Orpington Astronomical Society

Astronomy => Astrophotography => Topic started by: JohnH on Oct 23, 2021, 21:55:10

Title: NGC7000 in H Alpha
Post by: JohnH on Oct 23, 2021, 21:55:10

This was taken on 21st October. Unfortunately, I have not got the hang of meridian flip and the line caused by slightly different alignment pre and post shows up badly. With luck I will get some RGB data to see how it looks.

Site - near Bromley South Station Bortle 7 or 8 (full moon).
Camera - ZWO1600 Pro at -15C
Telescope - Sharpstar 15028HNT (150mm f2.8 - focal length 419mm)
Mount - iOptron CEM25P
Control - ASIAir Pro
Filter - ZWO Ha

Subs - 15 * 460 secs

Processed in PI, Topaz AI Denoise and Topaz AI Sharpen as Photoshop filters.

(still annoyed about meridian flip!)



BTW, Really enjoyed the 40th celebration. Thanks to all involved in organising it.)
Title: Re: NGC7000 in H Alpha
Post by: Carole on Oct 23, 2021, 23:27:39
That's a very nice image John, regardless of the flip problems.  What was the cause of that?  Did you let the Scope do it automatically and it didn't land in the same place.
I must confess I have problems with Meridian flips because there is rotation due to cone error, and I have never attempted to try to put this right as I know I will make a mess of it, so I just live with it.

So I either align as close as I can and crop out the overlaps, or if it is large object where I can;t afford any cropping I simply try to avoid imaging both sides of the meridian.

Shame you didn't introduce yourself at the Anniversary meeting. 

Title: Re: NGC7000 in H Alpha
Post by: Apophis on Oct 24, 2021, 07:34:34
I must admit when i first started to image i avoided any target that would flip say midway through imaging as i was uncertain how to do it. Now i am using platesolving all the time i set a timer to let me know what time the flip is needed and pause imaging and guiding then park the scope, i wait the few minutes for the target to cross the meridian and then platesolve back to it and its perfectly aligned.
Pixinsight handles the rotation in its integration.
Obviously totally manual but wouldnt trust automating it.
Spotted a sat trail , just wondering how you stack "Sigma" should take them out?


NB: It might help to slew to a star in the centre of your target before and after the flip as i had great alignment when i ldid it last.
Title: Re: NGC7000 in H Alpha
Post by: garrick on Oct 25, 2021, 07:34:48
Hi John,

I have to admit that I really like Black and White Images. Your image as it is is fantastic with lots of detail. I use the ASI Air as well and the whole process is automated. Any alignment issues are dealt with with you do image registration in Pixinsight. Are you using the WBPP script in Pixinsight?

Looking at the stars I don't think the Meridian flip caused your issue. You have slight star trailing in the middle of your image which could either indicate a polar alignment issue or likely a tracking issue.

Clear Skies.

Title: Re: NGC7000 in H Alpha
Post by: JohnH on Oct 25, 2021, 12:34:34
Thanks all,

The meridian flip was an accident. I was trying out a new ASIAir feature "Plan". I thought I had calculated the number of exposures to stop pre flip but missed by 2 (I had not allowed enough time for the Autofocus routine). When I looked I actually preferred the post flip framing and did an "Autorun" with the new centre.

I have looked at the log -

2021/10/21 18:56:13 Plan Nth American N 1 Start
2021/10/21 18:56:13 [Autorun|Begin] NGC7000 Start
2021/10/21 18:56:13 [Guide] Stop Guiding
2021/10/21 18:56:15 [AutoCenter|Begin] Auto-Center 1#
2021/10/21 18:56:15 Mount slews to target position: RA:21h0m5s DEC:+44°36'9"
2021/10/21 18:56:19 Exposure 2.0s
2021/10/21 18:56:22 Plate Solve
2021/10/21 18:56:25 Solve succeeded: RA:21h0m9s DEC:+44°36'20" Angle = -5.25385, Star number = 1841
2021/10/21 18:56:25 [AutoCenter|End] The target is centered
2021/10/21 18:56:25 Start Tracking


2021/10/21 19:53:05 [Guide] Stop Guiding
2021/10/21 19:53:05 Stop Tracking
2021/10/21 19:53:05 [Meridian Flip|Begin] Wait 10min8s to Meridian Flip
2021/10/21 20:03:13 Meridian Flip 1# Start
2021/10/21 20:03:21 Solve succeeded: RA:21h10m11s DEC:+44°37'45" Angle = -4.72179, Star number = 89
2021/10/21 20:03:21 [AutoCenter|Begin] Auto-Center 1#
2021/10/21 20:03:21 Mount slews to target position: RA:21h0m5s DEC:+44°36'9"
2021/10/21 20:03:55 Exposure 2.0s
2021/10/21 20:03:58 Plate Solve
2021/10/21 20:04:00 Solve succeeded: RA:21h3m42s DEC:+44°31'46" Angle = 175.195, Star number = 92
2021/10/21 20:04:00 [AutoCenter|End] Too far from center, distance = 35%(0.650108°)
2021/10/21 20:04:01 [AutoCenter|Begin] Auto-Center 2#
2021/10/21 20:04:01 Mount slews to target position: RA:21h0m5s DEC:+44°36'9"
2021/10/21 20:04:04 Exposure 2.0s
2021/10/21 20:04:07 Plate Solve
2021/10/21 20:04:08 Solve succeeded: RA:21h0m12s DEC:+44°34'26" Angle = 175.833, Star number = 110
2021/10/21 20:04:09 The Mount has flipped
2021/10/21 20:04:09 [AutoCenter|End] The target is centered
2021/10/21 20:04:09 [Guide] Calibration data Flipped
2021/10/21 20:04:09 [Meridian Flip|End] Meridian Flip succeeded
2021/10/21 20:04:09 Start Tracking
2021/10/21 20:04:09 Wait for Mount Settle
2021/10/21 20:04:14 Start Tracking

So it seems that there is a tolerance in centring the image after the flip. Normally, this would not matter but NGC7000 is huge and occupies the whole frame. It is, apparently, possible to use an image for plate solving rather than a target which may give a better result. Also, I note the different star count when plate solving. Pre flip it is 1841, post flip it is 110. This seems a lot, even allowing for the fact that the top left of the image is very star rich. I suppose a lot could have been cut off in the new framing.

I do think I have a number of other niggles to sort out:

1 I am not certain of my collimation. I think that I have a lot of coma around the edges.

2 Generally, my guiding is within a 2" range but it does occasionally throw a bit of a wobbly and spike to 4". I have not checked the PHD log (PHD is on the ASIAir) but I should be within 20" - 40" on PA according to the alignment routine.

3 I used ESD as pixel rejection with the default settings. Probably this required a tweak and maybe was the wrong method.

4 The telescope sits on grass, good to dampen vibration but soft and there could be a slight degree of settling particularly during meridian flip.

If and when Home Observatory UK come and build my observatory I can set up more accurately and permanently.


Title: Re: NGC7000 in H Alpha
Post by: garrick on Oct 25, 2021, 16:12:16
Noticed you are running 480 second subs. Try reducing this to 180 or 120 seconds. It could resolve your tracking. I am using 180 subs at the moment and am considering going down to 120s. At 180 seconds my stars are over saturated. Also noticed your scope is f2.8. This is RASA territory. You may need to use fast filters designed for fast telescopes like yours. I am not sure the ZWO Ha Filter you are using can handle f2.8.


Title: Re: NGC7000 in H Alpha
Post by: JohnH on Oct 25, 2021, 16:47:28
Thanks Garrick,

I am still feeling my way around and so appreciate all the advice I can get.

I am presently using 0 gain to keep noise to a minimum (I hope). I could up the gain value but I found when trying to image M42 that at Gain 143 my Lum subs were saturating stars at 7 seconds (I think that these Gain readings are arbitrary and do not refer to a measured unit). This is how the histogram looked during my imaging run (just pre flip):


I suspect this is underexposed. Should I take my gain back up? If so I think that ~ 140 gain is optimum for the ASI1600.


Title: Re: NGC7000 in H Alpha
Post by: Roberto on Oct 25, 2021, 16:52:01
Very nice image John! Nicely processed 👍
Title: Re: NGC7000 in H Alpha
Post by: garrick on Oct 25, 2021, 21:53:03
Hi John,

M42 is extremely bright with a high dynamic range and in this respect its a challenging object. A lot of the images you see for M42 has the core totally blown out. At f2.8 you will saturate star very quickly. You are correct, unity gain for the 1600 is 139, but 140 should work fine. Unfortunately you cannot alter the offset in the ASI Air App.

The only way to learn is through experimentation. With M42, you may have to do really short exposures to capture the star correctly, then do longer exposures to get the nebulosity. The image sets can then be combined in Pixinsight. One of the major issues with long 4-5 minute exposures in lighted polluted skies is that you are capturing a lot of the light pollution in your images. If you reduce exposure time and capture more subs, not only do you reduce the light pollution, but when you stack you will increase the signal to noise ratio. If you can capture say more than 60 subs, the Winsorised Sigma algorithm in Pixinsight does a good job on the integration.

Additionally, you may already do this, but dithering your subs using the ASI Pro will help remove noise.

Hope this helps, but as it stands your image is great.

Clear Skies,