional SCT with heavy FR/FF (or just FR/FF, I haven't checked this thoroughly yet).

 

-Jinux

Hi Jinux,

   I think that's a very good explanation, i.e. it is vignetting. The only way the spikes disappear is from an obstruction. If you rotate the mask, I bet that the missing spike returns and the other one disappears. I had originally thought that the obstruction is on the edge of the "tube" shown here:

 

 

If the op can take a flat, I bet you can see the "obstruction". My guess is that the "tube" is either not straight (sagging) or is displaced from the optical centre.

 

cytan

Edited by cytan299, 16 July 2020 - 01:05 PM.

[Jinux]

Ben,

 

I rotated your center image and converted to gray, adjusted contrast/brightness so that my program can detect the lines properly and did analysis. (You can only trust the direction and amount, not the numbers as I don't have necessary infos like your 1:1 image,focal length nor sensor pixel size.) 

 

It shows Red side is tilted to right and Blue side is off center to left. Please refer this with rotated image and your original to find corresponding patterns of yours.

You can actually visually see middle spikes are not in the exact middle in your corner images. This is 'part' of the problem of your corner stars and you can improve it a bit. But I think one of the main reason you're not satisfied on corner stars are from clippings by FR. 

I attached corner crop captured by full frame DSLR when I tried F/4 FR/FF stacking. (You can find full image from the link I put on previous post) You can see blue circled star is reasonably round but red circled ones are not round with flat side facing to outer edges. If your corner stars has clipping enough to missing spikes, I guess your star shape will be similar to this at the corners. Interesting thing is, stock EDGE optics and their plain 0.7x FR shows that much of clipping. I may have to try 0.63x FR/FF tonight to see how much clipping is happening on my conventional SCT.

 

-Jinux

 

ps. My program has too many flaws and precarious to distribute to anybody. Working on it with one volunteer as alpha tester but we hit immediate road block from ASI ASCOM driver is not responding to my expectation where I don't have any ASI cameras. Dang.

 

Attached Thumbnails

• 
• 
• 

[meegja]

-Jinux

 

ps. My program has too many flaws and precarious to distribute to anybody. Working on it with one volunteer as alpha tester but we hit immediate road block from ASI ASCOM driver is not responding to my expectation where I don't have any ASI cameras. Dang.

If you need one more, tiny bit experienced, alpha tester with 8" f/10 SCT and ASI385MC and ASI294MC Pro camera ... I volunteer

Edited by meegja, 16 July 2020 - 03:02 PM.

[Ben Diss]

Hi Jinux,

   I think that's a very good explanation, i.e. it is vignetting. The only way the spikes disappear is from an obstruction. If you rotate the mask, I bet that the missing spike returns and the other one disappears. I had originally thought that the obstruction is on the edge of the "tube" shown here:

 

track.jpg

 

If the op can take a flat, I bet you can see the "obstruction". My guess is that the "tube" is either not straight (sagging) or is displaced from the optical centre.

 

cytan

 

I'll go do that. What do you think the flat will look like if there is a center obstruction? 

[cytan299]

I'll go do that. What do you think the flat will look like if there is a center obstruction? 

I think you will see that the centre will be brighter than the edges but this bright centre will look displaced from the centre. My guess from your photos is that you'll see a larger darker band at the top than at the bottom. That's just a guess -- I'm not putting any money on it

 

cytan

[Ben Diss]

I think you will see that the centre will be brighter than the edges but this bright centre will look displaced from the centre. My guess from your photos is that you'll see a larger darker band at the top than at the bottom. That's just a guess -- I'm not putting any money on it

 

cytan

 

****. Yea. That's what my flats look like and that's what CCDInspector curves look like.