Help with calibration frames and processing - NGC 2903

[Alexp]

Hey guys,

I require some help here as I unfortunately do not know what the issue is or how to resolve it :/. Recently I shot 185 frames @60 seconds, 100 gain, 13 offset and -10°C on NGC 2903. I am using a SW 150PDS with an ASI533 mc pro, a Baader MPCC Mark III with a screw on adapter to the SW150PDS to avoid the potential tilt of the MPCC. All of this is sitting on an AVX. With that I also shot 50 flat frames (I used the CCD flats aid in APT and set the target ADU to 19000) and 50 darkflat frames. In this case the flat frames and the darkflat frames had a length of 0.0062 seconds. Master dark I used from my dark library (50 darks, same length, gain, offset, and temp as lights). Since the ASI533 mc pro is very new - I am using it since two months only - the dark library should still be very fine to use.

 

Since I have the ASI533, this is the way I have always shot targets and so far this has always worked. Due to bad weather though, i only have two other finished pictures on the Crab Nebula and the Iris Nebula but everything has worked out fine with imaging and creating calibration frames this way.

For anyone interested, here is a wetransfer link to the .xisf file. Due to the filesize I have only uploaded the non drizzled master light, the issue is still the same though.

After having done calibration and integration of NGC 2903 I ended up with a horrible result (I used Dynamic Background Extraction on it, then stretched it and saved as JPEG to make the issues visible) that I have attached to this post. I am highly assuming that something is not working correctly with the calibration frames as I can see remaining dust motes and a horrible lot of noise... I hope there is someone here who could maybe point me in to some direction as to what I may be doing wrong. I believe that either I have done something wrong in APT with shooting the flat frames or something in the calibration and integration process is wrong. Although the only thing that really "changed" recently is that I updated APT to the newest version and therefore lost all my settings.. 

If anyone could help me here and point me in the right direction I would be very thankful for any help! 

For image calibration and integration I usually do the following steps (part of it comes from Kayron Mercieca's excellent lightvortex astronomy site, part of it comes from Linda's Youtube video on calibrating with darkflats:

• integrate darkflats to make a master darkflat 
  • combination: average
  • normalization: no normalization
  • weights: don't care
  • rejection algorithm: linear fit clipping
  • normalization: no normalization
  • uncheck evaluate noise
  • uncheck clip low range
• calibrate flats with master darkflat
  • uncheck master bias and master flat
  • use master darkflat as master dark
    • uncheck calibrate and optimize
• integrate calibrated flats to make a master flat
  • combination: average
  • normalization: multiplicative
  • weights: don't care
  • uncheck evaluate noise 
  • uncheck clip low range
  • rejection algorithm: linear fit clipping
  • normalization: equalize fluxes
• calibrate light frames with master dark and master flat
  • uncheck master bias
  • use master dark from dark library
    • uncheck calibrate
    • uncheck optimize
• cosmetic correction on calibrated light frames 
• debayer light frames
• run through subframe selector to reject some lights and put weights on the remaining ones
  • in this case I ended up with 164 light frames to use out of the 185 I shot
• run the weighted lights through star alignment (also create drizzle data, rest of settings are standard)
• create local normalization files, standard settings
• integrate light frames with drizzle data and local normalization data
  • combination: average
  • normalization: local normalization
  • weights: FITS keyword 
  • leave evaluate noise checked 
  • rejection algorithm: linear fit clipping
  • normalization: local normalization
  • uncheck clip low range
• use drizzle integration afterwards on updated drizzle files with local normalization files

This usually gives me an image file to work with. First thing I then do is Dynamic Crop, then Channel Extraction and LinearFit to the lowest signal channel (in this case it was the green channel), before moving over to Dynamic Background Extraction, where I am having the issues in this case.  

 

I hope that I am simply missing something as it would be sad to not being able to use these light frames at all! 

Cheers,

Alex

 

 

 

 

[wimvb]

The bright edge and dust motes indicate that your flats are over correcting. This is most likely caused by faulty calibration of the flats, or by bad darks. The flat exposure time of only 6 millisconds is very short. I suggest that you reshoot the flats with a longer exposure time (cover the scope with white fabric or paper). As for your calibration/stacking process, try with the wbpp script. Adam Block has made an excellent tutorial about this. Available on Youtube.

The second issue is walking noise. You avoid it by dithering between exposures. Make sure you use a large enough dither step.

Btw, with the ASI533 and the 150PDS,  you don't gain anything by drizzle integration. Your image pixel scale is 1"/p (3.76*206/750). Drizzle works on undersampled images, which you most likely don't have.

[Alexp]

Thanks Wim for you input! 

Quote

The bright edge and dust motes indicate that your flats are over correcting. This is most likely caused by faulty calibration of the flats, or by bad darks. The flat exposure time of only 6 millisconds is very short. I suggest that you reshoot the flats with a longer exposure time (cover the scope with white fabric or paper). As for your calibration/stacking process, try with the wbpp script. Adam Block has made an excellent tutorial about this. Available on Youtube.

Oddly enough this is how I have been doing everything so far and it has worked out up until now. I guess I'll try and change that somehow. I am working with a flat panel, but unfortunately with one that cannot be dimmed, hence the short exposure time. I guess I'll try to find a workaround to use the flat panel and still increase the exposure time on the flats. I guess I'll try the WBPP and check out Adam's video as well, thank you!

Quote

The second issue is walking noise. You avoid it by dithering between exposures. Make sure you use a large enough dither step.

I forgot to add that I have been dithering every third shot as well. I do not have the numbers on dithering distance in my head but I have worked with this way up until now and also not have had this problem yet. I do have to crop the sides of the pictures because of dithering so I assumed up until now that the dithering steps were fine. It seems I'll increase the dithering distance and I will also dither between each shot. 

Quote

Btw, with the ASI533 and the 150PDS,  you don't gain anything by drizzle integration. Your image pixel scale is 1"/p (3.76*206/750). Drizzle works on undersampled images, which you most likely don't have.

Also here thanks for the input! My train of thought was to drizzle so I can increase the dimensions of the pictures in case I'd like to print them out on e.g. A1 format since I thought the 9mp resolution of the camera is rather low. But to be honest I am not really sure if my train of thought is correct here or not o_O.

I still find it very odd that the way I have been shooting images has worked so far and now suddenly it doesn't anymore. I have attached a picture of M1 here to show the difference (there are other issues in this picture, but definitely not the same issues as the ones I am currently having with NGC 2903)

Thank you very much for your input, I will try to somehow dim the flat panel to increase the exposure length and I will start to dither between every frame and increase the dither distance.

Cheers,

Alex  

[Alexp]

Come to think of it, since I updated APT and lost all my settings, it is highly likely the dither settings were reset and unfortunately I do not remember to what I had changed the dither settings to before updating APT.... I guess I'll research this again to work out the proper dither settings with my guiding setup resolution and my imaging resolution.

 

Since my guiding setup image scale is 4.3 arcsec/pixel and my imaging setup image scale is, as you already mentioned, 1.03 arcsec/pixel, I left the PHD2 dither scale at 1 and set the APT dithering distance to 4. If I am not mistaken this should move the scope around 12 pixels (= (4.3/1.03)*3) when dithering, which I hope should be enough. 

 

Cheers,
Alex

Edited February 26, 2022 by Alexp
Update

[wimvb]

3 hours ago, Alexp said:

I will start to dither between every frame and increase the dither distance.

If you use blink on the images before image registration, then you shouldn't see the stars drifting in any one direction. If they do, then your dither step is too smal.

You can use a few layers of white cloth (t-shirt) over the scope to dim the light from the flat panel. Try to get an exposure time longer than 0.1 s, preferrably close to 1 s.

[Alexp]

49 minutes ago, wimvb said:

If you use blink on the images before image registration, then you shouldn't see the stars drifting in any one direction. If they do, then your dither step is too smal.

You can use a few layers of white cloth (t-shirt) over the scope to dim the light from the flat panel. Try to get an exposure time longer than 0.1 s, preferrably close to 1 s.

Thanks for the tip. Currently image calibration is running on M1 that I shot the same night. If everything is as it is I should be at least able to see the same issues there as well with the same flat frames and with the same dithering settings that I had. This may be seen as unnecessary but I want to see if I have the exact same issues there. I will afterwards blink the pre registered images of M1 and NGC 2903 to see if the stars drift in any one general direction as second confirmation.

Concerning your tip with the t-shirts, I did exactly as you said shortly after I wrote my last comment. With three t-shirts I at least managed to get the flat frames to 0.14s and I already shot the corrseponding darkflats at -10°C to those as well. I did not move anything in the optical train since having shot both targets so I assume the new flats should also work. After reproducing the error on M1 and after blinking the images I will try and calibrate the light frames with the new flats and darkflats to see if only the walking noise is left. 

I thank you for your input so far! 

Cheers,

Alex 

[Alexp]

My laptop just finished integrating M1 from the same night with the same flat frames and the same dithering settings. The image is stretched a bit more than NGC 2903 but both overcorrection from the flatframes and the walking noise is still visible. I will now try and calibrate the light frames with the new flat frames to see if this problem will be mitigated then, which also means that only walking noise should be left and that problem I can only resolve in the next imaging session.

Cheers,

Alex