Problems with Helix

[Gavin1234]

Hi everyone, I don't know what it is about the Helix nebula but I always seem to have problems with this target.

Its been one of two targets that I've been focusing on for a while now. The other being horsehead, and I only have these problems with Helix.
I seem to have heaps of noise in my frames and my flats don't seem to work as well. The quality of the image just looks grainy and pretty bad in general whenever I try to process it.

Was wondering if anyone had the time to take a look at these and maybe even have a crack at processing them?

There are a couple of big circle marks on them that my flats usually take care but not on this occasion. I think that might be because I've taken the data over a number of nights and while I take new flats most nights I've just used one nights worth of flats for this integration. Maybe I had the camera orientation slightly off on one night? I guess I could try and figure out which batch of lights was causing it and take them out. Or could I create separate master lights with their own set of flats and then integrate them? Anyway I'm not too worried about the big circle marks (they are on every light frame I've ever taken) its the rest of the issues I'm more concerned about.

All of the frames were taken using SGP however, some of them were taken a while back when I was much worse at taking subs, only the recent ones were taken using the framing wizard, with dithering and plate solving.

I stacked them using PixInsight and I started off with 573 x 5 min frames. After using blink and filtering out the ones that didn't look as good I ended up 331 subs. Then I used subframe selector and took out anything with a FWHM score of less than 3.5 which left me with the 259 subs I've used. I created drizzle data and stacked those files for this image. I also did a normal version without using the drizzle files and it was the same.

The details are on Astrobin but this is 259 X 5min subs. Using a Skywatcher Esprit 100 ED mounted on a CGX.

This version is just the stack I haven't done anything at all to it.

https://astrob.in/full/369735/0/

This version I have just cropped and ran a dynamic background extraction on.
https://astrob.in/full/369734/0/

The image I've attached to this has been stretched and processed a bit to give you an idea of what I'm ending up with.

[cometcatcher]

That does look a bit too noisy for 20+ hours. Is this with the ASI071MC Pro or one of the Canon cameras? Do you have a lot of light pollution in the direction of the Helix?

[ChrisV]

The Helix is dim and you have pulled out a lot of detail. But as Kevin says 20hr should be a lot less noisy. Taking note of another thread - I don't want to start giving misguided advice, and trying to process jpgs is not a great idea (I think), but a few things

- You say you rejected anything with FWHM < 3. Is that right? Wouldn't you reject FWHM values that were too large.
- The uneven background could be light pollution, or light coming into the scope somewhere.
- What do your flats look like as there a dust/water droplet like blob.
- Maybe you could just stack the more recent data where you say you got better at it.

[multiweb]

Can you post your master flat?

[Gavin1234]

Thanks everyone.


Yes these were all taken with the ASI071MC. There isn't too much light pollution well no more than any other target anyway.


I rejected anything that had a FWHM higher than 3.5.

Ill try stacking just the newer ones as suggested, its a bit of a shame to waste all that data but if its garbage data then I guess I was never going to be able to use it anyway.

This is a link to a flat contour plot (this is from a few nights ago though).
https://astrob.in/full/369818/0/

This one is my current master flat.
https://astrob.in/full/369820/0/


So what about integrating and calibrating each nights lights with their own calibration frames and then integrating the master lights together? Reckon that would work?

[RickS]

How many calibration frames of each type are you taking? The calibration process actually adds noise and if your calibration masters aren't good quality this can be significant.

Making nightly integrations and then combining them is unlikely to produce a better result. The more subs you can integrate at once the better the rejection algorithm will work.

Are you using the default rejection parameters for ImageIntegration? You may be able to get a better result if you optimise them.

Cheers,
Rick.

[Gavin1234]

Thanks Rick, I think that might be the problem. I usually use 50 flats, 50 darks (sometimes 20) and 250 bias.

On some nights I've taken darks when it was still fairly bright outside. Maybe some light leaked into the image train.

[multiweb]

I can see the dust mote at 4 o'clock from the center of the field, so it's not an internal reflection in your pic, it's your flat over correcting, which points to a scaling problem with your master bias. You could try some pixel math to scale it in PI or ccd stack.

[RickS]

What gain are you using, Gavin? I just did the calculations for my ASI1600mm and ASI294MC to figure out how many flats were needed to get reasonable SNR and the numbers were a bit scary at high gain. I have been running the 1600 at gain of 200 and it turns out I need 252 flats to get my minimum preferred SNR of at least 50:1. The ASI294 at unity gain of 120 is not so bad at 68 flats.

Cheers,
Rick.

[Gavin1234]

I’m using 160 gain. I take flats 0.78 to 0.98 sec exposures depending on what the flats calibration wizard in sgp tells me. Usually I use a white T-shirt with a white iPad screen behind it.

To be honest the artifacts and motes are usually not so much of a problem. When I do the flats right all that stuff comes out as far as I can tell (see my other post just above this one on horsehead). It’s the noise and general graininess of the image that’s got me stumped.

[RickS]

My back of the envelope calculation suggests around 70 subs at half full well, so 50 is not too bad. Too few won't cause artifacts but it will add noise to the calibrated result. Sounds like this isn't the problem though...

[Gavin1234]

Probably worth mentioning that it still happens when I use DSS for integration instead of PixInsight.

I’ll try do very small batches of lights starting with the newer stuff. At least I might be able isolate a certain group of lights.

[ChrisV]

Have I got this right ? You take 50 darks each night. 50x 5min is over 4 hours !

[Gavin1234]

Yeah sometimes I just do 20 and if the weather is good and I don't have to move my scope at all I don't do any. I often start them before it gets dark which is what I was talking about before with potentially some light getting into the camera.

[RickS]

A couple more thoughts...

Have you looked at the SNRWeight of the subs (using SubframeSelector) to see if there is much variation? Do you blink and reject cloud affected subs?

Cheers,
Rick.

[Gavin1234]

Thanks again Rick, yep I blink out anything affected by clouds, ufos and anything that’s significantly brighter than the rest.

I have looked at SNRWeight but I don’t even know if higher is better than lower or vice versa.

[RickS]

Higher is better so long as it's not because of clouds :)

[Gavin1234]

Ran a heap of tests today. I'm getting a bunch of extra noise when I calibrate my lights with my master flats. It doesn't happen when I calibrate with only my master dark and a superbias. Only when I introduce the flats.

So basically I have a choice between lots of noise or dust motes.....

Even when I use flats for a week ago it still happens.

Another problem is the gradient and background rubbish although I think I might have found the cause for this. For a large part of the time I'm taking helix subs my scope would be pointing towards the general direction of an IR security camera on the back of my house. I think this might be causing some of the issues.

[RickS]

Are you taking the flats targeting ADU values around 30K or so? It's a bit tricky with an OSC camera because you'll get different peaks for each colour unless you have a light source that matches the sensor perfectly. You could try increasing the number of flats and see if that helps - as a test make a flat master out of flats from several nights. At high gain you don't get many electrons due to the low full well and you will need a lot of flats to get good SNR in the master flat.

Cheers,
Rick.

[Gavin1234]

Thanks Rick, yeah I’m using flats calibration wizard in SGP aiming for ADU of 29500. I put a white shirt over the scope and I put a white screen on my iPad which I place over the shirt. This usually results in about 0.78 to 0.98 exposure times. I’ve always done 50 flats. I’ll try about 100 flats to see if that helps. Also covering the IR camera has got to help. I use gain of 160.

Not sure if it’s related or not but when I integrate my lights for Helix I often get an error on some of the frames where it says “insufficient signal on frame xx< empty frame?”. I figured it was just because of the lack of stars around helix.