I decided it was going to be too cold to setup my own telescope, so I brought out my camera bags and my Vixen Polarie. My first plan for the evening was to test out a polar alignment rig I cobbled together for the Polarie that I'm going to see if it'll work for my trip to Chile next summer.
I spent part of the afternoon looking through all of the hardware in my house to see what I could put together for the Polarie. My master plan is to hack together basically a DIY Polemaster, which is a camera from QHY that helps do computer-based polar alignment. I've been doing pretty much the same thing with my camera attached to my telescope using SharpCap's polar alignment routine, so rather than spending $300 for the Polemaster, I started looking for ways to get a lens onto my QHY5 and get it attached to the cold shoe on the Polarie. Then I could just use the polar alignment routine in SharpCap with it. I bought a M42-to-C-mount converter with the intent of getting like a 35mm or 50mm CCTV lens, but then I figured I could also use my little Orion 50mm guidescope (focal length 162mm). I had recently bought a dual mount for it so I could have both the guide scope and my red dot finder mounted onto my Borg refractor at the same time instead of having to swap them out, and I noticed that it also had a standard 1/4-20 tap on the bottom. I also recently purchased a couple of camera shoe-to-1/4-20 converters. So I attached one of them to the bottom of the dual finderscope mount, and then attached the guide scope to that. It's not the firmest of connections, and the whole thing is kind of heavy, but for just testing the concept, it'd work.
In addition to the polar camera idea, I also needed a more solid base with fine adjustment knobs, and luckily Vixen makes one especially for the Polarie. It has three altitude settings for 0-30, 30-60, and 60-85, and at each setting altitude can be changed +/-15 degrees. Azimuth can also go +/-15 degrees. The base can attach directly to a tripod either with a 1/4 or 3/8 screw, and it comes with both kinds. I wanted to attach it directly, but I don't think the heads come off of any of the three of my tripods. I'm already shopping for a short, lightweight but sturdy one that does have a removable head. For the moment, I attached it like I would a camera to my sturdiest tripod, my Orion Paragon.
They thought of everything with this mount head - it even has a circular attachment that screws into the bottom of the Polarie so you can just pop it in and out with a set screw. The top where the Polarie attaches is just wide enough for the Polarie, and has raised edges so that it won't rotate by accident.
Next, I replaced the ball head that came with the Polarie with a taller one I picked up from an acquaintance of mine from area star parties while I was at the Hidden Hollow Star Party, Dave. This one is taller and beefier. Then I attached my dual finder mount to the cold shoe on top of the Polarie.
Finally, I put my DSLR on the ball head and my guide scope and guide camera on the finder mount to act as the polar alignment camera.
My Nikon D5300 with a Nikon 55-200mm lens attached to my Vixen Polarie tracker on the Vixen polar fine adjustment mount head, and my Orion 50mm mini-guidescope with my QHY5 red hockey puck guide camera attached.
Here are the links to all the stuff I bought:
It was still bright outside, so I went inside and ate dinner, and then did a quick training to re-qualify on using the memorial dome since we had the new Celestron CGX-L in it. I had pulled out the CGEM when I first got there only to realize that they had indeed already mounted the Meade 127mm apo in the memorial dome, so I couldn't get in there to image until everyone was done re-qualifying and looking through the scope.
Once I was done eating and it was darker, I plugged my QHY5 into my tablet and opened up SharpCap. After replacing the nosepiece on the QHY5 with the one I had a parfocal ring on already adjusted for that particular guide scope, I was getting nice stars, and opened up the polar alignment routine. Using the hole on the side of the Polarie got me close - Polaris was already in the FOV.
I quickly realized a fatal flaw in my plan, however. Normally for this polar alignment routine (and the QHY Polemaster), you rotate the mount 90 degrees in RA, and then it calculates how much you were off. This doesn't make sense on the Polarie, unless I use the main imaging camera for polar alignment, which wouldn't make sense either because it could be pointing anywhere. So instead, I just used the plate solving it does in the first step to show me where the north celestial pole is, and I put that in the center of the camera's FOV.
I'm not entirely convinced that this is a scientifically sound way of doing it because the guide scope is not necessarily boresighted with the polar axis of the Polarie, but it may yet be close enough for 100-200mm lenses for 2-5 minutes.
Once that was done, I carefully removed the guide camera and scope (because it was getting in the way of me seeing through the viewfinder of my DSLR), and then started looking for the Heart & Soul Nebulae, using the bottom star of the W of Cassiopeia as a reference in the viewfinder, and then the Double Cluster as a reference with some 5s exposures. I got lucky this time and nailed focus dead-on with only a few tries, woot! I had a hard time finding it, but eventually I got it mostly in there, or at least as far as I could tell using SkySafari and reference stars. I set up the intervalometer to take 2-minute subframes, but it's been finicky for a while now, and it was acting up again and not working. So finally I just hooked up the camera via USB to my Surface 3 tablet and used BackyardNikon to trigger exposures. I was just starting to re-adjust the target position when the last person was finally finished with the memorial dome, so I grabbed my other laptop and cameras and moved in.
Meade 127mm apo ED riding atop the Celestron CGX-L beefcake mount inside of the club's memorial dome. This is from the end of the night when everything was frosted over.
I hadn't used the laptop in a while, and it has a weird fault where it will run at 100% disk usage for the first 10 minutes or so after it wakes up from sleep or hibernation. This issue persisted through a hard drive replacement and at least one, possibly two clean Windows installs. I aligned visually with an eyepiece while I waited for it to work, and the polar alignment still needs a little work - the first star, Vega, was fairly significantly out of the FOV of even the finderscope. It took a while to find it; we're going to install a Telrad on there too soon, which will help with that situation. Finally I was able to kind of use the laptop (very slowly while the disk was still at 100%). I couldn't get my ZWO camera to show up at first, so I tried reinstalling the driver. It still wasn't working...and then I realized I forgot to plug in the USB cable. -.-
Next, I set to work on focusing the guidescope, but I couldn't get my QHY5L-II to talk to my laptop either for some reason, so I had to swap back to my QHY5. I remembered that I had to use a mirror diagonal to get enough backfocus to focus it, and this made the short USB cable that plugs into the back of my ZWO camera for help with cable management too short, so I had to run back out to my other bag and grab a longer cable. Finally I got it focused, did final focus with the Bahtinov mask on the precise goto star for my chosen target, galaxy M74, and then calibrated guiding. The calibration went quickly, and soon I was off.
Perfect focus with the Bahtinov focusing mask!
At first I was frustrated at how late it must be and how much time I'd wasted, but then I glanced at the computer clock and did a double-take. I thought I might have it still set for Texas time from the Texas Star Party, but I checked my watch and it also reported that it wasn't 9 PM yet! One nice thing about wintertime astronomy...way more time! It was 8:45 when I finally started getting images in.
Above is the guide graph for the Celestron CGX-L we have in there. Keep in mind that polar alignment probably needs some work, and seeing is never particularly good around here. The CGX-L is belt-driven, so backlash is nonexistent. The total RMS error hovered around 1 arcsec, which is pretty good! I'm going to work on that polar alignment next time I'm out there though, since my 5-minute test frames showed a little elongation in the stars. For that reason, I decided to stick with 3-minute frames.
By this time, it was 22F outside, and my ZWO ASI1600MM Pro camera was happy to oblige my request that it cool itself to -40C. Now, I don't even have dark frames at that temperature, which is why I haven't used it much, but hopefully I can attempt to get some. They're hard to get because even when it's in the teens outside, the heat from my apartment leaks out on the porch and the camera has to run at full tilt to make it that cold. However, I could try putting it in the fridge or the freezer, especially now that it's drier and condensation hopefully won't form. It's supposed to be able to do -45C below ambient. My fingers were getting very cold indeed - I can use a stylus with my tablet and leave my gloves on, but my laptop's touchpad needs fingers. I was also wrapped up in a blanket sitting on the camp chair in the dome.
While the ZWO was taking luminance frames (with much better-looking stars than Tuesday night, might I add), I went back over to the Polarie to get it rolling. Frost had formed on the camera's lens already, despite the presence of a toe warmer, so I added a second one that had been in my pocket. This inevitably messed up the focus, despite the fact that I had been so careful in putting it on! So I had to re-focus. This is much easier to do on the camera screen because there's no download time, and I just look for when the stars look pixelated when I'm super zoomed in. After getting it refocused, I plugged it back into my tablet, but my tablet wouldn't recognize it was there (even though the camera told me it was plugged in). Shortly thereafter, the tablet shut off - it was too cold for the poor thing. With a dead intervalometer and dead tablet, I had to resort to using the camera's internal timer, which I can only go as long as 30s with. I knew I wasn't going to be able to get the Heart & Soul Nebulae with only 30s, so I switched to the Pleiades, which were now rather high in the sky. I saw an image on AstroBin recently with the Pleiades and the California Nebula in the same view - I hadn't realized how close they were! So I zoomed out to 55mm and got them both in the field (or at least, the star pattern nearby, since I couldn't see the actual nebula in subframes). While I was getting that target in view, the polar fine adjustment tripod head kept slipping around on where I'd attached it to the tripod - I must not have tightened the screw enough - and I lost all of my hard polar alignment work. :( So I just did it the old-fashioned way and lined up Polaris through the little hole in the side. Then I just let that roll. I was so cold and grumpy that I finally stormed inside for a while!
I checked on the memorial dome periodically and rotated the dome so that the telescope could still see through the slit. Around 11 PM, it was time for a meridian flip - the CGX-L can probably go a ways past the meridian, but the Meade scope is so long (especially with how far back I have to put the camera to focus) that the camera was already almost hitting the laptop tray attached to the pier right at when it was crossing the meridian. So I told the mount to prefer west so I could get it to flip, but unfortunately all of the precise goto stars were in places that would have made the camera hit the laptop tray. At one point it did accidentally hit it when I hit a wrong key, and since it was still trying to slew but wasn't actually moving, the mount lost alignment! So then I had to spend 5 minutes re-aligning, but luckily I found Deneb quickly this time as the first star, and then the second star was in my camera's FOV. I only did two stars in the west and none in the east because I was only going to be in the west for the rest of the night getting M74. By the time alignment was done, the precise goto stars were far enough west that the camera wasn't going to hit the laptop tray, so I got M74 back in view and carried on with the luminance frames.
Still need to clean off the dust spot.
Around 1 AM, I flipped to the blue filter, which was really hard because the metal in the filter wheel was cold and expanded. As a result, M74 ended up closer to the top of the frame (and the dust spot) than before, and I didn't catch it till later. Sigh...
At 2 AM, I decided to call it a night. I packed up the memorial dome first, then helped my minion Miqaela pack up her gear, and got my Polarie put away. Comparing Tuesday night's images to tonight's, there was definitely some kind of seeing or focus issue, so I think I'm going to have to re-take the red and green filter data, although I'll give it a shot at stacking first just to see what happens.
My gear was very frosty by the end of the night!
My gear was very frosty by the end of the night!
I was feeling pretty frosty too! 20 degrees right now feels frigid. At least the wind was dead, otherwise it would have been really miserable. I had to run the hair dryer on my camera lens a couple times, and whenever I took my glasses off to get close to the viewfinder, they froze over too and took forever to clear!
Since the Polarie wasn't perfectly polar aligned, the DSLR images drifted over the course of the night, with the Pleiades, which were originally on the right side of the frame, moving toward the left.
I don't have cold enough darks, but I do have a set of 30s, ISO-1600 darks at 24 degrees, which is close enough. I keep forgetting that I don't have a bias frame library built up yet though, so I had to go without for stacking. Bonus points though, I got 400 light frames though, which should really help my SNR (signal-to-noise ratio)! This amounts to about 3h20m!
I stacked in DeepSkyStacker and had it choose the best 90% (after I deleted images where I shined light on the lens, etc), and I honestly wasn't expecting much. I stretched it a bit in Photoshop and tried a few things, but wasn't getting anywhere. But while I was messing around in Photoshop, I kept thinking of all the things I could do in PixInsight instead...so I imported it to PixInsight and got to work. A few hours later, my jaw was nearly hitting my keyboard - I just couldn't believe it!
Date: 10 November 2018
Object: M45 Pleaides Cluster
Attempt: 9
Camera: Nikon D5300
Telescope: Nikon 55-200mm lens @ 55mm, f/4.5
Accessories: N/A
Mount: Vixen Polarie
Guide scope: N/A
Guide camera: N/A
Subframes: 359x30s (2h59m30s)
Gain/ISO: ISO-1600
Stacking program: DeepSkyStacker
Stacking method (lights): Auto-adaptive weighted average
Post-Processing program: PixInsight 1.8.5
Darks: 20 (24F)
Biases: 0
Flats: 0
Temperature: 22F
See on AstroBin
See on AstroBin
I did not think I could ever go this deep from my light-polluted location, much less with 30-second exposures, much less with an unmodified DSLR and no light pollution filter. Yes that is dark nebula you are seeing throughout the image - part of the Taurus Dark Nebula Complex. Simply amazing! Especially when you consider one of the single subframes:
Single 30s subframe
It's all about getting lots of exposures and doing good processing! I applied the following processes, in order, with guidance from the amazing Light Vortex Astronomy tutorials:
- Stacked in DeepSkyStacker (I'm going to start stacking in PixInsight, but I didn't have bias frames, and I wanted to run this one fairly quickly)
- Cropped out black edges with DynamicCrop
- Nuked background with DynamicBackgroundExtraction
- Cropped again, for vignetting that was revealed
- Applied ColorCalibration to balance colors properly (PhotometricColorCalibration wasn't working for some reason, but the older algorithm still worked great)
- Applied MultiscaleLinearTransform for noise reduction
- Stretched the image (took from linear to non-linear brightness data)
- Applied HDRMultiscaleTransform to increase contrast
- Applied CurvesTransformation to increase contrast and boost saturation a tiny bit
- Did a little more denoising in Photoshop Camera Raw
The reason that M45 is a little off-center is because I was trying to also grab the California Nebula, but my not-excellent polar alignment made the camera FOV slowly drift, which cut the nebula from the image after not too much time. I made a second stacked image using "mosaic" mode, so I'm going to run that through PixInsight as well and see if I have enough SNR there to see it. It would be super cool if it showed up!
- Stacked in DeepSkyStacker (I'm going to start stacking in PixInsight, but I didn't have bias frames, and I wanted to run this one fairly quickly)
- Cropped out black edges with DynamicCrop
- Nuked background with DynamicBackgroundExtraction
- Cropped again, for vignetting that was revealed
- Applied ColorCalibration to balance colors properly (PhotometricColorCalibration wasn't working for some reason, but the older algorithm still worked great)
- Applied MultiscaleLinearTransform for noise reduction
- Stretched the image (took from linear to non-linear brightness data)
- Applied HDRMultiscaleTransform to increase contrast
- Applied CurvesTransformation to increase contrast and boost saturation a tiny bit
- Did a little more denoising in Photoshop Camera Raw
The reason that M45 is a little off-center is because I was trying to also grab the California Nebula, but my not-excellent polar alignment made the camera FOV slowly drift, which cut the nebula from the image after not too much time. I made a second stacked image using "mosaic" mode, so I'm going to run that through PixInsight as well and see if I have enough SNR there to see it. It would be super cool if it showed up!
Since I don't have an gain=300 dark frames at -30C or any dark frames at -40C on my ZWO ASI1600MM Pro, the M74 galaxy image is going to have to wait until it gets colder at night. I'll get back to you on that!
All told (including some trouble with frozen gate locks), I got into bed at 4 AM! I'm in for a loooong cold winter of imaging, but once the CGX-L gets dialed in and we get a good alignment model in there, it will be very quick indeed to get setup night-to-night. I've got a long list of winter targets yet to image, so get ready!!
[Update December 1, 2018]
I finally got around to processing all the data I had on M74! It was tough to process, since the red and green data were a little out of focus, the blue and luminance data were drifting due to not-quite-complete polar alignment on the mount, and there were a ton of dust spots on the objective (remind me to bring my multi-coated optics cleaner next time I'm out there), not to mention the fact that I had two different temperatures on the camera, but here it is!
- Generated master superbias for both temperatures
- Generated master dark for both temperatures
- Skipped CosmeticCorrection because there's not really any hot pixels
- B and L frames calibrated with -40C dark and superbias
- R and G frames calibrated with -30C dark and superbias
- Local normalization files generated
- Decided not to use the local normalization because it duplicated dust spots and did some blurring
on the RGB channels
- Integrated each channel's subframes, using SubframeSelector to pick the better ones
- Combined RGB channels
- Cropped both RGB and L images
- DynamicBackgroundExtraction on RGB, used same points for L
- Color calibrated RGB with PhotometricColorCalibration, with background neutralization
- Clone-stamped out some enormous dust spots in the RGB
- Denoised RGB with MultiscaleLinearTransform, with mask
- Clone-stamped enormous dust spots out of L
- Denoised L with MultiscaleLinearTransform, with mask
- Denoised RGB again with MultiscaleLinearTransform, with mask
- Stretched L and RGB
- Combined LRGB
- Applied Deconvolution with star model from DynamicPSF, with range_mask - star_mask mask
- Applied MultiscaleLinearTransform for noise reduction
- Tried HDRTransformation, but didn't like it
- Applied CurvesTransformation and HistogramTransformation to boost galaxy, kill background a bit more
You know, I would hear from fellow astrophotographers about taking hours to process a set of data, b but I could never understand why - it would only take me about an hour to stack in DeepSkyStacker and do everything I knew how to do in Photoshop. But now I understand! There is so much to do in PixInsight, and even just getting the images stacked takes about half the time. I would have tried BatchPreprocessor, but I had two different sensor temperatures of data, in addition to having four different filters. You can label groups of images so that you can link a specific set of flats with a specific filter, etc, but I'm not really sure yet how to use BatchPreprocessor with multiple channels of monochrome data yet. Next time!
[Update December 1, 2018]
I finally got around to processing all the data I had on M74! It was tough to process, since the red and green data were a little out of focus, the blue and luminance data were drifting due to not-quite-complete polar alignment on the mount, and there were a ton of dust spots on the objective (remind me to bring my multi-coated optics cleaner next time I'm out there), not to mention the fact that I had two different temperatures on the camera, but here it is!
Date: R,G: 6 November 2018, B,L: 10 November 2018
Object: M74
Attempt: 1
Camera: ZWO ASI1600MM Pro
Telescope: Meade 127mm f/9 ED/apo (club's)
Accessories: Astronomik LRGB Type 2c 1.25" filters
Mount: R, G: Celestron CGEM (club's), B,L: Celestron CGX-L (club's)
Guide scope: R, G: Orion 50mm mini-guider, B,L: Celestron 102mm
Guide camera: QHY5
Subframes: L: 32x180s (1h36m)
R: 13x180s (39m)
G: 14x180s (42m)
B: 14x180s (42m)
Total: 73x180s (3h39m)
Gain/ISO: 300
Stacking program: PixInsight 1.8.5
Stacking method (lights):
Post-Processing program: PixInsight 1.8.5
Darks: 40 @ -30C, 40 @ -40C
Biases: 50 @ -30C, 50 @ -40C
Flats: 0
Temperature: R,G: -30C (sensor), 44F (ambient); B,L: -40C
Here's all the stuff I did in PixInsight:
- Generated master bias for both temperatures- Generated master superbias for both temperatures
- Generated master dark for both temperatures
- Skipped CosmeticCorrection because there's not really any hot pixels
- B and L frames calibrated with -40C dark and superbias
- R and G frames calibrated with -30C dark and superbias
- Local normalization files generated
- Decided not to use the local normalization because it duplicated dust spots and did some blurring
on the RGB channels
- Integrated each channel's subframes, using SubframeSelector to pick the better ones
- Combined RGB channels
- Cropped both RGB and L images
- DynamicBackgroundExtraction on RGB, used same points for L
- Color calibrated RGB with PhotometricColorCalibration, with background neutralization
- Clone-stamped out some enormous dust spots in the RGB
- Denoised RGB with MultiscaleLinearTransform, with mask
- Clone-stamped enormous dust spots out of L
- Denoised L with MultiscaleLinearTransform, with mask
- Denoised RGB again with MultiscaleLinearTransform, with mask
- Stretched L and RGB
- Combined LRGB
- Applied Deconvolution with star model from DynamicPSF, with range_mask - star_mask mask
- Applied MultiscaleLinearTransform for noise reduction
- Tried HDRTransformation, but didn't like it
- Applied CurvesTransformation and HistogramTransformation to boost galaxy, kill background a bit more
You know, I would hear from fellow astrophotographers about taking hours to process a set of data, b but I could never understand why - it would only take me about an hour to stack in DeepSkyStacker and do everything I knew how to do in Photoshop. But now I understand! There is so much to do in PixInsight, and even just getting the images stacked takes about half the time. I would have tried BatchPreprocessor, but I had two different sensor temperatures of data, in addition to having four different filters. You can label groups of images so that you can link a specific set of flats with a specific filter, etc, but I'm not really sure yet how to use BatchPreprocessor with multiple channels of monochrome data yet. Next time!
That's very informative.... Keep up the good work
ReplyDelete