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

Image Artifacts

  • Please log in to reply
27 replies to this topic

#1 MountainAir

MountainAir

    Vostok 1

  • -----
  • topic starter
  • Posts: 120
  • Joined: 02 Oct 2019

Posted 16 December 2019 - 03:37 AM

Hello all,

 

I'm slowly getting experience with astrophotography, but every night I run into something I can't explain.

 

Friday night I gave up on the scope/imaging software I was using (temporarily, anyway) and started using KStars & Ekos.  This software looks more complicated, but it "just worked" when I tried getting all my components working with it and "just worked" when I tried it tonight for actual observing (the first clear night in weeks).  I decided to capture some good images of Andromeda (because it's bright, easy for a beginner).  I knew it wouldn't fit in my FoV, but still thought I'd get great images -- this was the first time I successfully auto-guided.  The problem is I ended up with 10 x 300s exposures that show stars "oddly" like there's a halo on one side.  I don't seem to see this in other photos I've taken, so it was either specific to KStars/Ekos or the auto-guiding of the mount.  Or could it be dithering?

 

Celestron Edge HD 9.25 on CGX

.7x Focal Reducer

KStars and Ekos controlling the mount, focuser, guide cam ASI174MM and main cam ASI294MC-Pro.

The main cam was set to abort exposures if guiding deviated by > 4".  The default is 2, but even though my guiding is generally under 1" it would spike often enough to abort most of the 5 minute subs.

 

See attached image.  Note that because CloudyNights won't let me upload a FITS image, or any image over 500KB (how 1986 is that?), I had to screen shot just two stars south of the M31 Core, near the center of the field.  They look the same in the screen shot as they do in the source image, so don't assume it's lossy JPEG compression or artifacting.  It appears on all the stars across the field.  Any ideas?

 

Thanks!

Attached Thumbnails

  • Screen Shot 2019-12-16 at 12.35.31 AM.png

Edited by MountainAir, 16 December 2019 - 03:40 AM.


#2 Dynan

Dynan

    Skylab

  • *****
  • Posts: 4,187
  • Joined: 11 Mar 2018
  • Loc: NOLA

Posted 16 December 2019 - 04:05 AM

Looks like collimation error since the star looks round (good guiding) but the intensity is off-center.

 

Just my penny.gif penny.gif


  • Swordfishy likes this

#3 sg6

sg6

    Cosmos

  • *****
  • Posts: 8,727
  • Joined: 14 Feb 2010
  • Loc: Norfolk, UK.

Posted 16 December 2019 - 04:26 AM

I suspect file limits are simply to prevent huge detailed images being posted along the lines of "Look what I can do" and the classic "My image is bigger then your image". And it will happen.

 

As to so 1986, most people here are likely pre 1986, so be careful.

 

Looks litl the scope is out of collimation from the images, might be the corrector not flat - the front plate is a corrector not a flat plate so if not seated could cause problems.


  • Astrola72 likes this

#4 Tapio

Tapio

    Gemini

  • -----
  • Posts: 3,084
  • Joined: 24 Sep 2006
  • Loc: Tampere, Finland

Posted 16 December 2019 - 04:28 AM

I agree with Dynan - collimation would be first thing to check.

Also I would check reducer camera distance too. And focus.

(And needs debayering  as well).



#5 MountainAir

MountainAir

    Vostok 1

  • -----
  • topic starter
  • Posts: 120
  • Joined: 02 Oct 2019

Posted 16 December 2019 - 10:26 AM

I suspect file limits are simply to prevent huge detailed images being posted along the lines of "Look what I can do" and the classic "My image is bigger then your image". And it will happen.

 

As to so 1986, most people here are likely pre 1986, so be careful.

 

Looks litl the scope is out of collimation from the images, might be the corrector not flat - the front plate is a corrector not a flat plate so if not seated could cause problems.

My 1986 reference was to AOL, not anyone's age, and was said out of frustration after having to save 4 different versions of the same image and then still having to take a screen shot of it.  There's no need to worry about massive image sizes; most applications resize on upload.  No disrespect meant.

 

 

I agree with Dynan - collimation would be first thing to check.

Also I would check reducer camera distance too. And focus.

(And needs debayering  as well).

 

I should have said, my collimation should be spot-on.  Well, I checked it Friday (on a 15" laptop with Vega de-focused to fill almost the entire screen, the center was off by less than 1 mm).  I suppose it could have changed in one observing session, but I just checked some images I also took last night with M57 and they look much better.  Note, M57 was pretty low in the sky, maybe 25 degrees or so, while Andromeda was near Zenith.  See attached for a star from M57.

 

It's really odd that this happened in one set of images and not another.  IIRC I did use Ekos' autofocus before both sets of images.  I did NOT use a Bahtinov mask last night.

 

I did also check the corrector for dew, but it was clear.  I was using a dew shield set on full, though it ended up not hitting the dewpoint and wasn't needed (I really need that Pegasus Astro Ultimate Powerbox to control this automatically).  I wonder if that could have introduced some kind of distortion, but it barely gets warm enough to feel and I'm pretty sure I had it on for M57, too.

 

Screen Shot 2019-12-16 at 7.18.28 AM.png


Edited by MountainAir, 16 December 2019 - 10:46 AM.


#6 Peregrinatum

Peregrinatum

    Apollo

  • *****
  • Posts: 1,404
  • Joined: 27 Dec 2018
  • Loc: South Central Valley, Ca

Posted 16 December 2019 - 11:29 AM

Do all the stars appear like this or just in the corners?

 

If you just have tear drop shaped stars in the corner it means you are not far enough out to reach the backfocus distance.  Need more spacers between the sensor and the reducer, I have the same scope and reducer, the backfocus is 146mm from the top thread of the FR baffle nut to the sensor.



#7 einarin

einarin

    Vanguard

  • *****
  • Posts: 2,244
  • Joined: 23 Dec 2016

Posted 16 December 2019 - 11:30 AM

Just to make sure:

- you should check collimation without reducer

- your star should be center of screen and not fill the screen



#8 MountainAir

MountainAir

    Vostok 1

  • -----
  • topic starter
  • Posts: 120
  • Joined: 02 Oct 2019

Posted 16 December 2019 - 02:27 PM

Do all the stars appear like this or just in the corners?

 

If you just have tear drop shaped stars in the corner it means you are not far enough out to reach the backfocus distance.  Need more spacers between the sensor and the reducer, I have the same scope and reducer, the backfocus is 146mm from the top thread of the FR baffle nut to the sensor.

Yep, it's all the stars.  My image train from the flange of the focal reducer to the camera sensor is about 155mm, maybe a fraction of a mm more (I need a longer digital caliper, it's 1mm too short).  Since you can't measure the total image train distance with the OAG screwed onto the reducer, I added the distance from the top of the reducer threads to the base of the flange.

 

Just to make sure:

- you should check collimation without reducer

- your star should be center of screen and not fill the screen

I didn't think to check collimation without the reducer, but would that really make a difference?  Also, keep in mind that this didn't seem to affect the earlier picture of M57.

 

The star was centered, but I de-focus it so it's large enough to reduce the effect of measurement error against a tiny image.  I hold a ruler to the screen and adjust collimation until the secondary mirror blind spot is directly in the middle.

 

I don't think this is a collimation issue, but I will definitely check it again the next time I'm out.  If it were, though, it would have affected both sets of images on the same night.  All I did was slew from one target to another.  The only other thing I thought is that maybe a mirror shifted position when I slewed, but that would be a major design flaw.  The collimation screws (OEM) are tight.



#9 OldManSky

OldManSky

    Skylab

  • *****
  • Posts: 4,418
  • Joined: 03 Jan 2019
  • Loc: Valley Center, CA USA

Posted 16 December 2019 - 09:40 PM

As to SCT mirrors shifting position being a major design flaw...well, sure, but that’s what they do. All of ‘em. Even the ones with mirror locks.

And actually the shot you posted from M57 shows the exact same issue, just to a slightly lesser degree.

Pretty darn sure this is collimation and mirror shift.

 

Finally, if you save a JPG with a size of 1200 pixels on the long side, it will pretty much always be under the 500k limit here. :)


  • Dynan likes this

#10 MountainAir

MountainAir

    Vostok 1

  • -----
  • topic starter
  • Posts: 120
  • Joined: 02 Oct 2019

Posted 17 December 2019 - 12:09 AM

As to SCT mirrors shifting position being a major design flaw...well, sure, but that’s what they do. All of ‘em. Even the ones with mirror locks.

And actually the shot you posted from M57 shows the exact same issue, just to a slightly lesser degree.

Pretty darn sure this is collimation and mirror shift.

 

Finally, if you save a JPG with a size of 1200 pixels on the long side, it will pretty much always be under the 500k limit here. smile.gif

I'll check collimation on the next clear night (Wednesday, if the weather holds).  I attributed the slight skewing on M57 to its low altitude above the horizon (atmospheric dispersion), but you all may be right.  Maybe the OTC got knocked, or the dramatic temperature change caused collimation drift.  I'll kick myself if that's the case; since I've owned this OTA I've really never had to collimate it (I've made very minor adjustments 3 times in 4 months, the last being 2 nights prior to these images).

 

BTW: I should have said I did use mirror locks after focusing.  Yep, I know they're not perfect but they help.

 

Thanks everyone, I'll report back mid-week.


Edited by MountainAir, 17 December 2019 - 12:15 AM.

  • OldManSky likes this

#11 WadeH237

WadeH237

    Fly Me to the Moon

  • *****
  • Posts: 6,054
  • Joined: 24 Feb 2007
  • Loc: Snohomish, WA

Posted 17 December 2019 - 09:37 AM

Can you be a bit more clear about your question?

 

There are a few things going on with the stars in the first image.  The most obvious are coma and the Bayer matrix grid.

 

If you are asking about the flares to the upper left, that is coma.  There are three potential causes:  Collimation, spacing and focus.  If stars across the field are all flared roughly in the same direction, then it's collimation.  If the stars in the center look good, but the stars on the edges and corners have flares that point away from the center of the field, then it's spacing between your focal reducer and the camera sensor.  If the stars in the center of the field are soft, potentially with dim centers, then focus is involved.

 

The checkerboard pattern is the Bayer matrix grid and is normal for a raw image from a one-shot-color camera.  When you use your imaging software to Debayer the frame, it will convert the checkerboard pattern into color information.

 

And finally, regarding the above conversation about collimation, you can only do a rough collimation with an out of focus star.  A defocused star is really good at hiding collimation errors.  I would only defocus to a large donut to find a really gross collimation error.  You need to precisely center a star in the field, and check to make sure that it's concentric.  If it's not concentric, then make an adjustment with the collimation screws.  Before deciding if the change helped or not, recenter the star.  This is important.  It is normal for an SCT to show collimation errors for stars that are not in the center of the field (the EdgeHD scopes are better about this, but you still want to always recenter the star after making an adjustment).  Once the star looks concentric, then slowly bring it closer to focus while taking images.  As the donut gets smaller, it will reveal residual collimation error that was too small to see when the donut was bigger.  As soon as you can see the error, then adjust it as above.  Keep doing this iteratively until the star is fully in focus and centered, showing no collimation error.


  • Peregrinatum likes this

#12 Madratter

Madratter

    Voyager 1

  • *****
  • Posts: 12,625
  • Joined: 14 Jan 2013

Posted 17 December 2019 - 09:44 AM

Can you be a bit more clear about your question?

 

There are a few things going on with the stars in the first image.  The most obvious are coma and the Bayer matrix grid.

 

If you are asking about the flares to the upper left, that is coma.  There are three potential causes:  Collimation, spacing and focus.  If stars across the field are all flared roughly in the same direction, then it's collimation.  If the stars in the center look good, but the stars on the edges and corners have flares that point away from the center of the field, then it's spacing between your focal reducer and the camera sensor.  If the stars in the center of the field are soft, potentially with dim centers, then focus is involved.

 

The checkerboard pattern is the Bayer matrix grid and is normal for a raw image from a one-shot-color camera.  When you use your imaging software to Debayer the frame, it will convert the checkerboard pattern into color information.

 

And finally, regarding the above conversation about collimation, you can only do a rough collimation with an out of focus star.  A defocused star is really good at hiding collimation errors.  I would only defocus to a large donut to find a really gross collimation error.  You need to precisely center a star in the field, and check to make sure that it's concentric.  If it's not concentric, then make an adjustment with the collimation screws.  Before deciding if the change helped or not, recenter the star.  This is important.  It is normal for an SCT to show collimation errors for stars that are not in the center of the field (the EdgeHD scopes are better about this, but you still want to always recenter the star after making an adjustment).  Once the star looks concentric, then slowly bring it closer to focus while taking images.  As the donut gets smaller, it will reveal residual collimation error that was too small to see when the donut was bigger.  As soon as you can see the error, then adjust it as above.  Keep doing this iteratively until the star is fully in focus and centered, showing no collimation error.

Top notch post and agree with his advice about collimating.



#13 MountainAir

MountainAir

    Vostok 1

  • -----
  • topic starter
  • Posts: 120
  • Joined: 02 Oct 2019

Posted 18 December 2019 - 10:47 PM

The attached images represent my current collimation, which was used to capture the coma-happy images of Andromeda.  They are screen shots of the Ekos FITS Viewer window, and were taken at 3000-tick focuser increments both inside and outside of focus.  When I was close to focus I zoomed in, with the grid on and off.  How do these look?

 

OK, scratch that.  Apparently even screen shots of the FITS Viewer window are too large @ jpeg 90% compression.  These darn retina displays.  I'll upload the ones that fit.

 

 

Attached Thumbnails

  • Screen Shot 2019-12-18 at 7.00.32 PM.jpg
  • Screen Shot 2019-12-18 at 7.11.17 PM.jpg


#14 MountainAir

MountainAir

    Vostok 1

  • -----
  • topic starter
  • Posts: 120
  • Joined: 02 Oct 2019

Posted 18 December 2019 - 10:48 PM

More pics.  Interesting, can you see the reflection of the star on the left in the smaller image?  Hmmm.

Attached Thumbnails

  • Screen Shot 2019-12-18 at 7.11.10 PM.jpg
  • Screen Shot 2019-12-18 at 7.11.44 PM.jpg

Edited by MountainAir, 18 December 2019 - 11:03 PM.


#15 MountainAir

MountainAir

    Vostok 1

  • -----
  • topic starter
  • Posts: 120
  • Joined: 02 Oct 2019

Posted 18 December 2019 - 10:49 PM

And more.

Attached Thumbnails

  • Screen Shot 2019-12-18 at 7.13.28 PM.jpg


#16 MountainAir

MountainAir

    Vostok 1

  • -----
  • topic starter
  • Posts: 120
  • Joined: 02 Oct 2019

Posted 18 December 2019 - 10:50 PM

Last one, I don't want to push my luck with the admins.

Attached Thumbnails

  • Screen Shot 2019-12-18 at 7.15.55 PM.jpg

Edited by MountainAir, 18 December 2019 - 10:53 PM.


#17 MountainAir

MountainAir

    Vostok 1

  • -----
  • topic starter
  • Posts: 120
  • Joined: 02 Oct 2019

Posted 18 December 2019 - 11:01 PM

My take on these is that they look pretty good.  I can detect a slight deflection from center, but it's so close I would be afraid to try to correct it.  That's what happened the last couple times I collimated -- I tried correcting a very small error and messed it up, taking many, many iterations to get it fixed.

 

I'm going to tear down the scope and take these in for closer analysis in Photoshop.  I'll use pixel measurements to see how far off-center it is.



#18 Dynan

Dynan

    Skylab

  • *****
  • Posts: 4,187
  • Joined: 11 Mar 2018
  • Loc: NOLA

Posted 19 December 2019 - 01:33 AM

I agree on both counts...very, very slightly left of center (last pic) and maybe not worth another try. You might be talking a few hundredths of a turn...

 

Personally, when I see ANY off center indication, I obsess...which is why so many good seeing hours have been lost.lol.gif



#19 WadeH237

WadeH237

    Fly Me to the Moon

  • *****
  • Posts: 6,054
  • Joined: 24 Feb 2007
  • Loc: Snohomish, WA

Posted 19 December 2019 - 12:40 PM

Those stars are too out of focus to for any critical analysis, plus they are stretched so that they are too bright.

 

When you are this far out of focus, a miscollimation will show better as slight brightening somewhere in the donut.  To see this, you want to stretch them so that the donut is really dim.  If you look at the second image in post 13, you can see that there is a faint star on the left side.  Notice how the left side of that star is slightly brighter than the right side?  That's what I'm talking about.  Note that since that star it off center, it's normal for it to show some coma.  I just wanted to point it out, since it's exposed better than the star in the center for this purpose.

 

When I approach final collimation on my imaging scope, the last step before and in-focus check shows the star stretched so that it's a small, very faint ring.  At this point, miscollimation usually shows up as one bright spot on the ring.  Note that you might need to use a fairly dim star for this.

 

If you've ever used FocusMax, it turns out that it does a pretty good job of stretching to the right amount to see miscollimation.  You could use that as an example if you use it.


  • Madratter likes this

#20 Madratter

Madratter

    Voyager 1

  • *****
  • Posts: 12,625
  • Joined: 14 Jan 2013

Posted 19 December 2019 - 03:20 PM

Those stars are too out of focus to for any critical analysis, plus they are stretched so that they are too bright.

 

When you are this far out of focus, a miscollimation will show better as slight brightening somewhere in the donut.  To see this, you want to stretch them so that the donut is really dim.  If you look at the second image in post 13, you can see that there is a faint star on the left side.  Notice how the left side of that star is slightly brighter than the right side?  That's what I'm talking about.  Note that since that star it off center, it's normal for it to show some coma.  I just wanted to point it out, since it's exposed better than the star in the center for this purpose.

 

When I approach final collimation on my imaging scope, the last step before and in-focus check shows the star stretched so that it's a small, very faint ring.  At this point, miscollimation usually shows up as one bright spot on the ring.  Note that you might need to use a fairly dim star for this.

 

If you've ever used FocusMax, it turns out that it does a pretty good job of stretching to the right amount to see miscollimation.  You could use that as an example if you use it.

+1



#21 MountainAir

MountainAir

    Vostok 1

  • -----
  • topic starter
  • Posts: 120
  • Joined: 02 Oct 2019

Posted 19 December 2019 - 06:57 PM

Those stars are too out of focus to for any critical analysis, plus they are stretched so that they are too bright.

 

When you are this far out of focus, a miscollimation will show better as slight brightening somewhere in the donut.  To see this, you want to stretch them so that the donut is really dim.  If you look at the second image in post 13, you can see that there is a faint star on the left side.  Notice how the left side of that star is slightly brighter than the right side?  That's what I'm talking about.  Note that since that star it off center, it's normal for it to show some coma.  I just wanted to point it out, since it's exposed better than the star in the center for this purpose.

 

When I approach final collimation on my imaging scope, the last step before and in-focus check shows the star stretched so that it's a small, very faint ring.  At this point, miscollimation usually shows up as one bright spot on the ring.  Note that you might need to use a fairly dim star for this.

 

If you've ever used FocusMax, it turns out that it does a pretty good job of stretching to the right amount to see miscollimation.  You could use that as an example if you use it.

I probably didn't upload all the best images.  I took photos of that star (which was near Zenith, like Andromeda was) at focus points of 12000, 14000, 16000, 18000, 20000, 22000, 24000, 26000, 28000, 30000.  Focus at HFR 1.67 was at 25,125.  But when the donut is that small, such as 22k, 24k 26k and 28k I have to zoom to see what's going on.

 

You're definitely providing more detail than the other collimation guides I've seen out there.  I've used KStars/Ekos twice now, so I'll figure out how to turn off auto-stretch.  Is there a particular gain you'd suggest I use?  A particular exposure length?  Should I aim for a particular histogram level?  How many X bigger than the focused star should the donut be?

 

Please note:  That light artifact to the right of the star is exactly that -- it is not a star at all.  I think it was some kind of internal light reflection.  I didn't spend much time trying to track it down, but I can tell you when I slewed the scope so the donut was slightly off-center (say, 10 px) that artifact would shift to an entirely different side of the field.  A good wind gust even shifted it around dramatically.

 

Last night was just about testing collimation.  I had a somewhat rough polar alignment and the scope was not well-aligned in general, but after seeing the collimation pictures above I decided to try a shot at M31 again. Attached are two files... the 120s version was, I *think*, unguided.  The 300s version was guided, but it had to try a couple times so the guiding was pretty rough with the alignment.  I was only able to get one before the dew struck (I didn't have my dew shield on last night).  These images -- with the same collimation I used in the crazy images above -- seem much better to me.  Opinions?  Obviously, ignore the drift that's fairly evident.


Edited by MountainAir, 19 December 2019 - 09:26 PM.


#22 MountainAir

MountainAir

    Vostok 1

  • -----
  • topic starter
  • Posts: 120
  • Joined: 02 Oct 2019

Posted 19 December 2019 - 06:58 PM

And the 300s version.  Ugh, these look horrible with 10% JPEG compression!

Attached Thumbnails

  • Light_300_secs_2019-12-18T20-44-53_001_d.jpg

Edited by MountainAir, 19 December 2019 - 07:01 PM.


#23 WadeH237

WadeH237

    Fly Me to the Moon

  • *****
  • Posts: 6,054
  • Joined: 24 Feb 2007
  • Loc: Snohomish, WA

Posted 20 December 2019 - 10:22 AM

Please note:  That light artifact to the right of the star is exactly that -- it is not a star at all.  I think it was some kind of internal light reflection.  I didn't spend much time trying to track it down, but I can tell you when I slewed the scope so the donut was slightly off-center (say, 10 px) that artifact would shift to an entirely different side of the field.  A good wind gust even shifted it around dramatically.

We're looking at different things.  Here is the faint star that I mentioned.

 

As for the camera settings, I don't have any settings for your specific setup.  I wish that I had some images of what I'm describing, but I've never saved any of the stream of images while I'm focusing or collimating.  The brightness of the of the star by the arrow is what you should be looking for.  As you close in on focus, it will get much brighter.  When it does, you will need to change the stretch to dim it down again (you may or may not need to reduce the exposure time, depending on how close it gets to saturation - saturation is bad).

 

Also, I will say that when you say you need to zoom in, that is a good thing.  I'm usually scaling my view to 400% when I get close.

Attached Thumbnails

  • donuts.jpg


#24 MountainAir

MountainAir

    Vostok 1

  • -----
  • topic starter
  • Posts: 120
  • Joined: 02 Oct 2019

Posted 21 December 2019 - 10:51 AM

You were right.  As I shrunk that donut down so the star was only slightly out of focus, the centroid did offset a bit.  It wasn't bad (I failed to capture a picture), but I decided to try to correct it anyway. I'm not entirely happy with it, as it's still slightly off-center (though a little closer than where it was).  This was using my ASI294MC-Pro as a camera whereas I should have used my 2.8 micron Celestron Skyris instead, but I had just quickly thrown the scope up to make an attempt between bad cloud cover.

 

BTW, I had to capture these with AstroImager.  I couldn't figure out how to turn off AutoStretch in KStars (each time I pressed the button it tried to stretch more).

 

Magnified and unmagnified (note: there will be another post to follow):

Attached Thumbnails

  • Screen Shot 2019-12-20 at 10.20.33 PM.jpg
  • Screen Shot 2019-12-20 at 10.21.15 PM.jpg

Edited by MountainAir, 21 December 2019 - 11:08 AM.

  • Dynan likes this

#25 MountainAir

MountainAir

    Vostok 1

  • -----
  • topic starter
  • Posts: 120
  • Joined: 02 Oct 2019

Posted 21 December 2019 - 11:08 AM

Unfortunately this didn't improve my images (see below).  This 300s guided exposure doesn't seem any better than the original.  It looks like total cloud cover here for the next 2-3 days.  I'll see if I can collimate properly on the next clear night, but I have to say it was really difficult to get this close and the images are still horrible.  I will do this properly with no reducer (I just didn't have time to change the image train last night), a good Barlow and my Skyris camera with smaller pixels/more mag.

Attached Thumbnails

  • M 31_Light_300_secs_2019-12-20T22-53-03_001_d.jpg

  • Dynan 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