Thursday, March 28, 2019

#179 - Tuesday, March 26, 2019 - Finally, a little success!

After some growing pains with my astronomy club's new Celestron CGX-L, I finally saw a glimmer of hope!  I am successfully acquiring images.  Short images, but images nonetheless!

I arrived out at the observatory about 20 minutes after sunset and immediately started setting up a timelapse.  It's fun to watch the dome move as the sky darkens and the stars slide across the sky.  Next, I got the memorial dome opened up, my cameras attached, computer hooked up, and covers off.  I turned on the mount, and the hand controller gave me the same error as last time after I input the time and date -- it said no stars were available above the filter (I checked, it's set for 20-90 degrees in altitude).  So I did an audacious thing to do with a Celestron telescope -- I unplugged the hand controller!  Tonight was going to be a computer-controlled night only.

Goodbye hand controller!

Hello USB!

I got the new Celestron PWI app booted up (it only works for their two USB-enabled mounts, the CGX and CGX-L, at the moment) and connected directly to the mount via the built-in USB-B port (woo hoo Celestron is now in the 20th century!  No more telephone serial ports!)



Then closed out of the alignment routine and opened the Slew dialog so I could rotate the mount 90 degrees for accomplishing polar alignment with SharpCap.  Polar alignment was a little over one arcminute off from last time, which isn't terrible, but isn't perfect either, so I adjusted it a bit. 

Before

After

Once that was done, I flipped back over to the Celestron app and had it split-screen with SharpCap for alignment.  While slewing around, the hilariously short power cable that came with my ZWO ASI1600MM Pro almost broke off of my camera, and it prevented the mount from slewing twice as I furiously tried to get it unhooked from the mount.  There is no good place on the mount to place the power converter box where it can actually reach at all angles!  I'm going to buy an extension cable. It uses the standard 2.1mm barrel connector, so it shouldn't be too hard to find.  To be safe, I re-homed the mount, powered off and on, and re-started alignment.

The camera is a loooong way from the mount.

Once I had about five stars added, I hit the "view alignment" button, but it froze the app!  I waited for several minutes, but it was still frozen.  So I begrudgingly force-quit the app, restarted it, and then had to perform the whole alignment again because I didn't save the first one before trying to click that button.  Lesson learned!  I saved it as soon as it was complete the second time.  That is one nice thing about using the app instead of the hand controller -- as long as no one else touches the polar alignment, I can plug in any one of my computers and re-load the last alignment model.  This will be really helpful for when other people use the telescope and don't hibernate it when they're done!  Should save me some time.


With alignment complete, I had a choice to make.  I wanted to image with my new hydrogen alpha filter, and I really wanted to image the Horsehead Nebula.  It's getting pretty far in the west now, however, and I would only be able to get two hours on it, at best. Another option was the Rosette Nebula, which is a good bit higher, but it's too large for the 0.6x0.9 degree field of view of my ZWO camera on the Meade 127mm ED apo f/9 that's in the memorial dome.  It has a focal length of 1143mm (which, as a refractor, as part of the reason my camera is so far from the mount -- and the fact that it has so much backfocus, the focuser is nearly all the way extended!).  To the east, it's Galaxy Season, mostly the itty bitty ones.  Some of those would love a hydrogen alpha channel, but I don't want something like that for my first Ha target.  So I decided to image it anyway and get what I could, and then maybe find another target at 11 PM.

I slewed over to the Horsehead, but I don't see a Precise Goto option in the Celestron app, so I opened up Sequence Generator Pro to give it a try.  I couldn't figure out how to make that plate solve, though, so I will need to do some further reading.  I also had AstroTortilla handy, however, so I swapped over to that.  As it turned out, it was quite close - only off by about 2 minutes of declination, and some fraction of a second in RA.

I turned on guiding, then took a 3-minute image.  And there it was!


Next, I took a 5-minute image.  It looked great!  Until I zoomed in on the stars...


I checked guiding, and it wasn't absolutely awful, but it wasn't great either, especially with that long focal length.


I took a second 5-minute frame just to check, and it was equally bad.  So I set the exposure time back to 3 minutes, which hurt to do with a hydrogen alpha filter, but hopefully I'll still get some good signal-to-noise ratio once I stack.

I went and checked on it after about a half hour, and the frames still looked good.  I rotated the dome and went back inside -- my thermometer read 35F!  But no breeze, so that was good.  I took a moment to admire the sky.  Earlier in the day, I kept an eye on the distant buildings on the horizon, which I use as a gauge of atmospheric condition for imaging.  It was unusually clear today, which was very promising.  The sky was a rich, clear blue all day, and the Clear Sky forecast predicted 3/4 seeing and 4/5 transparency, and zero clouds, at least until about 1 or 2 AM.  This all meant that the light pollution haze in the west wasn't as bad as usual, and there were quite a few stars and star clusters visible.

While I was warming up in the warm room, I scrolled through available targets in SkySafari.  I came up with two options: M82 (Cigar Galaxy), or the Cone Nebula.  I have been dying to image the Cone Nebula, but didn't get out much this winter to do it.  I did try earlier this winter, but had some issues getting on target.  It is getting lower in the sky was well, but it's higher than than Orion's Belt, where the Horsehead Nebula is.  On the other hand, M82 would just be crossing the meridian, and it has two massive jets that glow brightly in hydrogen alpha light that I have been unable to capture on my DSLR or even my ZWO camera.  I do have a pretty robust dataset on the ZWO camera of LRGB data that I could add the Ha to...but, it's on the Vixen telescope that used to be there, which had a shorter focal length (800mm) and a larger field of view, so I would have to like crop the image right around M82 or something and process with that...more complication than I have time for.  Cone Nebula it is!

The hard part is slewing to it.  It's actually a dark nebula structure right next door to NGC 2264, the Christmas Tree Cluster, and the whole Cone Nebula area is known as NGC 2264.  It's much larger than the FOV of my current setup, so slewing to that would not have the actual cone part of the Cone Nebula within my FOV.  I couldn't get SkySafari to select one of the stars near it either to get coordinates to enter into the Celestron app to slew to.  Then I remembered that I had just installed Cartes du Ciel over the weekend, an observing planning application that can also control your telescope.  So I searched for NGC 2264, zoomed in on that area, downloaded the DSS (Digitized Sky Survey) image for that area, and saw where the Cone itself was.  I right-clicked, told it to center that object on the screen, connected the telescope to the app (really the Celestron PWI app, which is also an ASCOM driver for the mount -- very handy!), and then told it to slew to that spot.  And off it went!  Right to that spot!  I took a test image, and there it was!



Pretty dim, but hopefully it will come out in stacking.  Unfortunately, for some reason, when I had Cartes du Ciel open, it turned off my camera's cooler, and once I closed the app, the cooler started again.  I didn't think CdC talked to the camera, and SGP was still able to trigger it and download images, so I'm not sure what's up there.  As soon as I closed CdC, the cooler kicked back on (I could see the temperature dropping in SGP and hear the fans).  Weird.

I wrapped up at midnight since I needed to be at work the next morning!  It was too bad, since the sky was gorgeous, and the moon wouldn't cross the horizon until after 1 AM.  Unfortunately the weekend is looking rainy, but there may be a night or two next week that clears.

[March 28, 2019]

Processing

Horsehead Nebula

I did a quick-process on the Horsehead Nebula data last night since it was getting late and I needed sleep, but boy did it come out awesome!  

Date: 26 March 2019
Object: Horsehead Nebula
Attempt: 2
Camera: ZWO ASI1600MM Pro
Telescope: Meade 127mm ED APO f/9 (club's)
Accessories: Astronomik Hydrogen Alpha CCD T2 12nm filter
Mount: Celestron CGX-L (club's)
Guide scope: Celestron 102mm
Guide camera: QHY5
Subframes: 17x180s (51m)
Gain/ISO: 139
Stacking program: PixInsight 1.8.6
Stacking method (lights): 
Post-Processing program: PixInsight 1.8.6
Darks: 20
Biases: 0
Flats: 0
Temperature: -30C (chip), 27-35F (ambient)

When this came out of stacking, my jaw hit the floor!  It was so exciting to see a group of dim, noisy images become a thing of striking beauty. I got such better signal on the glowing backdrop of the Horsehead than I have before!  Here are a few previous attempts, in wideband:

Date: 4 November 2016
Object: Horsehead Nebula
Camera: Nikon D5300
Telescope: Celestron C11
Accessories: Orion SkyGlow filter, f/6.3 focal reducer
Mount: Celestron CGE
Guide scope: Orion ST-80
Guide camera: QHY5
Subframes: 17x300s (1h25m)
ISO/Gain: ISO-3200
Stacking program: DeepSkyStacker
Stacking method (lights): 
Darks: 8
Biases: 20
Flats: 20
Temperature: 37F

Date: 17 February 2017
Object: Flame & Horsehead Nebulae
Camera: Nikon D5300
Telescope: Vixen na140ssf
Accessories: Astronomik CLS filter
Mount: Losmandy Gemini II
Guide scope: Celestron 102mm
Guide camera: QHY5
Subframes: 25x180s (1h15m)
ISO/Gain: ISO-1600
Stacking program: DeepSkyStacker
Stacking method (lights): 
Darks: 30
Biases: 20
Flats: 20

I also got much sharper detail this time.  This is all despite the fact that it was well inside the light pollution area of the sky in the west that I don't usually image in, and it was getting low on the horizon!  

The Horsehead Nebula is part of the much larger Orion Molecular Cloud Complex.  The head-shaped structure itself is a dark nebula -- a cloud of molecular gas that absorbs light.  Molecular clouds include both organic and inorganic molecules, and lots of interesting low-pressure chemistry happens there that can't happen on Earth.  It lies about 1,500 lightyears away.  Behind the Horsehead is a cloud of hydrogen gas that emits on the hydrogen alpha wavelength my filter was built for.  The shape of that glowing nebula (also known as IC 434) is due to the stellar winds of stars in that vicinity that sculpt it.

Here are the steps I did in PixInsight:
- SubframeSelector to cut less-than-ideal frames
- BatchPreprocessor
- Bias: Average, linear fit clipping
- Darks: Average, linear fit clipping
- Reference frame: frame5
- Calibrated, registered
- ImageIntegration
- Combination: average
- Normalization: additive
- Weights: SSWEIGHT keyword
- Rejection: Winsorized Sigma Clipping
- Discovered image files out of BatchPreprocessor had lost all data but stars...master dark and bias look fine. Registered and calibrated images look weird
- ImageCalibration for calibration, with master dark & bias from BatchPreprocess
- Frames came out dark again.  Okay, hand-making master dark and master bias
- ImageIntegration for biases
- Average combination
- No Normalization
- Don't care for weights
- Linear fit rejection algorithm
- Saved bias and created superbias
- Calibrated darks with master superbias
- Integrated darks into master dark with ImageIntegration
- Calibrated lights with master dark & superbias with ImageCalibration
- Still weirdly dark!!
- Calibrated with just bias - still dark
- Calibrated with just dark - looks fine.  Something's wrong with the bias
- Pressing on
- Registered with StarAlignment
- ImageIntegration for registered, calibrated lights
- Same settings as above
- Cropped with DynamicCrop
- DynamicBackgroundExtraction
- Denoising with MultiscaleLinearTransform, with stretched lum mask
- Stretched with HistogramTransformation
- Cropped again (vignetting)
- Denoised with ACDNR
- CloneStamp for dust spot
- Done...for now

Cone Nebula

The raw frames for this one looked even dimmer -- I had to crank up the auto-stretch in order to even see it!  But I believe in the magic of stacking...let's see what it can deliver!

All right, so here's a single, raw, screen-stretched frame.  (Screen stretch means there's an estimated stretch that is applied to the image, but it's not actually stretched - the view of the image is stretched so I can see what I'm doing).

Single raw 3-minute subframe of the Cone Nebula

I used just the master dark for calibration again because there is something up with my bias frames.  I'll have to re-take them next time I'm out (even though they look fine).
Calibrated and registered single frame

Next came stacking -- and the air went out of my lungs!
Integrated image (15 frames)

Wow!  It just came right out!  So exciting!!

Then I did the rest of my processing:
- SubframeSelector: didn't eliminate any, but calculated weights
- frame3 has highest weight
- BatchPreprocessor for calibration and registration
- frame3 for registration reference
        - Only used master dark, since something wrong with bias
- ImageIntegration for stacking
- Combination: average
- Normalization: additive
- Weights: SSWEIGHT keyword
- Rejection: Winsorized Sigma Clipping
- Cropped with DynamicCrop
- Denoising with MultiscaleLinearTransform, with stretched mask
- DynamicBackgroundExtraction
- DynamicPSF generation for Deconvolution
- Deconvolution
- Made range mask on stretched copy
- Made star mask on linear image
- PixelMath for range_mask-star_mask to target bright areas of the image
- Used PSF and deringing
- Stretched with HistogramTransformation
- Denoised with ACDNR
- CloneStamp to remove dust spot
- Another round of denoising with MultiscaleLinearTransform
- CurvesTransformation touch-up
- One more round of MultiscaleLinearTransform denoising...with a mask with clipped shadows

Aaaaaaaaaand:
Date: 26 March 2019
Object: Cone Nebula
Attempt: 3
Camera: ZWO ASI1600MM Pro
Telescope: Meade 127mm ED APO f/9 (club's)
Accessories: Astronomik Hydrogen Alpha CCD T2 12nm filter
Mount: Celestron CGX-L (club's)
Guide scope: Celestron 102mm
Guide camera: QHY5
Subframes: 15x180s (45m)
Gain/ISO: 139
Stacking program: PixInsight 1.8.6
Stacking method (lights): Average, winsorized sigma clipping
Post-Processing program: PixInsight 1.8.6
Darks: 20
Biases: 0
Flats: 0
Temperature: -30C (chip), 27-30F (ambient)

So awesome!  The noise is a bit high in this image, but only three minutes with a narrowband filter is less than ideal.  I will need to figure out how to improve the guiding on that mount!

More to come!



No comments:

Post a Comment