A Year's Experience with the ZWO ASI1600MM-Cool and a Year's Progress

I bought my ZWO ASI1600MM-Cool a year ago, and after using it for a year I still love this camera. It produces far better images than I could ever get with my old Meade DSI IIc. Some of this performance is no doubt because of the smaller pixels and the fact that it's a monochrome camera so doesn't suffer from the limitations imposed by a Bayer matrix. It also true that this better performance comes at a cost; not just the higher cost of the camera, but also in much greater effort required for processing the data (at least in my hands). The much higher number of pixels alone means that processing time is much longer (I'd estimate a factor of 3-5 times at least) and data storage requirements are much greater (more than a factor of 30--my primary drive is filling up fast!). In addition, it has been a real adventure (and actually a lot of fun) learning how to do LRGB, SHO, HSO, and various other combinations to produce color images. I've still got a lot to learn about that!

I've now used this camera with my AT8IN 8-inch imaging Newtonian, with my Orion ST-80 very-non-apochromatic refractor, and with an old Tamron 135 mm camera lens bought for $15 on e-Bay, and it has done a good job with each. The much larger imaging area does mean that the optical flaws in my systems are much more apparent than they were with the DSI IIc; its tiny sensor only sampled the nearly flawless center of the image plane, whereas the ASI1600MM-Cool finds lots of flaws out near the edges of the field of view and I see plenty of effects from vignetting that I never saw before. I've got one dust mote on the sensor that I either have to correct with flats or correct in software, but I've pretty much figured out how to do that.

I've experimented with various gain/offset settings, and find myself gravitating back toward the "default" unity gain (-139) offset 21. I really don't have a scientific reason for this. It's just what I've found most comfortable with the exposure times and dynamic range expectations I have. I think higher gain combined with shorter exposures is an approach with a lot of promise, as it seems like it would allow better statistical treatment and would minimize any frame-to-frame drift from differential flexure (I see a lot when using my finder/guider with the Tamron lens, and even see some when using an off-axis guider). A future goal is to get a bit more scientific with these experiments.

The only persistent problem I've had with this camera appears to be in the USB-3 cable connecting the camera to a hub located at the telescope mount. Occasionally the connection to the camera and filter wheel drops, with the result that the next attempt at downloading an image hangs. I've had quite a few sessions end prematurely because of this. Often it happens while doing a meridian flip or right after a filter change. I'm sure it's not the hub or the USB cable back to my living room, because the guide camera is also on that same hub and never has problems. When I originally set up the ASI camera I noticed it only seemed to work well with the cable that shipped with it, so it seems to be a bit picky about the cable. I might try getting another cable to see if it makes a difference. Usually, jiggling the cable and doing an unplug/replug fixes the problem.

All of that said, I've learned a lot over the last year. As evidence, I offer the following two images. The first was done a year ago, and was the first time I had used the ASI1600MM-Cool with the Tamron 135 mm lens. Because it was a "first time" in many ways, maybe this isn't exactly a fair comparison. At any rate, I was delighted with the image at the time, because there was no way I could have done anything like that with my prior setup.

This image was a first in many ways. It was a first attempt at this large, dim, difficult target. It was the first use of the Tamron 135 mm f/2.8 lens I bought for $15 on e-Bay. It was my first try at using high gain with the ASI1600MM-Cool, and it was first light for the Tamron 135+ASI1600MM-Cool combination. I didn’t get any blue data (the neighbor’s darn tree got in the way again), so I cheated: I used the Hα data both for luminance and for R, and I put the R data in the B channel. All things considered, it could have turned out worse. I had to process the heck out of this to get a result that was anywhere near reasonable. The supernova is reported to have gone off 40,000 years ago!

Date: 7 Jan 2017

Subject: Sh2-240 (Simeis 147) Spaghetti Nebula supernova remnant

Scope: Tamron 135 mm f/2.8 lens

Filters: ZWO 31 mm RG, ZWO 7nm Hα

Mount: EQ-6 (EQMOD)+PEC

Guiding: Finder/Guider+DSI IIc+PHD 2.6.2.4 (Win 10 ASCOM)

Camera: ASI1600MM-Cool, -20 °C, acquired Hα 1x1 RG 2x2, Gain 300 Offset 50 

Acquisition: Sequence Generator Pro 2.6.0.1

Exposure: 80x180 s Hα, 55x45 s R, G

Stacking: Neb 4.1.2, flats & darks, trans+rot align, Nebulosity 1.5σ stack.

Processing: too extensive to detail.

The second image is my most recent, just finished today. In the last year I've learned a better way to mount the camera on the lens, and I've built, tested, and used an automated stepper-motor-based focuser (based on Robert Brown's designs and using his software) for the lens. I've learned to do bicolor combinations and have benefitted greatly from Annie's Astro Actions for Photoshop, which make this easy. I've gotten better at picking good stars for doing alignment and stacking in Nebulosity (I tend to use dimmer stars that have good shapes, and I always magnify the image when I choose them, avoiding ambiguity in which star is picked). I'm trying to be a bit more subtle in my processing, so my images probably are not as stark from trying so hard to pick up the really dim stuff. To me this newest image is far better than the one from a year ago. One of the things I really love about astrophotography is that it's easy to see progress!

One motivation for this image was to compare with the image I took a year ago, which was the first using the Tamron 135 mm lens; I wanted to see how much my technique and experience had changed over the last year. The new image suggests these have changed a lot. I had intended this to be HSO/SHO, but the session failed while trying to focus the SII filter (which was the last in the sequence) long after I had gone to bed, so it ended up being only bicolor. It looks like the weather is turning stormy at last (winter is coming pretty late this year in Utah) so I’m going with the data I have rather than wait for more clear skies. These data were tough to process because this target is quite dim, hence the original data required very extensive stretching and were consequently pretty noisy. Nearly all the structure shows up only in Hα (only the barest hint of the filaments in OIII), though I’m curious whether anything would have been visible in SII. I continue to have trouble with my Tamron 135 flats (the current ones overcorrect), so I ended up not using them and just let StarTools’ Wipe vignetting module handle things. I went back and forth a few times about how dark to make the background, and ended up leaving it fairly dark. I understand this nebula is thought to be a supernova remnant. The structure in the filaments is amazing—it looks to me more like a wad of fine spun gold with the color choices in this image, rather than spaghetti—leading me to wonder what caused it. Most likely it was just random turbulence at the time of the supernova explosion, but it is fun to speculate that maybe it was somewhat shaped by a planetary system around the star, destroyed at the time but imprinted on the remnants (now that we know that most stars do have orbiting planets). Of course I have no evidence for such romantic speculation!

Date: 19 Dec  2017

Subject: Sh2-240 (Simeis 147), Spaghetti Nebula

Scope: Tamron 135 mm f/2.8 lens stopped to f/4

Filters: ZWO 31 mm diameter unmounted Hα, OIII (7 nm bandpass)

Mount: EQ-6 (EQMOD 2.000j)+PEC

Guiding: Orion 9x50 Finder/Guider + DSI IIc +PHD 2.6.4.5 (Win 10 ASCOM) using predictive PEC algorithm

Camera: ASI1600MM-Cool, -20 °C, Gain 139 Offset 21 

Acquisition: Sequence Generator Pro 3.0.0.4

Exposure: 40x300 Hα, 36x300 OIII

Stacking: Neb 4.1.6, darks only, trans+rot align, Nebulosity 1.5σ stack and align.

Processing: StarTools 1.4.328: Used StarTools’ “Wipe” module to correct for vignetting, then stretched and deconvoluted each channel separately in StarTools. Aligned the processed layers in Nebulosity then combined in Photoshop using Annie’s Astro Actions’ HO Bicolor module. Multiple rounds of Photoshop Curves and Levels, combined with several rounds of Carboni’s Make Stars Smaller, then Deep Space and Space Noise Reduction and a few rounds of Less Crunchy More Fuzzy. AstroFrame.

2 Minutes, 17 Seconds

11:09 AM

Ever since reading about them as an elementary school student, I have longed to see a total solar eclipse. It has been a “bucket list” item. At long last, I have finally checked that one off. On Sun. 20 Aug I traveled with my wife Denise, my son Brent and my son-in-law Eric Mansfield to Rexburg, Idaho to view the 21 Aug 2017 “Great American Solar Eclipse” in totality. We had originally planned to go in the afternoon, after church, but I guess I got over-anxious about forecasts of huge crowds and traffic problems. The last thing I wanted was to be stuck in a traffic jam on I-15 somewhere south of the zone of totality, so we left around 7:30 AM and drove up to Idaho Falls, just a few miles south of the zone, where we went to church. We then drove on up into Rexburg, more or less to the center of the zone. We saw no heavy traffic at all. We had a reservation to camp in Riverside Park in Rexburg, where they had taken a large grassy area and partitioned it into hundreds of tent sites across a little river from a parking lot, which was entered by permit. I had bought this from a guy online who couldn’t go, so I couldn’t help but worry a little that he might have sold it more than once, but those worries were completely unfounded; all was perfect. 

11:14 AM

There were probably a thousand people camped in the park. In the parking lot they had food trucks and lots of booths selling various memorabilia; the sister missionaries from the Mormon Church (mine) were there with a family history booth. I’d say most of the people who were there, who were mostly families, were from Utah, but visitors were clearly there from all over the world. I bought a T-shirt. A few people had telescopes and were practicing projecting images on tent sides, and one guy had what looked like a 12” Dobsonian. I anxiously watched my weather apps (Weather Channel, ICSC, & Observer Pro), worrying we might get clouded out right at totality, and Sunday afternoon it was looking like we were going to have some clouds at the crucial moment and were certainly going to have some overnight. Sure enough, around sunset it clouded up pretty well, so the Dobsonian guy didn’t get to use his scope.

My bio-alarm (a full bladder) woke me around 6:15 Monday morning to a perfectly clear sky, which remained through the eclipse and beyond, so that worry was also unfounded. We packed up our stuff and drove over to the home of my college friend, Steve Ott, who is on the faculty of BYU-Idaho. Again, there were worries that the whole town of Rexburg would gridlock with visitors, but that did not happen. I would say the city was extremely well prepared for the eclipse and probably got fewer visitors than expected. The predicted gridlock in Rexburg didn’t happen. Cell phone service was not interrupted, although the traffic cameras did become inaccessible. So all in all, the infrastructure held up pretty well. At the Ott’s I stayed busy trying to make a solar filter for my iPhone/2x telephoto attachment (didn’t work well at all; too many nasty internal reflections and no ability to set the phone camera’s exposure times short enough) and setting up my point-and-shoot to try and image totality. 

11:22 AM

The eclipse started about 10:15 AM, with a little nick appearing in the side of the sun. We had lots of pairs of eclipse glasses, and I had a pair of solar binoculars from Celestron. Eric had made filters for his binoculars using some Baader astro solar film I had left over from making filters for my scopes for the transit of Venus. We also hastily assembled a pinhole camera and Steve had a small (maybe 50 mm) old telescope with a somewhat dangerous “solar” filter that screws onto the eyepiece. The eclipse glasses gave a nice sharp view and were easy to use, showing the Sun in an orange color, but the binoculars gave a much closer if whiter, less colorful, view. There was a nice line of sunspots across roughly the center of the Sun that allowed us to mark the progress of the Moon across the Sun. These were easily visible through the binoculars. I liked the simplicity and sharp view afforded by the eclipse glasses. One thing I noticed right away that was different about this total eclipse vs. the partial eclipses I’ve seen before was the symmetry: the Moon’s motion across the Sun was centered right on a diameter. I was a little disappointed with the pinhole camera because its image was small and not particularly bright. What did not disappoint, and one of the highlights of the experience, was looking at light coming through the trees in Steve’s front yard. I got lots of pictures of this, and it was easy to follow the course of the eclipse by looking at how deep the crescents were.

11:23 AM

I had read that you should look for diffraction fringes (“shadow bands”) moving across the ground just before and after totality, so we did and we saw them; a conveniently located white car, belonging to one of Steve’s family members, provided the perfect screen to view the shadow bands. These were spaced maybe an inch apart, they moved rapidly, and were very faint. One of our group said it looked like mirage ripples or like seeing light through rippling water and I think that’s a good comparison, although it was dimmer than what you see through water. 

I also wondered if we would be able to see the Moon’s shadow approaching. Just before totality I looked out to the west-northwest, where we had previously determined the shadow would come, and sure enough I saw a shaft of darkness coming down from space. It was impressive. 

11:25 AM

Through most of the eclipse you couldn’t tell any difference in the brightness of the sunlight, but for perhaps the last 5 minutes prior to and after totality the light was obviously dimmer and the temperature was also noticeably cooler. Some of our group saw the street lights coming on, but I didn’t because my attention was elsewhere.

11:37 AM

Then came totality (11:33 AM), with cheering from all over town and lots from me. It was about as dark as an hour or so after sunset, dark enough that the brighter stars and Venus were easily visible. And the corona, which was a beautiful silvery white and full of easily-seen filaments and asymmetric, was naked-eye visible out to 4-5 solar diameters. I fumbled to get this in the field of view of my point-and-shoot camera and hit the exposure button with the intention of taking continuous 5-sec exposures at maximum zoom through the 2 min 17 sec period of totality. I hit the start button without starting the program, so I got nothing photographically from that effort. In retrospect, 5-sec exposures would have been far too long anyway. I did have my 70 mm binoculars ready and prefocused, so I got them on the totally eclipsed sun and was rewarded with a beautiful and somewhat unexpected view of several large prominences along the top and trailing side of the eclipsed sun. These were bright orange to contrast with the silvery-white corona. Steve Ott had a phone app running that counted down to the Sun's emergence from behind the Moon and also mentioned out loud what to look for; this was very helpful. This allowed me to get the binoculars down in time. We got an awesome “diamond ring”/Bailey’s beads effect as totality ended. Pictures do not do this justice as they can’t capture the huge dynamic range between the darkness of the Moon’s disk and the actinic brightness of the Sun coming through Lunar topography. The time of totality was unearthly, beautiful, and over way too quick.

It got light and warmed up, and we saw the diffraction ripples again. Everything happened in reverse from there. I got some pictures of very narrow shadow crescents. I hung on to the bitter end, trying to see 4th contact. I lost the Moon from the Sun’s edge about 20 seconds before the app said the eclipse ended.

We decided to drive down to Ririe, Idaho, where Denise’s dad grew up, and also visited the Ririe-Shelton cemetery where her grandparents are buried. All the entrances were locked up to prevent people from driving in, but we were able to walk in. We saw lots of this: public parking lots and the like were blocked all over the zone of totality, presumably to force people to rent spots from the farmers who were gouging outrageous prices. 

11:42 AM

We then started heading south along back roads, which were all jammed. We (and probably everyone else) were using Google Maps to try and avoid the mess, and just kept getting into the mess. This was the worst traffic jam I have ever experienced in my life. Idaho Falls was a congested chaos, with many freeway on ramps closed. We eventually got on I-15 and moved at about 5 miles per hour average. Things finally lightened up a little south of Pocatello, but then jammed up again north of Tremonton, Utah and stayed heavy until Ogden. We didn’t really get to full speed until the Salt Lake Valley. While it took about 4 hours to drive up to Rexburg, we needed more than 9 to come home. The Eclipse Expedition was an adventure, and was completely worth it. Now I want to do it again!

As I continue to reflect on the experience, this was among the most impressive public sights I have ever seen, up there with watching men walk on the Moon (on TV, of course) and witnessing a space shuttle launch from close enough that I not only saw it, but heard and felt it. 

12:21 PM

And I can’t help mentioning a spiritual insight that came from watching the eclipse, which after all was a very spiritual experience for me. We read in Mormon scripture about how a mortal human cannot behold God with the natural eyes. It occurred to me that the Sun is a good analogy to this. We can’t look at it unaided for more than an instant without severely damaging our eyes (you’d better have those eclipse glasses or something equally good!). Because of this, we go through most of life not comprehending how truly grand our star is. Seeing the eclipse gave me a whole new perspective on the Sun; it’s not just a brightly glowing disk, but its corona and magnetic field extend way further into space in ways we can’t normally see and of which we are not normally aware. Of course I knew all this and more before, but seeing it gives it far deeper impact and meaning. In the same way, we mortal humans cannot bear to look directly at God. A finite mortal mind cannot perceive or indeed even conceive of a superintelligent God, yet at the same time, just as we can understand that our Sun nurtures all life on Earth, we can understand that we are all God’s children and that he loves us enough that he sent his Son in loving, atoning sacrifice for all of us. I’m grateful to have this knowledge.

1:59 PM (post eclipse)

Automated Focusing of the Tamron 135 mm f/2.8 Lens

My second-iteration MyFocuser Pro control box

I thought I’d start writing up my experience building a motorized focuser for the Tamron 135 mm f/2.8 lens. After using the Tamron 135 mm f/2.8 lens a time or two, it has quickly become apparent that the various filters I’m using are not parfocal, which maybe should have been obvious given that the lens uses refractive optics and refraction is wavelength-dependent. Therefore, if I want to get good images with this lens I’m going to need to refocus at least every time I make a filter change. That would badly interfere with sleep. I recalled that Robert Brown’s do-it-yourself MyFocuser Pro project includes description of a setup for using  his Arduino Nano-controlled stepper motor focuser on a camera lens using a belt-pulley system to move the lens. After a little investigation, I decided to build a second MyFocuser Pro system to implement motorized focusing with the Tamron 135 mm f/2.8 lens. Experience with my original MyFocuser Pro has shown me I don’t use either the LCD display or the manual buttons, and a little calculation showed that with 1/32 stepping I don’t need the more expensive planetary gear stepped-down stepper motor. This meant I could build the minimal MyFocuser Pro and use it on the inexpensive NEMA 17 stepper motor and expect reasonable performance. 

I ordered the parts, waited for the slow boat across the Pacific to arrive, then assembled them. After a little Googling I realized I didn’t need to buy a special LED to run at 12 V for the “power on” indicator; rather, using two 1-kΩ resistors I already had on hand, wired in parallel, would give me the 500 Ω resistance needed to provide proper current limitation to a regular cheap LED. After initially building a board that had the transmit terminal of the Arduino shorted to +5 V (which of course could not talk to my computer), I found and cut the shorted trace, loaded the minimal MyFocuser Pro software, and got the motor moving. 

T-adapter attached to Tamron lens, with
toothpicks holding the aperture open.

I bought some toothed pulleys with a long length of toothed belt originally intended for do-it-yourself 3D printer assembly. The ends of the belt are glued together to make a loop of the proper size. I tried to get the belt length right by trial and error and found considerable tension was needed to get it to adjust the lens focus without slipping. 

Toothed belt glued to lens focusing ring.

Camera, filter wheel, lens, and focus motor assembled.

When I finally got this rig on the sky, I found the autofocusing routines in Sequence Generator Pro worked quite well with this setup. The attached images show the difference between manually focusing once then never refocusing (in the Spaghetti Nebula image) and using autofocusing every 10 subframes and on filter changes (in the Christmas Tree Cluster/Cone Nebula image). Refocusing helps a lot in my opinion. 

The setup worked pretty well on the first night out, but much more poorly on the second night. That second night I also noticed that I could see the region of stars that were in focus march across my image from bottom right to upper left as I adjusted the focus, strongly suggesting that the optical image plane and the plane of the imaging chip were not parallel. After the session ended I found that the tight belt around the lens had put a sideways force on the lens that pulled apart the glued joint holding the lens to the T-mount ring. No wonder the lens seemed misaligned with the camera; it was! 

I cleaned off all the hot glue I had used before to make the joint, and replaced it with Gorilla Glue Epoxy reinforced  with duct tape (the magical tool of all science), which I hope will hold better. Because I had a lot of toothed belt left over, I decided to try a different way of connecting the stepper motor to the lens. I hot-glued a ring of toothed belt around the focusing ring of the lens. This meshes with the toothed belt from the pulley, making a firm, non-slip connection between the motor and the focusing ring without needing a lot of tension. Initial testing suggests it moves the focus  reproducibly without a lot of backlash or hysteresis. We’ll see what really happens next time I get a chance on the sky (which may be a while, given recent weather; even with a beautifully clear day like today, the forecast says it’s going to cloud up again tonight!).

Image taken without refocusing. The blue channel is not focused as
well as the others.

Automated refocusing using Sequence Generator Pro and the MyFocuser Pro
motor assembly. Notice how much smaller the stars appear.

Resurrecting an Old Camera Lens for Astrophotography, Part 2

Some time ago I decided to try using my old Quantaray zoom lens, once used on a Minolta film camera, for astrophotography. The experiment worked reasonably well, but at f/5.6 at best, the Quantaray lens is fairly slow. In addition, that lens was designed for autofocus use. This presents some problems for astrophotographic applications because moving the focus manually just a little makes a big difference and the manual focus is very sensitive to just a slight touch.  

Tamron 135 mm f/2.8 lens with T-adapter glued to the top

I have seen a number of nice images on Astrobin taken by John Leader using a Tamron Adaptall f/2.8 camera lens (in fact, this is what originally inspired me to try the Quantaray). Seeing a few more images motivated me to look into how much the Tamron lens might cost, and I found that used lenses are available for low cost on e-Bay from time to time. Knowing that Tamron invented the T-mount to allow their lenses to be mounted on a variety of cameras, I assumed the lens would come with T-threads, which fit my astrophotography cameras. I found a Tamron 135 mm f/2.8 lens with a Canon adapter of some sort on e-Bay for $15. Its condition was not specified, but pictures of the lens looked good and the price was so low I figured it was worth a try so I ordered it. When it arrived I found its condition was excellent, but my assumption that I could just pull off the Canon adapter and expose T-threads was ill-founded. I decided I could do something similar to what I had done earlier with the Quantaray, which was to just glue a T-adapter to the back of the lens. I used a hot glue gun and a liberal amount of glue to attach the adapter, then wrapped the joint in black rubber waterproof tape to secure the joint and discourage light leakage. A spring-loaded toggle switch enables the f/stop adjustment, but for my initial experiments I left it free so the lens remains wide open at f/2.8 (later on, I can easily shim the spring-loaded toggle so I can use the f/stop ring to stop down the lens if I want).  The attached picture shows the adapter on the lens.  I got a mounting ring for my ASI1600MM-Cool camera with electronic filter wheel. The mounting ring screws onto a dovetail bar, which also holds a finder/guider so I can put the whole thing on my mount and do autoguided LRGB and narrowband astrophotography. 

Lens, filter wheel, & camera on dovetail bar with guide scope

I chose a very challenging target, Sh2-240, aka Simeis 147, the Spaghetti Nebula, to test the setup.  This target is about 3° in diameter, much larger than I can image with any other optics I have, but it is quite dim and benefits greatly from narrowband filters, especially Hα. Therefore, this target would present a good challenge for this lens. The image was a first in many ways. It was a first attempt at this large, dim, difficult target. It was my first use of the Tamron 135 mm f/2.8 lens. It was my first try at using high gain with the ASI1600MM-Cool (gain set to 300—much greater than unity—and exposures of only 45 sec for LRGB and only 180 sec for Hα), and it was first light for the Tamron  135+ASI1600MM-Cool combination. I didn’t get any blue data (the neighbor’s darn tree got in the way again), so I cheated: I used the Hα data both for luminance and for R, and I put the R data in the B channel. All things considered, it could have turned out worse, and the lens passed with flying colors. It was easy to focus and seemed to stay focused. I won’t say that the stars looked perfect all the way to the edge, but they looked pretty good especially in the region I cropped to for the attached image. I did the cropping primarily so I wouldn't have to deal so much with the extensive vignetting produced by my 31 mm diameter unmounted filters on the ASI1600MM-Cool imaging chip. I had to process the heck out of this to get a result that was anywhere near reasonable. The supernova that produced this remnant is reported to be  about 3000 light years distant and to have gone off 40,000 years ago!

Date: 7 Jan 2017

Subject: Sh2-240 (Simeis 147) Spaghetti Nebula supernova remnant

Scope: Tamron 135 mm f/2.8 lens

Filters: ZWO 31 mm RG, ZWO 7nm Hα

Mount: EQ-6 (EQMOD)+PEC

Guiding: Finder/Guider+DSI IIc+PHD 2.6.2.4 (Win 10 ASCOM)

Camera: ASI1600MM-Cool, -20 °C, acquired Hα 1x1 RG 2x2, Gain 300 Offset 50 

Acquisition: Sequence Generator Pro 2.6.0.3

Exposure: 80x180 s Hα, 55x45 s R, G

Stacking: Neb 4.1.2, flats & darks, trans+rot align, Nebulosity 1.5σ stack. Processing: too extensive to detail