Saturday, April 16, 2016

Adventures in Building a Stepper-based Autofocuser

Sequence Generator Pro has many powerful capabilities for automating the process of capturing images. One of these is the ability to use a motorized focuser to automatically focus based on the apparent sizes of the star images. This is important because the focal position can shift significantly during a session, especially if the temperature is changing and the optical tube undergoes thermal expansion or contraction. Fortunately, here in Utah the temperature usually remains pretty stable at night once the sun goes down, but this is not always the case. So until now I have tended to focus at the beginning of my session, perhaps refocus before going to bed, and then just take my chances. Obviously it would be better to refocus periodically, especially if the temperature is changing. SGP has the capability to do that if you have hardware that is supported.

When I first began using SGP, I attempted to autofocus with the Orion Accufocus + Shoestring Astronomy FCUSB controller I have used for a long time. While this system works fine for remote, manual focusing, this is a DC focus motor with some hysteresis, meaning that focus positions are not terribly reproducible. Shoestring has an ASCOM driver with an option to mimic an absolute position-capable focuser (which a DC focuser is not, but can sometimes approximate if well adjusted). I guess some people have been able to make that work, but I couldn’t successfully do automatic focusing with the DC motor, and in fact when I tried I tended to make my focus a lot worse.

So I looked into stepper motor driven autofocus systems. In contrast to DC motor-based systems, stepper-based systems retain absolute focus positions so can repeatedly return to the same focal distance (within limits, of course). Commercial stepper-based systems tend to be relatively expensive, however; the least expensive systems I found in online searching cost more than $400.

Googling also found a “do-it-yourself” solution, MyFocuserPro, developed by Robert Brown in New Zealand and posted on SourceForge as "Arduino ASCOM Focuser Pro DIY." This system is controlled by an Arduino Nano, which has an on-board USB port to talk with the computer that controls the session. This project was attractive to me because it looked considerably easier to do than the PicGoto I had already done, and because it comes complete with a Windows-based control application and most importantly, ASCOM drivers that would allow me to use it with Sequence Generator Pro or other control software. On the other hand, I am a complete novice to Arduino-controlled systems and went into this not even knowing how to load code into an Arduino. 

One of the difficulties with this project is that it comes with a huge range of options in Arduinos, motor controllers, stepper motors, wifi vs. Bluetooth vs. IR vs. USB, etc. Frankly, it was pretty confusing. So deciding on which options I wanted was a big issue. I initially decided to go with the recommended “HW203_DRV8825” option, which uses an Arduino Nano and a DRV8825 motor power board and a high-torque, extensive microstep-capable NEMA17-PG27 stepper motor. Because I have experience with vero board and none with making my own printed circuit boards, I decided to go the vero board route. I wanted a temperature probe so I could refocus when the temperature changes, but didn’t care about an LCD display too much, so I decided to build the “minimal” option + temperature probe.

I started ordering parts, mostly from Ebay and mostly coming from China. I rounded it out with a bunch of parts from Mouser. It is simply amazing what you can get in electronic parts for very little money. More than half the total cost was for the stepper motor, and the whole thing adds up to less than $100 (and I could have done better, had I been a little more careful). As I started assembling things, I realized I really would need pushbuttons to focus at the scope, and before long I had ordered the rest of the parts for the “Full” option with pushbuttons and an LCD display.

The board was pretty straightforward to assemble, but I made a few mistakes. One of these involved forgetting one trace cut on the vero board, which left the “receive” pin on the Arduino shorted to ground. Needless to say, when I powered it up I couldn’t talk to the Arduino! I also tried to use a 5 V LED in a place where a 12 V LED was called for, and I discovered that 5 V LEDs burn out under those conditions.

Then I loaded up the Arduino programming environment and started testing, and started having problems. As I mentioned above, at first I couldn’t even talk to the Arduino. I thought maybe that was because I was trying to program from Mac OSX El Capitan and there are apparently known problems running Arduinos under the El Capitan operating system. So I loaded the Windows software in my Windows 7 VMWare virtual machine and had the same problems. That's when I discovered the forgotten trace cut. Once that was sorted out, the LED/pushbutton test worked and the test of the temperature probe worked, but I could never get the stepper to move or get any characters on the LCD. After fumbling around and not getting anywhere, I eventually decided to go on the SourceForge discussion group and start asking questions.


This is when I discovered another wonderful thing about the MyFocuser Pro project: Robert Brown gives absolutely phenomenal support! He has patiently walked me through troubleshooting that has established that the DRV8825 board and the LCD are both bad (I probably damaged the DRV8825 in my initial fumbling around, and the LCD came with a short on the board that fried it, so I guess in a sense it was defective from the beginning). But these parts are so cheap that there is essentially no risk (at least if you are patient). I finally have the Windows application talking to my board and reading temperatures, but still can’t move the motor (because the motor board is dead). I have now mounted my parts in a project box and am waiting for the new DRV8825 and LCD components to arrive on the slow boat from China. I’m confident that with Robert’s help I’ll soon have this thing working. I’ll post pictures once I get it all going (but be forewarned that my “fit and finish” are not too pretty!).

Saturday, January 9, 2016

Stepping Up to the EQ-6 and Sequence Generator Pro

I have known since pretty early in my astrophotography adventure that the single most important component of your equipment is your mount. When I got my old, used, no-go-to CG5 mount, it was a huge step up for me. My AT8IN scope is really too heavy for the CG5, but taking steps to minimize the weight (like using a finder/guider) and begin careful about balance I could usually make it work and get reasonable results. However, I’ve long had a desire to get a better mount (along with a desire to get a cooled monochrome camera). Neither of those exactly fits with the “cheap astrophotography” theme, but then this hobby is known for growing on people and subconsciously inducing them to spend money. Back in November I became aware that Skywatcher USA was putting their EQ-6 mount on sale for a pretty good price. That’s one of the mounts I’ve been quite interested in for a long time, particularly because it is supported by the EQMOD project, an open-source mount control software project that greatly extends the capabilities of the mount. So I bit and ordered one right after Thanksgiving when Black Friday prices made the deal even better. It arrived a couple of days later. I was immediately impressed with how solid and massive and finished the EQ-6 looked. I unboxed it and put the mount head on the tripod, then hunkered down to finish the semester and get ready for Christmas. I didn’t have time to play with it until after Christmas, but I ordered an EQDIR cable for direct computer control, downloaded EQMOD and EQASCOM, and built a Windows 10 virtual machine from which to run everything. That last turned out to be a bust, because I couldn’t get the Meade DSI drivers to work under Windows 10, so I fell back to my Windows 7 virtual machine. 

Finally I got a chance to try out the new mount the day after Christmas, with all my optical gear loaded on it (AT8IN with ST-80 refractor on top with the finder/guider on that and the Quantaray zoom lens bolted to the ST-80 rings). I got all the scopes and the Telrad aligned on a mountaintop through the living room window. This mount is pretty heavy to lug outside, but I can just manage with the counterweights removed. This optical stack also is a bit unwieldy. Anyway, I used a bullet level to set the mount’s home position, got polar aligned, and used Cartes du Ciel with the EQASCOM driver to start pointing the mount. It is wonderful to be able to do fast slews. I first went to Capella and it was within the field of view without doing any alignment. I slewed around and synched on 5 or six stars in various parts of the sky, then decided to do a Hα image of the Crab Nebula (the Moon was only 1 day past full). The tracking performance of the EQ-6, even prior to any periodic error training or correction, was much superior to what I’d been getting with the CG5. The resulting image is one of the sharpest I’ve ever taken, far superior to any of my prior Crab Nebula images.
 I haven’t applied any PEC yet, but the guiding performance was far superior to what I’ve been seeing with my old CG-5. I have a lot to learn about how to best use this mount, but I’m really impressed so far. The main thing that went wrong on this image was the remote focuser failed (I suspect a loose linkage or bad connection, but haven't had time to troubleshoot yet).

Date: 26 Dec 2015
Subject: M1, Crab Nebula
Scope: AT8IN 
Filter: Baader 7 nm Hα
Mount: EQ-6 (EQMOD)
Guiding: 9x50 mm finder/guider + DSI Ic + PHD 2.5.0.8 (Win 7 ASCOM)
Camera: DSI IIc, no chiller CCD -10 °C
Acquisition: Nebulosity 4.0.4, no dither
Exposure: 46x600 s
Stacking: Neb 4, bad pixel map, bias included, 33 flats, extract R channel, match histograms, square, trans+rot align, 40-60-tile stack, Catmull-Rom spline 2x resize
Processing: StarTools 1.4.305 + CS6 Astronomy Tools: Crop; Wipe; Develop: 84.75%; HDR: Optimize; Deconvolute: 3.5 pix; Track: Smoothness 88%, 4.5 pix; Levels; Astroframe.  
I used RGB data from 19 Dec 2012 and 21 Nov 2014, combined with Hα from 26 Dec 2015, to make this image. I played around with processing a bit. I healed out the stars from the Hα and combined the starless Hα with the R channel from the RGB, using a layer mask so it has the RGB data for the stars but Hα for the nebula. I used this to replace the RGB R channel, and boosted the R channel a bit using the channel mixer. I also used the Hα as luminosity for the whole image. This mostly was done to address the “blue star” problem I usually have with these images. I’ve still got a little of that close to the nebula but it’s better than before, and I have far more structure within the nebula than I’ve ever gotten with RGB alone.

After posting this image and mentioning that I was using a new mount, several Astrobin users mentioned that I should give Sequence Generator Pro a try. I downloaded it and started the 45-day free trial the next day. The first night out with it, I seriously underestimated what it would take to set it up and couldn’t get anything to work right so I fell back to just taking an image using my old Nebulosity workflow. I imaged up to the meridian, where the mount’s meridian limit kicked in and stopped the tracking, then performed a meridian flip, reacquired, and restarted imaging. I set the EQASCOM timer to park the mount just after my target set and went to bed. I woke up in the morning to find the mount nicely parked. It is really nice not to have to worry about it bumping. I also decided to reconfigure my optical stack, so I just put the ST-80 on top of the AT8IN to use as a guide scope, which is what I originally intended when I bought the ST-80. I have the finder/guider and the Quantaray lens mounted on a separate dovetail bar so I can use them on the new mount too.

I decided I needed to work a bit harder to get Sequence Generator Pro up and running. I looked at the excellent “First Week with Sequence Generator Pro” tutorial (http://www.mainsequencesoftware.com/content/SGP-The%20First%20Week.pdf), which was a big help. One of the challenges to using SGP, which I thought might be a deal-breaker, is that it doesn’t support the Meade DSI cameras. However, I was able to work around that by using Nebulosity to control the camera and invoking the Nebulosity 4 ASCOM camera driver from within SGP. I also downloaded the free PlateSolve2 software from PlaneWave to use as a plate solver with SGP. I got the camera working, then got plate solving working, then tried a test automated meridian flip using SGP. This was simply amazing! You start by unparking the mount and then either giving the target coordinates to SGP or controlling a goto from Cartes du Ciel. SGP takes an image after reaching the target coordinates and by plate solving automatically aims the camera to within a fraction of an arc second of the target. It then starts your previously-calibrated guiding program (PHD2 in my case) and commences taking images. It will also use your motorized focuser to automatically focus using half-flux radii measured for multiple stars in your field of view, but I don’t have that working yet. When the mount reaches the meridian, SGP shuts off guiding, performs the meridian flip, takes an image and uses the plate solver to again center on your target to within a fraction of an arc second, turns the guider back on, and continues imaging for the preset time, then parks the mount. These are truly amazing capabilities, well worth the $100 cost, which I plan to pay as my free trial ends. Conditions on the night I got all this going were terrible, with lots of gusty wind. It was the kind of night I normally wouldn’t even have tried to image, but with this much more massive mount I was actually able to get decent data. So in case you can’t tell, I’m thrilled with the new mount.

Monday, November 23, 2015

Resurrecting an Old Camera Lens for Astrophotography

I used to have lots of fun taking pictures with a Minolta autofocus single-lens reflex film camera. The camera was last used when one of my children took a photography class in high school; now everything has gone digital, for good reasons. But I still have the camera and its lens. I’ve seen a number of nice astrophotos on the Astrobin site taken using camera lenses. I’m particularly interested in being able to take wider field-of-view shots with my small-chip Meade DSI IIc camera, and the old camera lens isn’t being used any more, so in the full spirit of cheap  astrophotography I decided to give it a try.

The box for the Quantaray zoom lens for my Minolta camera

The available lens is a Quantaray 28-90 mm zoom lens, f 5.6 at 90 mm. So it isn’t particularly fast, but it certainly meets the “wider field-of-view” requirement. Because it is designed for use on an autofocus camera, this lens has an internal diaphragm iris that is spring-loaded, normally stopped way down. Therefore, I initially I had to overcome two problems to use the lens: 1) I had to find a way to keep the iris open; and 2) I  had to find a way to mount the lens to my DSI camera.




Fortunately, problem 1) was fairly easy to solve. The lens cap that fits on the back of the Quantaray (and therefore is designed to mount to the back of the lens) has a little plastic tab that pushes against the switch that opens the spring-loaded iris. So putting the rear lens cap on opens the iris all the way.

Problem 2) also wasn’t too bad. The front of the Meade DSI IIc has female T-threads, so I just needed to find a male T-thread adapter to mount the lens.

Lens cap (left) with T-thread adapter (right) glued in
Next, I needed to ensure that I could reach focus. I measured the distance from the mounting flange of the lens to the focal plane of the Minolta camera body, then got T-thread adapters to match that distance as closely as I could. 

Finally, I wanted to be able to use my 1.25” astrophotography filters with this rig. Fortunately, I already had adapters for inserting filters into a stack of T-thread pieces; these have T-threads externally and female 1.25” filter threads  inside, and   are available from places like Agena Astro and ScopeStuff online.

All that remained was to cut a hole through the lens cap, and glue in a T-thread component to mate with the proper set of extenders+filter holder to get the right focal distance to the imaging chip of the DSI IIc. The attached images show the T-thread adapter glued into the lens cap, and then attached to the lens with a filter adapter ready to screw into place. Finally, I show the whole assembly with my homemade adapter ready to screw into the T-thread on the front of the camera.
Lens assembly with filter installed
Cap and adapter attached to rear of lens, with filter adapter


























Initially, I used a C-clamp to attach the camera to the mounting rings of my ST-80 telescope and took a test image of the Pleiades. That was encouraging enough that I decided to put together a more stable way of mounting the lens + camera. I ended up getting a couple of 2.9” mounting rings from Agena Astro, which were just a bit too large for the lens, and wrapped the lens with duct tape to get it to fit in the rings. I drilled a couple of holes in the large mounting rings that hold my ST-80 scope, and bolted the 2.9” rings to the big mounting rings via an aluminum rail. This gives a nice, stable lens mount with no detectable differential flexure; I can guide either with the ST-80 or with a 50 mm finder/guider that I usually use, attached on the rear of the ST-80. I’ve now used this rig to get wide-field views of the Andromeda Galaxy and the Witch Head Nebula, and have also used it for Hα images of the California Nebula and of the Heart and Soul Nebulas together. This has been a fairly successful experiment.
Quantaray 28-90 zoom lens mounted on ST-80 rings





















This image is an experiment and something of a homecoming. One of the first deep sky targets I ever tried was the Pleiades, back when I was just using a Canon point-and-shoot camera with terrible amp glow issues.  I’ve come a long way since then. One of the challenges using a small-chip CCD camera like the Meade DSI IIc is trying to image larger objects; it’s just difficult the get them to fit on that little chip. About the only option is to use a shorter focal length. I’ve tried shooting through my finder scope with a focal reducer, but the optical flaws of that approach are formidable. So inspired by others like John Leader and Jammie Thouin I see using camera lenses, I decided to give that a try. I have an old Minolta autofocus SLR camera I never use any more, with an inexpensive Quantaray zoom lens we bought for use in a high school photography class. I got the distance close enough that I can get to focus at the 90 mm zoom setting (and down to about 65 mm currently). Mounting this chimera camera initially was very crude; I literally used a C clamp to hold the DSI camera body to one of the mounting rings on my ST-80 scope (which I used for guiding this image). This resulted in a lot of differential flexure and will need to be fixed with a better approach. Anyway, this is the first light image for this lens/DSI combination. I chose the Pleiades because I knew they would be easy to find and focus. Guiding failed a short way into the run, so most of this was unguided. Focusing this rig is difficult and very sensitive, but overall I’m happy with this as a first attempt.

Date: 10 Oct 2015
Subject: M45, The Pleiades
Scope: Quantaray 28-90 zoom lens at 90 mm setting
Filter: Baader Fringe Killer
Mount: CG-5 (Synta motors, PicGoto Simplificado)
Guiding: 9x50 Finder/Guider + DSI Ic + PHD 2.5.0.1 (Win 7 ASCOM)
Camera: DSI IIc, no chiller CCD 15 °C
Acquisition: Nebulosity 4.0.3, no dither
Exposure: 32x600 s
Stacking: Neb 4, bad pixel map, bias included, 33 flats, match histograms, deBayer & square, trans+rot align, DSS 2x drizzle 1.75 average κ-σ stack.
Processing: StarTools 1.4.305 Crop; Wipe 75%; Develop 81.77%; HDR optimize; Color: scientific, 275%; Deconvolute 4.8 pix; Life: heavy; Track 4.5 pix, Smoothness 88%; Magic: shrink 2 pixels. CS6 Astronomy Tools deep space noise reduction; increase star color; make stars smaller; levels & curves; Astroframe. 

More experimentation with the Quantaray zoom lens. Another large, easy target. I'm finding this lens is optically pretty slow compared to what I'm used to (I think it is f/5.6), so for the first time ever I went to 15 minute subframes (which is not hard when imaging at these large scales). I also used Nebulosity fine focus to try to improve the focus, which is more difficult with this lens and does not hold very well either. I had a set of 10 subs that were very tightly focused prior to performing a meridian flip; subs after the flip were not as good, but this image includes all of them less a few thrown out by Deep Sky Stacker's quality estimator. The results are encouraging enough I guess I'm going to need to find a better way to mount this rig.

Date: 10 Oct 2015
Subject: M31, Andromeda Galaxy
Scope: Quantaray 28-90 zoom lens at 90 mm setting
Filter: Baader Fringe Killer
Mount: CG-5 (Synta motors, PicGoto Simplificado)
Guiding: Orion ST-80 + DSI Ic + PHD 2.5.0.1 (Win 7 ASCOM)
Camera: DSI IIc, no chiller CCD 15 °C
Acquisition: Nebulosity 4.0.3, no dither
Exposure: 20x900 s
Stacking: Neb 4, bad pixel map, bias included, 33 flats, match histograms, deBayer & square, trans+rot align, DSS 2x drizzle 1.75 average κ-σ stack.
Processing: StarTools 1.4.305 Crop; Wipe 75%; Develop 79.51%; HDR optimize; Deconvolute 4.5 pix; Color: scientific, 250%; Life: moderate; Track 6.0 pix, Smoothness 88%; Magic: shrink 1 pixel. CS6 Astronomy Tools deep space noise reduction; space noise reduction; local contrast enhance; increase star color; make stars smaller; liquify/pucker; levels; enhance DSO and decrease stars; Astroframe.


An easy find using Astrotortilla plate solves by just pointing in the general direction, plate solving, and doing a goto from Cartes du Ciel. First light for the Quantaray 28-90 zoom lens mounted on 2.9" tube rings. This is also my first attempt to do a narrowband image with this setup, and it is my first-ever attempt to use 20-minute subs. Fortunately, it looks like my mounting arrangement is stable enough that I don't have too much differential flexure over 20 minutes, but of course using a 90 mm focal length makes the guiding trivially easy compared to what I'm used to. Overall, I'm quite happy with the result. It is really nice to have a way to get a 4° field of view with my little imager.

Date: 23 Oct 2015
Subject: NGC 1499, California Nebula
Scope: Quantaray 28-90 zoom lens at 90 mm setting
Filter: Baader 7 nm Hα
Mount: CG-5 (Synta motors, PicGoto Simplificado)
Guiding: Orion ST-80 + DSI Ic + PHD 2.5.0.5 (Win 7 ASCOM)
Camera: DSI IIc, no chiller CCD 10 °C
Acquisition: Nebulosity 4.0.3, no dither
Exposure: 15x1200 s
Stacking: Neb 4, bad pixel map, bias included, 33 flats, extract R channel, match histograms, square, trans+rot align, DSS 3x drizzle 1.75 average κ-σ stack.
Processing: StarTools 1.4.305 Crop; Wipe 75%; Develop 77.60%; HDR optimize; Deconvolute 4.6 pix; Track 9.0 pix, Smoothness 88%. CS6 Astronomy Tools deep space noise reduction; space noise reduction; levels & curves; Hα false color black space; Astroframe; create frame animation.


Realized I could get the Heart & Soul Nebulae together  with the Quantaray zoom lens set at 90 mm, so this is my attempt. Good conditions on Halloween night. Had a lot of trouble getting a good plate solve and getting good focus; you have to find a bright star to focus with this lens and the Hα filter. Eventually I plate solved with my finder to get close, then used that to get a bright star in the field of my imaging camera. I started getting drift/differential flexure near the meridian (I had gone to bed so wasn’t around to fix it). This image greatly benefitted from 3x drizzle stacking in Deep Sky Stacker.

Date: 31 Oct 2015
Subject: IC 1805 & Sh2-199, Heart & Soul Nebulae
Scope: Quantaray 28-90 zoom lens at 90 mm setting
Filter: Baader 7 nm Hα
Mount: CG-5 (Synta motors, PicGoto Simplificado)
Guiding: Orion ST-80 + DSI Ic + PHD 2.5.0.7 (Win 7 ASCOM)
Camera: DSI IIc, no chiller CCD 15 °C
Acquisition: Nebulosity 4.0.3, no dither
Exposure: 17x1200 s
Stacking: Neb 4, bad pixel map, bias included, 33 flats, extract R channel, match histograms, square, trans+rot align, DSS 3x drizzle 1.75 average κ-σ stack.
Processing: StarTools 1.4.305 Crop; Wipe 75%; Develop 71.13%; HDR optimize; Deconvolute 4.1 pix; Life: moderate; Track 8.0 pix, Smoothness 88%. CS6 Astronomy Tools levels & curves; deep space noise reduction; Hα false color black space; Astroframe; create frame animation.


I've never tried this object before because I didn't have the field of view needed and I felt it to be a difficult target. But the 90 mm Quantaray camera lens gives me the needed field of view so I decided to give the Witch Head a shot. I first synched on Rigel and used that bright star and surroundings to focus; focusing this lens is challenging. Then I did a PicGoto the target. With this orientation, Rigel is not in the field of view; I may want to try again with the camera in what I call the "zero" position, rotated 270° from this view, and include Rigel in the field. This was indeed a very difficult capture for me; even with 20 min. subs, the nebula is barely visible and required very heavy stretching. I wonder if the UHC-S filter would have been better, because contrast with the background sky is what is sorely needed here. But at least I got something.

Date: 14 Nov 2015
Subject: IC 2118, Witch Head Nebula
Scope: Quantaray 28-90 zoom lens at 90 mm, f 5.6 
Filter: Baader Fringe Killer
Mount: CG-5 (Synta motors, PicGoto Simplificado)
Guiding: 9x50 mm finder/guider + DSI Ic + PHD 2.5.0.7 (Win 7 ASCOM)
Camera: DSI IIc, no chiller CCD 10 °C
Acquisition: Nebulosity 4.0.3, no dither
Exposure: 16x1200 s
Stacking: Neb 4, bad pixel map, bias included, 33 flats, match histograms, deBayer & square, trans+rot align, DSS 3x drizzle with σ = 1.75 κ-σ stacking.

Processing: StarTools 1.4.305 Crop; Wipe 75%; Develop 76.2%; HDR optimize; Color: scientific, 200%; Deconvolute 3.9 pix; Track 7.3 pix, Smoothness 88%; Magic: shrink 4 pixels. CS6 Astronomy Tools deep space noise reduction; increase star color; make stars smaller; enhance DSO & reduce stars; layer masked levels; liquify/pucker; Astroframe.

Friday, May 15, 2015

Cleaning My Primary Mirror

Cleaning your primary optics is something they say you should only do as a last resort. I admit I haven't really noticed a degradation of optical performance, but my primary mirror has become quite dirty lately. Even though I'm good about keeping the dust covers on, dust gets in (likely some gets in when I'm imaging), and even though I rarely have dew issues here in desert Utah, I do get frost in the winter, which tends to concentrate the dirt. On top of that, I even had some dead bugs and airborne seeds down on the primary I couldn't just shake out, so a while back I decided to do some cleaning. It isn't hard to remove the primary mirror cell on the AstroTech 8" Imaging Newtonian; you just remove 6 screws around the base of the optical tube and slide out the mirror cell. It fits snugly in the tube so it takes a little prying to get it out. Here's the dirty mirror on the mounting cell after I removed it:
I've cleaned this mirror once before, after I left the telescope out the same night my sprinklers came on. That time I was extremely worried about scratching the mirror or removing coatings, so I used only distilled water applied with a spray bottle and some isopropanol to help in drying it out. I do think you need to avoid scratching your mirror, but I was a lot less cautious this time. I rinsed the mirror with tap water, applied a few drops of dish soap to the water-filled "dish" formed by the concave surface of the mirror, and after carefully washing my hands, rubbed the surface with my fingers. This took off the dust deposits nicely, with no apparent damage to the surface. I then thoroughly rinsed off the soap with tap and distilled water, and finished off with isopropanol to assist in drying, as before. Here's the resulting cleaned mirror:
I ran into a little difficulty getting the mirror cell back into the optical tube. As I noted above, the fit is pretty snug so it took some prying to get the cell back into place, with the screw holes in the cell aligned with the holes in the mounting tube. After a little effort, I got that done without dropping anything. I used a laser collimator to line up all my optics, and was ready to get back to imaging. To be honest, I have not noticed a difference in my images after going through the cleaning procedure. I imagine it would be manifest in higher image contrast, but with the galaxies I've been trying to image recently I think the darkness of the background sky is a much more important factor.

Saturday, May 2, 2015

The Season of Galaxies

I've had a very busy winter term and haven't taken the time to post in quite a while. But now the term is over and I've got the urge to write a bit.  This time of year is known as "galaxy season" among amateur astronomers, because in the Northern Hemisphere the galaxy-rich constellations of Ursa Major, Leo, Virgo, and Coma Berenices are all well-placed for viewing and astrophotography. As a result, I've been imaging galaxies during most of the time since my last post, and have stepped away from narrowband H-α for a bit because most galaxies are not best photographed in H-α (although it can certainly help in some cases). Most of what I've been doing is just one-shot color imaging, going as deep with the subexposures as I can (currently my longest are 600 s). I'm getting a lot more integration time these days than in past seasons (I almost always image up to the meridian, do a meridian flip, and continue until dawn or the setting of the target, whichever comes first), and this has allowed me to stretch the data more aggressively without bringing up as much noise. I'm also routinely trying both Nebulosity stacking and Deep Sky Stacker 2x drizzle stacking (using subframes preprocessed and prealigned in Nebulosity) for every image. Sometimes drizzling helps, and sometimes it doesn't.  Anyway, here are a few galaxies. M81 is a personal favorite, and I'm getting better at imaging it:

A beautiful moonless night, so I decided to go back to one of my favorite and most-attempted targets, M81. Because I'm routinely doing 600-s subframes now in Hα, I decided to try to go deeper with M81 by using 600-s subs and the UHC-S filter. Finding was via synch on α UMa followed by PicGoto acquisition. I had some trouble initiating guiding; calibration went OK, but then the guide star started drifting off in Dec. I think my Dec gear clutch was just loose. Imaged up to the meridian, then re-synched with Astrotortilla and reacquired and imaged until dawn. I was thrilled to pick up what appears to be a small satellite galaxy at top center. I wasn’t sure that wasn’t an artifact, but now I’ve seen it in others’ images so it must be real.

Date: 16 Feb 2015
Subject: M81, Bode’s Galaxy
Scope: Astrotech AT8IN+High Point Scientific Coma Corrector
Filter: Baader UHC-S
Mount: CG-5 (Synta motors, PicGoto Simplificado)
Guiding: 9x50 Finder/Guider + DSI Ic + PHD 2.4.1 (Win 7 ASCOM)
Camera: DSI IIc no cooling (about 0 °C)
Acquisition: Nebulosity 3.3, no dither
Exposure: 41x600 s
Stacking: Neb 3.3.2, bad pixel map, bias included, 33 flats, histogram match, deBayer & square, trans+rot align, 40-60% median combine.

Processing: StarTools 1.3.5.289 Crop; Wipe; Develop 87.49%; HDR:Optimize; Color: Scientific, 300%; Deconvolute 2.5 pix; Track 3.5 pix; Magic: 1 pix. Photoshop CC 2014 + Carboni Astronomy Tools Deep space noise reduction; Increase star color; Layer masked unsharp mask for dust lane enhancement; Astroframe.

M82 is nearby and also a favorite. This is one where I thought maybe adding in some Hα would help pick up the jets streaming from the galaxy, so I tried that too. I'm not sure it helped, but here are the results.
Since I was already in the neighborhood the night before (imaging M81), I decided to try and get a better image of M82 by using similar tactics. Began by synching on α UMa with a PicGoto slew to the target; it was in the field of view. Used the UHC-S filter because I'm looking north toward the Provo light dome, and used 600-sec subframes again to try and go as deep as possible. Started imaging, refocused a couple of times, and took a nap, with an alarm set for transit time. Got up and did the meridian flip and refocused, then went to bed.

Date: 17 Feb 2015
Subject: M82, Cigar Galaxy
Scope: Astrotech AT8IN+High Point Scientific Coma Corrector
Filter: Baader UHC-S
Mount: CG-5 (Synta motors, PicGoto Simplificado)
Guiding: 9x50 Finder/Guider + DSI Ic + PHD 2.4.1 (Win 7 ASCOM)
Camera: DSI IIc no cooling (about 0 °C)
Acquisition: Nebulosity 3.3.2, no dither
Exposure: 43x600 s
Stacking: Neb 3.3.2, bad pixel map, bias included, 33 flats, histogram match, deBayer & square, trans+rot align, 40-60% median combine.
Processing: StarTools 1.3.5.289 Crop; Wipe; Develop 87.49%; HDR:Reveal core; Color: Scientific, 230%; Deconvolute 2.5 pix; Life:Moderate; Track 4.0 pix; Magic: 1 pix. Photoshop CC 2014 + Carboni Astronomy Tools Deep space noise reduction; Increase star color; Layer masked unsharp mask for dust lane enhancement; Astroframe.
Here it is with Hα added:
I'm filling in these notes almost a week after the fact because I was so dissatisfied with the results that I wasn’t going to post them. There was a lot of moon the night these were taken, so I decided to see how M82 looks in Hα, thinking that might enhance the filaments that emanate from the galaxy. It didn't work very well, but it was fun to give it a shot. I had some guiding issues and also didn’t do a very good job of focusing, so the galaxy Hα experiment was not too successful. I did drizzle process these Hα images (in Nebulosity 3), and also tried drizzling my earlier (17 Feb 2015) UHC-S images (using Deep Sky Stacker; I couldn’t get it to look good using Nebulosity—DSS was just better for this image). Finally I scaled the two images in Nebulosity 3 and used a 50:50 combination of the R channel from the UHC-S image and the Hα as R in this final image, with most of the background from the UHC-S. Here it is for what it’s worth.


And then a few other galaxies:
Found very easily by initially synching on ψ UMa and using the PicGoto. This is a beautiful little galaxy. I'm wondering if I would have been better off withouth the UHC-S filter. I had a lot of guiding issues as well, probably due to non-optimal balance (balance is critical with my finicky old mount). Finally, my primary mirror has gotten absolutely filthy so I really do need to clean it. Nevertheless, the image turned out OK but not great. 

Date: 19 Mar 2015
Subject: NGC 3198, spiral galaxy in UMa
Scope: Astrotech AT8IN+High Point Scientific coma corrector
Filter: Baader UHC-S
Mount: CG-5 (Synta motors, PicGoto Simplificado)
Guiding: 9x50 Finder/Guider + DSI Ic + PHD 2.4.1 (Win 7 ASCOM)
Camera: DSI IIc no cooling (about 10 °C)
Acquisition: Nebulosity 3.3.3, no dither
Exposure: 23x600 s
Stacking: Neb 3.3.3, bad pixel map, bias included, 33 flats, histogram match, deBayer & square, trans+rot align, 40-60% median combine.
Processing: StarTools 1.3.5.289 Crop; Wipe: Develop: 87.49%; HDR:Optimize; Color:Scientific 248%; Deconvolute 3.0 pix; Life:Moderate; Track 5.4 pix; Magic:Shrink 1 pix. Photoshop CC 2014 + Carboni Astronomy Tools Deep space nr; layer masked space nr; Increase star color; Astroframe.

This galactic cluster always proves challenging for me. I thought I'd try it on an extremely dark night, go to 600-s subs, and use only the UV-IR cut filter. It is still challenging. I continue to have guiding problems; it is just hard to guide well enough with my rickety old mount when using my tightest imaging scale. I tried both my "normal" workflow in Nebulosity and using 2x drizzle stacking in Deep Sky Stacker; I think the DSS drizzling worked better, and here is the result. I didn’t want to darken the background too much, because the ring galaxy at upper left has a faint outer ring I didn’t want to disappear.

Date: 20 Mar 2015
Subject: HCG 44 / Arp 316, Galaxy Cluster in Leo
Scope: Astrotech AT8IN+High Point Scientific coma corrector
Filter: Baader Fringe Killer
Mount: CG-5 (Synta motors, PicGoto Simplificado)
Guiding: 9x50 Finder/Guider + DSI Ic + PHD 2.4.1 (Win 7 ASCOM)
Camera: DSI IIc no cooling (about 0 °C)
Acquisition: Nebulosity 3.3.3, no dither
Exposure: 29x600 s
Stacking: Neb 3.3.3, bad pixel map, bias included, 33 flats, histogram match, deBayer & square. DSS 2x drizzle app (1.5)-sigma combine.
Processing: StarTools 1.3.5.289 Crop; Wipe:76%; Develop: 83.33%; HDR:Reveal core; Color:Scientific, 200%, cap green to yellow; Deconvolute 2.0 pix; Track 6.7 pix; Magic:Shrink 2 pix. Photoshop CC 2014 + Carboni Astronomy Tools Deep space noise reduction (2x); Star color enhance; GradientXterminator; Astroframe.

More galaxies coming in future posts!

Saturday, February 14, 2015

A Celestial Christmas Tree

One of my favorite deep sky objects is NGC 2264, the Christmas Tree Cluster and Cone Nebula. It looks a lot like a Christmas Tree to me, even if it moves into optimal position in my sky a bit after Christmas. It is reported to be about 2600 light years away.

I first imaged this in RGB a couple of years ago, and was thrilled with the result at the time. I liked the combination of colors with bright stars showing diffraction spikes. I had to process this extensively, and the data were a bit grainy even after long exposure, but this is still one of my favorite images:

Found the Christmas Tree Cluster on the first try in the guide scope.  It is easily visible at 1 s exposures, and the star at the "trunk" is especially bright.  I'm battling the guiding a bit tonight, but I'm happy to be out as this is the first clear night we've had in about a month.  CCD is at 2 °C.  Decided to stop & check focus after 16 subs, and found it was way off.  It is sure nice to be able to use Nebulosity fine focus and Shoestring Focus to fix things like this from indoors at mid-run.  CCD temperature dropped during the run.  I imaged until this got into the Spanish Fork light pollution dome a bit too far.  

Date: 1 Feb 2013
Subject: NGC 2264, Cone Nebula
Scope: AT8IN+Antares 0.5x telereducer
Filter: None
Mount: CG-5 (Synta motors)
Guiding: 9x50 Finder/Guider + DSI Ic + PHD 1.14.0
Camera: DSI IIc (no chiller, T = 2 °C to -2 °C)
Acquisition: Nebulosity 3.1.3, no dither
Exposure: 45 x 300 s
Stacking: Neb 3, bad pixel map, bias included, normalize first, trans+rot align, 1.5 SD stack.
Processing: Neb 3 crop. StarTools 1.3 Develop, HDR:Reveal, Sharp, Deconvolute, Track, Magic:Shrink, Magic:Tighten, Life:Moderate, Rotate:90°. CS6 Astronomy Tools increase star color, AstroFrame.

Because I've now seen many examples of picking up much more detail in Hα light than you can see in unfiltered RGB, I've been looking forward to imaging this object in Hα, and finally got a chance on the 31st of Jan. this year. Unfortunately, the Moon was nearing full and was quite close by on that night, but I thought maybe it wouldn't affect my narrowband imaging so I went ahead. The result didn't show as much detail as I had hoped it would, but still showed more than I could see without a filter. This is an animated GIF transitioning between grey scale and false color.

Synched on Betelgeuse because  my Telrad is aligned well enough and I needed a bright star to focus on anyway. Used the PicGoto and as usual was right on target. A lot of Moon, very nearby; it was causing horrible shadows in my guide scope. Maybe I should have chosen a target a bit farther from the Moon. But I went ahead anyway because I haven't had many clear nights lately. I'm testing my new butyl rubber-tape-wrapped-connection USB cabling and it worked beautifully: no disconnects! (Admittedly, I need to test longer before I declare victory on this one). Imaged up to the meridian with relatively poor guiding performance, did a meridian flip and readjusted my counterweights, then continued. Used the Astrotortilla synch this time because my Telrad was frosted over and was at an awkward angle; it worked fine after initially getting a bad goto because I forgot to uncheck "inverted" on the PicGoto software after doing the flip. Guiding after the flip was much better. Started getting clouds immediately after flipping, but as I was heading to bed it looked like they were dissipating. Did a set of sky tee shirt flats in the morning. You can see the cone in these images but not much other detail in the nebulosity until processing brought a bit more out. I may want to try this again when the Moon is not so near just so I can compare how much it affects Hα. Not enough data; still pretty noisy.


Date: 31 Jan 2015
Subject: NGC 2264, Christmas Tree Cluster/Cone Nebula
Scope: Orion ST80 + Antares 0.5x telereducer
Filter: Baader 7 nm Hα
Mount: CG-5 (Synta motors, PicGoto Simplificado)
Guiding: 9x50 Finder/Guider + DSI Ic + PHD 2.4.1 (Win 7 ASCOM)
Camera: DSI IIc no cooling (about 0 °C)
Acquisition: Nebulosity 3.2.2, no dither
Exposure: 24x600 s
Stacking: Neb 3, bad pixel map, bias included, 33 flats, extract R, histogram match, square, trans+rot align, 1.5 SD stack.

Processing: StarTools 1.3.5.289 Crop; Develop 70.43%; HDR:Optimize; Life:Less=More; Track 4.3 pix; Magic 1 pix. Photoshop CC 2014 + Carboni Astronomy Tools Deep space noise reduction; layer masked space noise reduction; Hα false color black space; Astroframe; assemble into animated GIF.

So I decided to try again on a darker night and see what effects the Moon had on the image. It turns out the Moon had a huge effect, so I guess I've learned another lesson: watch out for a bright nearby Moon, even if you are doing narrowband imaging! I think you'll be able to see the difference easily:

Wanted to get a new set of images to  compare with the 31 Jan set that was taken right next to a bright Moon (this would give me an idea about how much the Moon matters). Synched on Betelgeuse and used the PicGoto; right on, as usual. Needed to refocus periodically, and the mount appeared to be suffering from stiction. Imaged up to the meridian, performed a flip, re-synched on Betelgeuse, reacquired, continued until the Cone/Christmas Tree set. The Moon matters a lot when it's as close as it was on 31 Jan 2015. This image is far better. HαRRGB coming

Date: 11 Feb 2015
Subject: NGC 2264, Christmas Tree Cluster/Cone Nebula
Scope: Orion ST80 + Antares 0.5x telereducer
Filter: Baader 7 nm Hα
Mount: CG-5 (Synta motors, PicGoto Simplificado)
Guiding: 9x50 Finder/Guider + DSI Ic + PHD 2.4.1 (Win 7 ASCOM)
Camera: DSI IIc no cooling (about 0 °C)
Acquisition: Nebulosity 3.3, no dither
Exposure: 38x600 s
Stacking: Neb 3, bad pixel map, bias included, 30 flats, extract R, histogram match, square, trans+rot align, 1.5 SD stack.
Processing: StarTools 1.3.5.289 Crop; Autodev; HDR:Optimize; Deconvolute 4.5 pix; Track 2.5 pix; Photoshop CC 2014 + Carboni Astronomy Tools Deep space noise reduction; Hα false color black space; Astroframe; assemble into animated GIF.

Finally, I combined the original RGB data from 1 Feb 2013 with my new Hα from 11 Feb 2015 (I didn't use any of the 31 Jan data for this). I aligned the two images using Nebulosity 3.3's "translate, rotate, and scale" procedure, then mixed the R channel from the RGB with the Hα, 25% and 75%, respectively. This HαR became the new R channel in the final image; I dialed it up 130% in the CS6 channel mixer to mitigate the "blue star" problem I tend to have. I also used the HαR as a luminance layer for the final image, at 60% opacity. I really love the colors in the result, although this blended version does lose some of the detail that can be seen in the Hα image (someday I'd love to get a monochrome camera and a full set of filters so I could have it all). Here's the full-color version:

Tuesday, January 27, 2015

Jupiter Triple Shadow Transit, 23 Jan 2015

Although I'm not much of a lunar/planetary astrophotographer, one thing that is exciting about taking pictures of nearby objects is that unlike deep sky objects, nearby celestial bodies show a lot of dynamic change. Jupiter is a great example: it is easy to see the rapid rotation of the planet, the orbital motions of its moons, and dynamic events like transits, eclipses, and shadow transits (the passing of a moon's shadow across the face of Jupiter).

Transits and shadow transits occur frequently. I use the "JupiterMoons" app from Sky and Telescope to find out about them. But last Friday night was special, because it was one of the rare times when the shadows of 3 of Jupiter's big Galilean moons were simultaneously visible transiting the disk of the planet.  These "triple shadow transits" happen roughly once in a decade; I believe the next one will not occur for another 17 years. Luckily, two good things happened last Friday: I saw a reminder of the triple shadow transit that afternoon, and the weather here was clear. So even though I'm not very skilled at lunar/planetary data acquisition or processing, I decided this was something I could not pass up.

I used a $30 Logitech c270 HD webcam as my imager. It has fairly small pixels (I believe they are 2.8 microns square) and has reasonably high data transfer rates with a USB2 interface. I modified it to fit my focuser by removing the lens on the front and replacing it with a 1.25" adapter (I just hot glued the adapter onto the camera). I did all this many months ago, so I have some experience using the camera. I use a 3x Meade Barlow lens to get a more magnified image than I could get with just the 800 mm native focal length of my AT8IN newtonian telescope; this puts me at 2400 mm focal length (still not great for planetary work, but the best I can do with my equipment). I had wanted to use SharpCap to control my imaging session, but as Jupiter rose above the mountains to my east I could not get VMWare/Windows 7 to recognize the camera. I was able to see it fine via Mac OSX 10.10.1, so rather than waste time fighting the software I decided to go with the Mac software I have, imaging with BTV and controlling the camera settings with WebCam Settings.

The next challenge was getting focused and getting Jupiter into the field of view. I focused by finding a nearby streetlight, figuring that would get me close enough, and it did. Then I took an image through my guide scope and plate solved with Astrotorilla, then used the PicGoto to slew to Jupiter's coordinates, obtained from Astroplanner. This got Jupiter into my guide scope field of view, but not into the imager (my guide scope was only roughly aligned). From there I just slewed around until I got lucky. I focused and started imaging.

The goal when shooting Jupiter is to get as many frames as possible so that your software (Registax in my case) can find the few where the seeing is good and stack them. At first I used the full field of view of the camera (1280 x 800, if I remember correctly) and 2 minute videos. This leaves lots of wasted black space around Jupiter and makes the data transfers (and the frame rate) relatively slow, so I cut back to 640 x 480 to make the videos smaller and improve the frame rate. In retrospect, I probably should have cut down even more to improve my frame rate further. I had intended to just make one long video and cut it into 2 minute segments for processing and stacking, but BTV apparently is limited to 3 minutes so I went with that and refocused every 3 minutes. This is a little too long because the planet's rotation is reported to blur images using more than about 2 minutes' worth of data, but I went with it. Seeing wasn't great so focusing was hard, and this is obvious in some of my stacked data.

After Registax processing using automatic alignpoints and 75 best frames per alignpoint, followed by wavelet sharpening, I got a series of frames that I combined using Photoshop CC 2014. I took a lot of time adding in timestamps for the various frames. Here's the result. It’s interesting that it is possible to see Europa’s disk (not the shadow, the actual disk) entering transit and Io’s leaving (both happen at about the same time, at about 00:06-00:010). It is also interesting to observe the relative speeds that the shadows move across Jupiter. At the beginning of the session, the shadows of Io and Callisto overlap near the center of the image, but Io's moves off quickly, reflecting the fact that Io is the innermost Galilean and so its orbital speed is high. Likewise, Europa's shadow (entering from the right of the image) moves more quickly than Callisto's because Europa is the 2nd closest Galilean to Jupiter whereas Callisto is the outmost big moon.
Despite my less-than-stellar planetary  astrophotography skills, I couldn’t pass up a chance to image the triple shadow transit. Seeing was variable, but in the best images such as the title frame the detail was reasonable given my equipment and skill level. It’s interesting that it is possible to see Europa’s disk (not the shadow, the actual disk) entering transit and Io’s leaving (both happen at about the same time, at about 00:06-00:010). I would have liked to use SharpCap to control this session, but I couldn’t get VMWare/Win 7 to recognize the Logitech webcam, whereas MacOS was seeing it so I went with the MacOS solution.

Date: 23 Jan 2015
Subject: Jupiter triple shadow transit
Scope: AT8IN, 3x Barlow(f/12, 2400 mm)
Mount: CG-5
Guiding: PHD2 (on a star)
Camera: Logitech c270 HD at 640 x 480
Acquisition: BTV+WebCam Settings
Exposure: 180 sec, ~15 fps mov
Stacking: Registax 6, automatic alignpoint selection, best 75 frames/alignpoint

Processing: Registax 6 wavelets, CS6 combine into animated GIF