The Veil Nebula is a stark test of your optical system. There is a huge brightness range among the various pieces. The size of the object is pretty staggering, too. And there’s a ton of wispy detail that is easy to get lost in noise, blown-out highlights, etc.

I have been playing around with the Dust-n-Scratches + gradient removal method of editing for the last couple of images. I’m not sure I have it down yet, but I’m getting there.

There is a very strange and disturbing “tearing” effect that happens with my images. It has something to do with the dark frame application I think. It does not seem to matter whether I use bias frames or not (I do not scale my darks). I end up with a kind of blotchy background which detracts a little from the gorgeous deep sky object.

This is 18x20m subframes, a full 6 hours.
There is no catalog that lists all the pieces of the Veil as a single object, so this image contains NGC6992, 6960, and several others.

For extra wispyness, check out the image inverted!

Sometimes you luck out when you’re “just testing”. I am still shaking out the kinks in my 350D workflow. I put the Ha filter back on the 350D last night. Focusing with my new commercially-produced Bahtinov mask was a snap. Those spikes were seriously sharp.

I thrashed around a little on what to shoot and &c, ended up doing some 20m subframes of IC1396. When I woke up in the morning, the sun was up and I lost 3 frames. But I ended up with 17 lights that are usable. Because it’s a new camera, I didn’t have any 20m darks yet, so I set it to shooting a bunch of 20m and 10m darks, which took all day.

I still need to shoot some flats, but I’ll get around to that.

One thing I have found is that I seriously need to clean the CCD. There is a dust bunny on there that is like 150 pixels across. You can clearly see it on the first-light NGC7000 photo, and I haven’t cleaned my optics yet.

The dew heater ran all night on a separate power supply form the rest of the rig. The scope was without dew this morning, but the finders (which had no dew prevention) were also not dewed up, so I don’t know that this is the best indicator of anything in particular.

One thing that I’ve definitely noticed is that I am having the same irritating thing happen with my FOV now that I had when I switched from film to CCD.

Right about the time I switched from film to CCD, I owned a ED80 refractor (600mm f.l.), then I bought a Meade SN6 (750mm f.l.). The SN was f/5 to the refractor’s f/7.5, but the additional f.l. meant that the FOV was ever-so-slightly smaller. Then I switched to CCD. 35mm film is 36x24mm. DSLRs use an “APS” size sensor, 23x15mm. So again, my FOV got compressed. And all of a sudden, instead of being able to just barely squeeze NGC7000 into the FOV of the film camera/ED80, there was absolutely no chance with the DSLR.

So I bided my time, waited for a nice 400mm ED refractor to come out, and finally picked up Pumpkin a couple years later. A few nights out with Pumpkin revealed that the field was badly curved, so I picked up the only field flattener available at the time, the William Optics 0.8x FR/FF, which turned the scope from a 400mm f/6 into a 320mm f/4.8. The difference in exposure length isn’t really noticeable. But the FOV was gigantic. It’s like twice the size of the FOV I got with the ED80. So I was happily ripping down these huge swathes of sky.

Then I made two critical decisions.

First, I decided that I wanted the field flat but not focal-reduced, so I picked up an Astro-Tech FF.

Second, I got an offer I couldn’t refuse on a 8mpx DSLR, so in comes the 350D.

So, now Pumpkin gets a ever-so-slightly smaller FOV with the new flattener, and the 350D (being more pixels on the same size sensor) gets an ever-so-slightly smaller FOV still. The difference seems small when considering the whole FOV: 4°x2° -> 3.2°x2.2°. But when you consider that a 1000px x 1000px crop out of the 300D/0.8x FR combo covers an area 78′ square, but the same crop out of the 350D/FF combo covers an area 55′ square, it’s big. The 350D’s square covers half the amount of sky (in area) the 300D’s square does. Whoa.

And it shows. Big objects that used to flop around in the middle of the frame with plenty of space around the edges now barely fit. And the stuff that used to barely fit, no longer fits at all. It’s a brave new world.

So this is how much of IC1396 I captured. It was spilling off the sides of the frame when I shot it last year.
I need to go grab some Veil, too, and see what’s what.

…and I may go back and buy a WO 0.8x again, as much as it pains me. sigh.

I’m still tweaking and tuning.

Here’s the image, 17x20m subframes, 5h 40m, in Ha from the observatory.

This is my first HaRGB shot. It’s a combination of grayscale luminance data that I captured while camping at Coyote Lake, CA over the weekend, and color data that I captured from the backyard on Tuesday night. Looking back, I probably should have done it the other way around. Oh well.

Also, while camping, I was using the Canon 300D DSLR, but the AC adapter for the 350D showed up Tuesday, so I switched to the 350D for the RGB data.

Because the 300D and 350D have different pixel counts (6mpx vs 8mpx) but the same size CCD (APS-size), the 350D magnifies the image slightly given the same optics. All this is to say that I had to do some scaling and rotating in Photoshop to get the two images to line up. It’s enough to make a person want a filter wheel.

The Ha data is 5 full hours, fifteen 20m subframes. That’s close to a record for my deepest exposure yet.
The RGB data is a little over 2 hours, twenty-six 5m subframes.

I’m not sure how to calculate a “total” exposure time off of that. shrug.

This is a full-frame crop of the original (as much of the full frame as was still useful after rotating)

This is my deepest image yet of NGC7000, 5 full hours (15x20m). This is with my normal Levels-and-Curves treatment in Photoshop, but it’s the full FOV. This was first light for my new Astro-Tech field flattener. It did … OK. But the field isn’t flat. I will try putting in some spacers to see if I can tighten up the stars in the corners of the image.

Normally I don’t post unedited images, but I was stunned at how nice the Red channel looked, even before I started tweaking it:

…and that’s why you go to a dark sky site, folks.

This is my first shot of NGC281. I attempted it a couple of years back, but came up essentially empty.
Normally, the Casseiopeia Milky Way is not available until autumn, but I was up really really late, so I decided to give this target a go.

I only got 4 subframes (20m each, 1h 20m total) before the sun came up.

This object deserves more attention, but I will have to wait a few months before it’s in prime shooting position.

In retrospect, I probably should have shot RGB from the dark sky site and left Ha for the backyard.
In addition, it was very windy, so I probably should have dumped Cassie in favor of Pumpkin for the night.

Having said that, I got a little over an hour of useable data (four 20m subframes) from NGC6888 tonight.

The guiding was all over the place in the wind, but the RMS was still down in the 1″ range. It’s just that Cassie shoots at 0.6″/px.

Another long clear night.

Since the webcam was already installed and focused, I started the evening working WCS to refine my polar alignment. Azimuth was essentially perfect. Elevation started around 200 and is now around 40. I’m done with polar alignment. The final elevation run went for a full 20 minutes, with up to 3″ of drift.

Then I had to reinstall the motor focuser and the main camera. That went fairly quickly. But I needed to do a FocusMax run, too, which is always a bit hit-or-miss. I found a spot close to focus by watching the numbers in FM. FocusMax was headed off to infinity, but it stumbled across a nice 4.78 HFD position, and I went with that.

I spent a little time trying to figure out what to shoot for the evening, and against my better judgement, I landed on UGC10822, The Draco Dwarf Galaxy. I shot 2 hours of it, 20m subframes. In order to get 20m frames, I have to shoot through the Ha filter, so this target is even less visible than it normally would be.

I have never seen the point in shooting open clusters; the photo usually looks a whole lot like empty space. But I have now found the idiot cousin of open clusters — dwarf galaxies. I swore I’d missed the object, until I looked up a photo of it online, and I then noticed the little teeny dim stars that make up the galaxy. So I “captured” it, but I am going with my better judgement and not posting the image.

By the time I’d finished on UGC10822, Cygnus had risen high enough for me to start shooting some nebulae. So I set up on NGC6820, which I’d only shot once before, and went to bed.

I woke up in the morning to a strange phenomenon; I had several subframes with trailed stars in them! Oh, the horror! I actually stacked the trailed ones into the first cut; median filtering did an OK job cleaning up the worst of it. But for the cut you see here, I only took the untrailed images, and ended up with 2 hours of data.

This is NGC6820 in Cygnus, 6x20m subframes, in Ha

Click here for a full frame version of this image.

I talked it over with Art, and he assures me that the seeing goes crazy every morning about 2 hours before sunrise, so I am not freaking out about the trailed subframes at this point. We’ll just see if this continues…

…with zero dropped subframes.

I spent a little time refining the polar alignment last night. I am still getting some drift while running unguided; I calculated it to be about 6′ from the pole now.

The guiding is even smoother now, if one could believe it possible.

I shot 20m subframes all night. I ended up with 5 frames of M51 (OK, but not worth posting). By the time M51 was finished (around 2am), Cygnus was up, and so I went and grabbed 10(!) more frames of Sh2-101.

The 3h 20m image came out pretty nice, but as always it could use some more integration time:

This is one of my standard 1500×1000 -> 900×600 crop-and-size jobs. There was a lot more context in the image, but Veronica is badly out of collimation at the moment, so there were some badly mangled stars; this is as good as I could get.

Things I noted last night:
- the pier topper needs to be rotated about 1/2″ clockwise, because the azimuth adjustment has run out of room to the west.
- Veronica needs to be recollimated. I am going to pull the long bolts and put the stock ones back in. I think I’m done with the RCC/OAG.
- I am not certain if it did anything bad, but I pulled the webcam out of the OAG and put it into the main focuser partway through the night. I would think that it doesn’t matter where in the FOV the star lies, but better safe than sorry.
- I will not be putting the mount on the tripod for tonight’s star party. I think it’s more impressive on the pier.
- One of the GOTOs last night was so accurate that I looked in the finder, couldn’t find the star, started pushing it around with the hand controller, and realized it was sitting behind the crosshairs. (:
- The mount put up more gaudy numbers; 0.4″ RMS for the night. wow.

A good night. And I got an hour more sleep and still hit a Cygnus target!

Buoyed with the success of the NJP with Pumpkin, I decided to start upping the ante. Pumpkin was mounted offset to one side, but I decided that Veronica should be balanced over the Dec axis. I was already a little concerned whether the DA’s would hold Veronica, and I wanted one less thing to worry about regarding balance. So I reworked the DA’s so they’d fit in the center of the mounting plate, and mounted up Veronica.

At the same time, I decided to mount a couple of finder dovetails on the mounting plate, so I would have a place to mount Guido. That was a measured success; it totally worked for guiding (see guide data below), but was not so cool for using Veronica’s finder; something I’ll fix presently.

Here’s the guide data. The mount is so smooth, it’s silly.

I had been having trouble last night with subframe length; skyglow was eating any images longer than 3 minutes. I wanted to test some 10m and 20m subframes, so I decided to add the Ha filter to the mix. This made focusing a challenge, although the 3-day moon did help out some by staying up late enough that I could use it as a focus target.

Getting FocusMax working again was a bit of a pain. It took about 2 hours (or 3?) before I was up and running and ready to start guiding.

The winter nebulae had set by then (it was about 11:30 or so), and the spring ones weren’t due up for hours, so I decided to run the Ha filter on the spring galaxies. M101 was the target of choice; because the polar alignment isn’t exactly perfect, my “bracketing” exposures missed the galaxy (because the GOTO was inaccurate; I need to add “re-synch on nearby star” to my workflow habit). Once I resynched, Temma put M101 on the CCD.

The bracketing exposures showed that 10m exposures were fine, but 20m are a little too long (I’ll work out later whether 15m can be done).

I had earlier figured out that I can go about 2 hours, 10 min. past Meridian before hitting the pier. I set this up in the mount safety limits, but I still need to make sure that’s actually doing something.

Because I was 30 min. before the Meridian, I set it up to run for 2.5 hours (20x10m exposures, allowing for download time), set an alarm clock, and took a nap. At 3:30a (yes, it took awhile to get started), I went out, to find that the telescope had hit the pier. It actually hit the pier during the last exposure before I got up, though, so not too bad.

I crunched the data once I woke up, and here’s what I got. I think that my data processing workflow could use some help; I stacked this using “SD Mask”, for reasons passing understanding, and got some strange diagonal noise added.

But, this is 16x10m images, 2h 40m total, of M101 in Ha through Veronica.

I re-synched the mount on a star rising in the East (Deneb), and slewed to NGC6888 to see how Veronica could do on a “real” Ha object. I set it up to run for 3 hours (18x10m), knowing that I would be up by 6:30a to turn everything off.

I went to bed, had a crazy bad dream about the mount running into the roof and somehow falling into a lake (wow), and woke up to the mount having lost the guide star because of the sunrise. I only lost one image there, too. Not bad again!

I was a little off on the GOTO again; it was 3:30 in the morning, give me a break. I only got 6 useable subframes of NGC6888; I think a cloudbank hit at some point in there.

NGC 6888, The Crescent Nebula in Cygnus, 6x10m, one hour total exposure, Ha

A reminder that the reason these are grayscale is that this is the red channel only, as the Ha filter only lets data into the red channel.

So, my impressions?
First, the NJP is absolutely destroying my circadian rhythm. I am so tired I could sleep standing up right now.

Second, the mount is so stupidly accurate that I don’t know what to do with myself. The first time I tried to cut the guiding data this morning, PemPro reported a PE curve of +/- 0.0. When I look at the raw data, other than the big jumps that happen in Dec whenever the shutter trips, all of the data falls between 0.0 and 0.4 pixels. From the previous night’s data, I had an RMS periodic error of about +0.3/-0.2, or a swing of about 0.5″. Tonight’s data looks similar. That’s a 20x improvement over the best performance I ever achieved with the G-11 (+/- 5″, 10″ peak-to-peak). NJP good? I’m sold.

Third, I think the RCC is a lost cause. It makes my telescope hard to collimate, and it doesn’t seem to do as nice of a job flattening the field as the MPCC does (take a look at the corners of the images). Now that I am using clip-in filters, and have a mount that is really guideable, I think that I’m ready to pull the RCC, get Veronica reset back to normal, and use the MPCC from here on out.

Fourth, I don’t know whether the IDAS LPS and Ha filter are not behaving properly together, or whether my optics need cleaning, or dew prevention, or image processing, or what, but these images seem a lot noisier and generally uglier than the Ha stuff I was doing last spring. I’ll get my Ha mojo back, just something to note.

Finally, I can’t wait to see what the mount can do once I’ve aligned it properly! Unguided? Crazy long focal lengths? Oh, I think we’re going to find out. Stay tuned.

With the newly-tuned mount, I decided to try Pumpkin on some more unguided imaging. After the “baseline” from the previous night, I now wanted to see what this MFWB could really do.

M42 was a bit far west and covered with clouds by the time I got Pumpkin up and running (I had a pretty funny motor stall that wigged me out momentarily — in my haste to get imaging after testing Trixie, I had forgotten to swap out the counterweights, so 10# Pumpkin was offsetting 43# of counterweight… oops), so I decided to see if I could work The Rosette for a bit instead.

This time, I centered the cluster that I could see in the finder. This proved to be a good decision:

That’s 12 five min subframes, one hour of integration time. Rosette, you are mine.

I picked up some wispy outer nebulosity, too, that is easier to see in inverse:

Still a bit noisy. I’d like to apply a full calibration set on this, and have about 3 hours of integration time to work with.

This was shot under an 8-day moon.

The image is unguided.