Dark lines emanating from bright stars (e.g. in M45)

[Epicycle]

Hi there,

I'm hoping some of you have come accross the following issue: when imaging objects with very bright stars

(like M45) the images I have obtained exhibit dark lines coming out of bright stars. An example is

attached. I've done some search and there some threads about this on CN and other places but it's not conclusive what's the cause nor how to deal with it. I fixed it at the very end of processing using the resynthesize feature in Gimp. An example with this fixed is in http://www.astrobin.com/31460/ where you can find also the details of my gear etc.

(although the image there was processed slightly differently from the one attached but it originates from the same stack).

The stack combines 80 exposures of 120s and were taken with an unmodded Canon 1100D.

Today I just finished stacking an image of ngc2403 and when I strech it a lot I can also see dark lines radiating out

of the brightest stars (although nothing as dramatic as in M45).

What bothers me is that I'm about to start guiding hoping to get longer exposures and since this annoying phenomenon

seems to intensify with exposure time, perhaps some of you might have some tips about how deal with it,

Thanks!

E.

[Epicycle]

Apologies for insisting, but there is no Canon user out there which had similar issues when imaging

M45? Thanks,

E.

[GlassWalker]

Can't say that has ever affected me on either of the 450D or 600D I use.

Out of interest, what settings are you using? e.g. ISO, exposure length.

One workaronud I can think of, isn't great, but should work, is to rotate the camera now and then. Thus the dark lines will get moved around, and a stack using some kind of rejection should only pick out the brighter ones.

[Epicycle]

Hi GlassWalker

thanks for the reply! I was using iso 1600 and 120s subs, but I saw the same dark lines with iso 800. I've seen the idea of rotating the camera to average the lines out

with stacking, the problem is taking flats afterwards: changing the orientation of the camera will mess them up...

E.

[GlassWalker]

The ISO and duration are similar to what I use also, was only wondering in case you're using something totally different.

If the optics are rotated with the camera, then I think that would keep flat integrity assuming the optical chain is solid enough not to flop about. Otherwise taking multiple sets of flats would add to the fun...

[russellhq]

I use a 1100D and have not seen this problem. Probably a long shot, but have you checked your USB cable, sometimes it can affect the data...

[Epicycle]

I doubt is the USB. I'll try with the camera rotated 90 degrees and see what happens...

E.

[StarRaver]

I think it could be something to do with dss, i had the same problem and changed a few things in program ( can't remember what now though ) and it sorted itself out

[Epicycle]

Individual frames show the same "bleeding", so I doubt is stacking. I also tried stacking with nebulosity with the same result. Rotating the camera had no effect, so it seem something intrinsic to the sensor...

[ArmyAirForce]

I would run a few subs, then rotate the camera in the focusser. look to see if the lines have moved with the camera, or stayed in the same orientation in relation to the stars. This should give you an indication whether it is a camera artifact, or something within the scope optics. If you have the opportunity, try the camera on someone elses scope and see if you get the same effect.

[Epicycle]

Done that, it's not the scope. Also bright stars of different objects change position, so it has to be the sensor...

[ArmyAirForce]

To create the artifact from each star, I find it odd that it is the sensor itself. Perhaps it is something to do with the signal transmission of say the blue channel from the sensor. Have you tried imaging something with more red and see if it does the same?

[Epicycle]

Here is another image (M101 wide field) which I overstreched to show the same phenomenon. This is a stack

of 24 subs of 5minutes (my first guided picture, yeah!).

It seems to be produced by overexposure of the pixels. What confuses me is that this is known to happen to CCDs but it's not supposed to happen to CMOS (goes under the name, "blooming", "smearing"...)

E.

[Psychobilly]

I used to see the same sort of thing witn my Nikon D200 but havent noticed it on my Canon 350/1000/500 or 7Ds

Anti blooming - although cmos sensor so shouldnt suffer from it...

I would say its an electronic rather than optical problem...

Sent from my GT-P7300 using Tapatalk HD

[Epicycle]

I agree with you Physcobilly, any tips about how to deal with it? (In the CN thread I read, some Canon users were reporting having the problem

and some didn't experience it, I don't know why some get it and some don't),

cheers

E.

[Psychobilly]

Heres a link to a 2006 thread on cloudy nights....

http://www.cloudynights.com/ubbarchive/showflat.php/Cat/0/Number/1264923/page/6/view/collapsed/sb/9/o/all/fpart/1

Sent from my GT-P7300 using Tapatalk HD

[Epicycle]

Right! That's the one I read, but unfortunately there are no tips about how to deal with it (which was my original question),

E.

[WaveSoarer]

To me it looks like streaks in the read-out direction of the CCD chip. If the shutter doesn't close before the chip is read out (basically the charge accumulated in each pixel is shuffled along pixel by pixel in rows to the edge of the chip where it is read) then anything bright, like the stars, will cause a line. We noticed something similar with our X-ray diffractometer due to a synchronisation issue between the shutter and the detector read out. It was fixed by ensuring the shutter was closed before the detector read out. The only slight doubt I have is that the streaks on your images are dark and I would expect them to be bright if the shutter was closing too late.

[Epicycle]

Hi WaveSoarer,

sounds plausible (although Canon sensors are cmos not ccds). I have no idea how to sychronise things with my Canon, but I suppose I could test your hypothesis just by putting the cap on the scope a few seconds before the exposure finishes,

E.

[jrowe30]

Hi Epicycle, 

I've just come across this topic while doing a search after experiencing the same suspected sensor issue with my Canon 1100d. I had the problem after imaging M45 too. Did you find a solution? 

Thanks

[lensman57]

Hi Epicycle, 

I've just come across this topic while doing a search after experiencing the same suspected sensor issue with my Canon 1100d. I had the problem after imaging M45 too. Did you find a solution? 

Thanks

Hi,

I had the same problem with my modded 1100d. The sensor had to be changed. I got the camera new from cheap astrophotography  and Juan replaced the sensor as soon as I had sent the camera back to him.

A.G

PS: Try extreme  dithering with a large number of subs and use sigma clipping for stacking and see if it helps, it won't hurt.

[jrowe30]

Thanks for the reply Lensman57, 

I had a feeling that might have to be the solution. I bought mine a couple of years ago new from PC World but I doubt it'll be under warranty any more, so could be time to look for a new DSLR. Thanks for the post processing tips, I'll give them a go!  

[lensman57]

Thanks for the reply Lensman57, 

I had a feeling that might have to be the solution. I bought mine a couple of years ago new from PC World but I doubt it'll be under warranty any more, so could be time to look for a new DSLR. Thanks for the post processing tips, I'll give them a go!  

I think that even if it were under warranty they would not do anything about it as the camera is highly unlikely to show such defect in terrestrial photography even at night time. In AP we really do put these devices through what they have not been designed for. I was reading about Atik 11000 the other day and the camera at nearly £4000.00 is specified with a grade 2 sensor which means that it will have column defects that dark and bias subtraction will sort out, the option of having a grade A sensor just renders the camera uneconomical. Perhaps if you tried dithering and a good dark and bias subtraction the problem might go away.

Regards and best wishes,

A.G

[Epicycle]

Sorry for not replying earlier, I just saw the thread was revived. I never got to the bottom of it and moved shortly after to ccd imaging...

E.

[Fred_76]

This is a common problem that only affects the Canon 1100D. About 1/3 of the 1000D are affected by these dark lines. It can also be visible on some daylight "high contrast" photos.

There is no solution and Canon never found this to be an issue. They don't replace the cameras nor they will issue a patch.

There is no real solution in the post processing except the brush... however big dithering (at least 5 pixels) will remove it, and some field rotation would also help.