Aligning images issue in Deep Sky Stacker

[James22]

I have about 2h of images on the whirlpool galaxy that I've put into deep sky stacker, all of similar quality to the 1 below (biases, darks and flats provided, the picture below is untouched). When trying to stack the alignment is dreadful and everything is blurred when viewing it in photoshop with an increased exposure. Why would this be? I have read that only 8 stars are needed for an alignment - around the galaxy itself I can make out 10 on the raw unreduced data. Maybe focus related? although I think strong enough. Isolating the best images with little trailing and stacking in DSS doesn't improve things either.

 

Equipment and settings:

Camera - Nikon D3500, ISO 6400, 30s exposure, white balance auto, raw format

Telescope stuff - EQ5 goto (~15arcmin accuracy on goto, target is kept in a similar position throughout observations with little trailing), Skywatcher explorer 200P with a 2x barlow

DSS - recommended settings, stacking parameters are median/kappa-sigma clipping (default values) for everything, alignment neither automatic or billinear. 2% detection with hot pixels searched for gives about 50 stars, 3% about 5-10 across the whole bunch.

Thanks in advance!

Edited April 1, 2021 by James22

[Adam1234]

I take it the image above a single sub? Could you post the stacked image?

[James22]

It is yes, out of 500 subs the best I have been able to get is 4 to stack with the rest's alignment info not being computed. Often I receive the message that only 1 frame will be stacked. I found a cheat on a forum saying to manually increase the brightness manually in the raw/fits settings tab to gain more stars. As far as I can tell it's made little difference only detecting noise as stars with a low threshold.

Here's the 'stacked' one, basically the same. 

[Jamgood]

There's something not right there. With those settings (ISO 6400, 30s exposure) you should see more than there is.

Have you tried imaging without the barlow?

Edited April 1, 2021 by Jamgood

[James22]

I agree - I was surprised too. I find issues getting the camera to focus without the barlow (everything is collimated but I haven't manually shifted any mirrors to gain focus), so have resorted to this

In case it's my settings, 30 subs and all master files are here, if anyone would be so kind to see if they get any offsets computed. They are around the same time. Given what you say @Jamgood I wonder if is simply too little light getting through

https://www.dropbox.com/sh/af3p135eh4ys65l/AABj2a3SP3LMl6jMONifWZLZa?dl=0

[The Lazy Astronomer]

DSS needs a minimum of 8 stars which are common to all frames in order to align and stack them. I.e. it could find 1000 stars in each sub, but if it can't find 8 which they share, it won't stack them. 

Also, I can see some coma in the edges (a optical aberration which is common with newts) - this is distorting star shapes to the point where DSS is unable to recognise them as stars. Possibly a coma corrector is required.

[tomato]

I would also suggest to try imaging without the Barlow, at least to start with. You will get many more stars in the FOV, and the alignment of the frames will be not need to be as rigorous to enable successful registration.

[alacant]

Hi

The master calibration frames do not work correctly. The bias is too bright and the flat over corrects. Not many of your frames have usable data.

Recommendations:

- lose the barlow

- use ISO200

- use a cc

- bias: 1/4000s in a dark room 

- flat: use a light panel with paper to adjust the exposure to over 2s with the camera on aperture priority

- don't use dark frames

- use Siril for calibration, registration and stacking

Good luck and HTH

 

      

Edited April 2, 2021 by alacant

[bottletopburly]

definetley lose the barlow ,you can focus a dslr on a 200p 

[James22]

Thanks for the help everyone. I did the tshirt+elastic band method for flats but didn't realise the flats were so poor. I'll definitely give it a go again without the barlow and see if I can get somewhere. Biases were all done on a 1/4000 (same iso, in the dark) exposure but were in quick succession so maybe they need spacing out more messing with the readout time/temperature of the detector.

Tyvm @alacant for doing that on you pc too appreciate it - I'll stick to all your advice. Why do you say not to do darks? Darks include the bias offset and a master dark is surely best to reflect the electronics for a 30s exposure. Mind elaborating?

[alacant]

12 minutes ago, James22 said:

not to do darks

Dark current is temperature dependent. It is impossible to match light and dark sensor temperature when using a dslr; the temperature varies during the exposures. Their use will therefore add more artefacts and noise. There are dark optimisation algorithms which can be used when all else fails; usually for old sensors with banding.

DSLR: take bias and flat frames and dither between the light frames. Stack with a clipping algorithm.

HTH

[James22]

Shall do. Thanks again!