rickwayne

Your estimate for autofocuser parts is in the ballpark. I started with a Pi running Ekos, added a Waveshare Motor HAT board to it (US$26), got a stepper motor ($30), pulley ($10), and a timing belt ($8). The big advantage is refocusing as the scope cools down overnight. 1℃ seems enough for my 336 f/5. I never knew when to stop tweaking the Bahtinov mask, though a quantitative figure of merit would have really helped.

Wow, I'm using a 183MM and I think I have it set to 2 pixels or something, which was the default in StellarMate. Really nice detail in M82. Last time I tried it I could barely make out that there was supposed to be hydrogen there!

Oh, there it is. Nice detail in the Whale!

The other issue with temperatures dropping during the night is the thermal contraction of bits in your equipment, especially any metal ones. Many of us refocus our telescopes every hour, or even oftener, or use software to monitor a sharpness figure of merit to trigger refocusing. For my 336 mm f/5 refractor, the zone of critical focus is all of 74 microns deep. Doesn't take a lot of contraction to exceed that.

Nothing obvious springs to mind. You could overexpose and look for stars in the image, see if they're soft or jump-around lines, might help you see if it's vibration or something else. You could also try using much higher ISOs to enable very short exposure times. Another dodge would be to try it with a terrestrial object during the day, when you can use super-fast shutter speeds even at lower ISOs, and really nail the focus. Finally, you could make or purchase a Bahtinov mask and use it to get stars absolutely spot-on focused, and work from there.

If you stay within its weight limits and don't overstress the geartrain, I've found an iOptron CEM25P (now CEM26) a very good starter mount, though it's certainly at the top of your range. I loaded it up with almost 8 kg and was consistently getting sub-second autoguiding with a lot of learning and tweaking. Eventually you will probably want a long scope or heavier load and want to upgrade, but you can get great results out of one. I love iPolar (I have since upgraded to a CEM70), but the polar scope plus something like Sharpcap will get your alignment dialed in very nicely.

In general, you want to do as much processing as possible while your data are linear, i.e. unstretched. So do everything you can in SiriL, then stretch and export as the last two steps. It sounds as if you're doing it pretty right (as your image demonstrates -- nice one!). One of the reasons I sprang for Astro Pixel Processor is its world-class gradient reduction feature. That's something you really, really, really do want to accomplish with linear data if at all possible. It might take a couple of goes, mind. Eliminating noise digitally without losing detail or introducing artifacts (e.g. the "plastic" look) is a real challenge and no mistake. I have struggled to find free software that really does the job; again, I resorted to the wallet, and picked up Topaz DeNoise AI. It's not magic, but it's pretty good. The noise-reduction tech in PixInsight is supposedly quite good, but that's getting practically into nosebleed territory, both financial outlay and the effort required to master it. To reiterate: If you're doing this well early in your imaging career, you should be proud and delighted.

If you think you might someday put together an autofocuser, the Waveshare Motor HAT board serves as a 12V power converter. It also has two stepper-motor outputs, and there's an INDI driver for autofocusing. 5.5x1.2mm barrel connector, pretty standard in the astro world. My autofocus setup is well under US$100; after the HAT, I needed a stepper motor, a pulley, a toothed drive belt, and a bracket to hold the motor, which I built out of some scrap steel.

Thanks for putting up the data. Would you mind also posting a screenshot or JPG giving us a visual clue to the problem?

Glows might also be amp glow -- what does a stretched dark look like for your camera?

It does have that "plastic" look common when you lean too hard on NR. It would not take much focus or zoom creep to soften things either -- you can check by loading subs from the beginning and end as layers in Photoshop, auto-aligning them, and playing with the opacity of the top one to blink-compare them. Two things about using a Bahtinov mask: You have to be super-exacting if you're shooting at low f-ratios, and even if you nail the focus to start, thermal effects can throw it off as the night progresses. For my 336mm f/5.5 rig, the critical focus zone is all of 74 microns deep and 1 degree C is enough to bollix the focus. I personally don't mind a bit of noise in an image showing good detail, though I do tend to pull that slider a little too enthusiastically. Of course, I was a late-70s photojournalism major so I perforce learned to tolerate a lot of grain. Obviously a nice job with polar alignment, BTW. Pixel-peeping reveals a bit of elongation in the RA axis but I don't see any in DEC (RA is vertical to the Horse and DEC along its base), so not a lot of drift going on there.

That is...completely whack. Why only those seven spots? Do the individual subs have stars there? Can you blink to see if any of them have them in a different position? (Still can't imagine why only those particular places would be affected, though.)

I'd be shocked if you can't zoom the live view. Two other dodges to maximize the signal you have to work with: Rack the ISO to something completely ridiculous, and use a long exposure now and then instead of live view to confirm where the spikes actually lie. Also, it can help to use a dimmer star for final tweaking, paradoxically enough. The spikes and diffraction "beads" get fatter with brighter stars and so give a less precise indication. Bias frames are short-exposure darks, and so needn't be shot in the same session. Dark current for such a short exposure should be negligible, so the temperature really needn't be close (although it's a good practice to use the same ISO). Since the sensor temperature varies during an imaging session, close enough is probably good enough for darks -- once you have a set at a given temp (use the EXIF data to find the actual sensor temp, instead of ambient), you can safely reuse them for awhile. Focus position doesn't matter for darks and bias, there's no light anyway. Flats, however, should be shot every session, at the same focus point. A cheapo LED tracing pad has a sufficiently consistent brightness across its width, I find, and can be used in the dark after you've achieved good focus, unlike twilight flats. And IMO it's less bother than a T-shirt, just point the scope straight up and lay the panel across it.

I assume that M51 is cropped a lot, given the field of view of the others, which is why it's so noisy. I'd say M81-82 is your most successful processing, the black point is maybe a little high. If you leave a smidge of light in the background (with 8 bits per color channel, something like 30R, 30G, 30B) you'll be able to retain more detail in faint nebulosity. If these are your first efforts, I'm pretty jealous! Late-model DSLR sensors often don't exhibit much dark current or amp glow, the two problems mandating dark frames. My Pentax, for example, seems to do fine without. If you do decide you need them, since dark current depends on temperature, and your rig is most likely cooling down during the entire imaging session, you might be able to either get away without darks or do them first thing. Most likely the temperature is changing more slowly at the end, of course. Folks with observatories and automation often set up a lighted panel of some sort on the wall, then have the scope slew to point to it at the end of the night and take their darks. Conceivably you could do something similar with a black panel mounted so that the end of the dew shield just grazes it, maybe a 3-sided box to minimize ambient light? I dunno, might just bend the dew shield or dink your mount if it weren't exactly right. Another strategy would be to take sets of darks at different temperatures (cooling the camera is left as an exercise for your ingenuity), then look at the EXIF data for your light frames afterward to determine which set of darks to choose. It would be tedious, but you could put the bagged room-temp camera in a freezer, let it shoot for awhile, take it out, warm it back up, repeat for a few cycles, then collate the resulting frames by EXIF sensor temperature into sets. You don't need a bazillion darks to yield a useful master dark for a given temperature, so a few cycles would probably give you darks for several temperatures. Note that I'm specifying EXIF data rather than, say, putting a thermometer in the bag, because what counts is not the ambient temp but the sensor temp. And that actually varies even if the ambient does not, since the electronics heat up in use and cool down when they sit. I suspect you'll find that flats give you a much bigger bang for the buck than darks, frankly. And you can do those before you start the sequence.

Oh, two other tidbits: For an unmodded terrestrial camera and a dim emission nebula, take LOTS of sub-exposures. You'll want hours of total integration time for the best results, which for 20-second subs means hundreds of frames. Yes, you're right, that will be a complete PITA to process, which is one big reason why the true obsessives spend kilobucks on mounts that allow 10- or 20-minute exposures. Also, you can use plate solving to really nail down where you're pointing. With Internet access, upload a JPG or FITS to nova.astrometry.net, within a few minutes you will have an exact solution. You can also download ASTAP, or PlateSolve2, or the offline version of astrometry.net, plus the data files, to do the same thing local to your computer.

Especially for dim emission nebulae, and looking through the built-in IR cut filter on your Nikon, the image preview gives you little useful data for judging exposure. Far better to go by the histogram, and the good news is that the rules aren't too subtle. Dim as it is, the nebula is going to be brighter than the background, which makes up most of the pixels in your image. So if the big peak is completely clear of the left edge, you won't be black-clipping anything. With stacking, that is actually enough right there. You can check for overexposure too, the histogram's tail shouldn't actually touch the right side either. If you can't satisfy both, that means the scene has more dynamic range than your camera can accommodate, so you have to either allow some black clipping, blow the brightest elements (stars) out to white, or resort to HDR techniques of shooting and combining exposures optimized for each end of the range. The Orion Nebula is notorious for this -- lots of very dim gas, but the Trapezium area is super-bright.

Don't want to sound like a broken record but to me, "brutal gradients" == "buy Astro Pixel Processor". Just sayin'.

Lovely, lovely detail in the galaxy.

And, as Charlie Bracken points out in the APP quick start, it helps to use it iteratively -- don't try to get it perfect in one go.

I spent the money for APP, but I'm pretty impressed with ASTAP. Lots of little things to like, e.g. the "blink" feature.

Well, expert enough to be correct with your advice, at least! If you want to get all quantitative about it, here is a calculator for image scale: astronomy.tools CCD calculator. As for field of view, Telescopius.com will help you previsualize what various setups will do, as will the Stellarium application if you download it.

To be clear, there are two different ways to set up guiding, one of which has a direct connection from the guide camera to the mount using an RJ-11 connector (like a phone). You don't want that. 🙂 Assuming that the 294 has a USB hub built into it, you can plug the guide cam's USB into that. Or you can plug it directly into the computer. Same same, other than the latter means two cables dangling off the scope instead of one, as noted above. Since cables can drag and snag -- it takes astonishingly little force to mess things up -- fewer cables are better. Can't help you with the software setup -- I use a single integrated package that does mount control, guiding, and sequencing.

Ekos definitely does #2. I use that all the time. "Load and Slew" button on the Alignment module.

Twenty seconds is a LONG time for Ekos to download an image. The Capture module should list the download times for you. That's something you can play around with in daylight, no need to waste dark-sky time. Especially with very short exposures, your sub-exposures will indeed look all dark. Check the histogram, though. So long as you're not up against the left edge, how each subexposure looks doesn't matter a bit. (If there are data right at the left edge, that strongly implies that there were dimmer pixels in the scene too.) Ideally the biggest peak -- the background or "skyfog" value, since dark sky is the commonest thing in astrophotos -- would be at about 1/3 of the range. But you haven't the margin for that kind of nicety.

All excellent suggestions. I would add that for best results, you should calibrate near the meridian and celestial equator, then run the Guiding Assistant on your target for a few minutes and see what it recommends.