About the Object
NGC 253 (Caldwell 65) is also known as the Sculptor Galaxy, the Silver Coin, and the Silver Dollar Galaxy.  It is a starburst galaxy undergoing a period of intense star formation.  It lies at a distance of 11.4 Mly, subtends 27.5' x 6.8', and has an apparent magnitude of 8.0.  It's large size, brightness, and complex dust structures make it a very popular target for amateur imagers at all levels.  It looks like there are hundreds of images of this target here on Astrobin so this image is my contribution to the pile.

Commissioning the 20"
This is the first image that I've been able to process from my 20" in Chile at Obstech.  I picked the Sculptor galaxy as a first target simply because it was high in the sky when we first set up the telescope.  It is large, bright and extremely detailed so it made a great target for commissioning the scope.  As we got things going, a lot of things were changing.  I switched between operating modes in the camera, the camera was rotated by almost 90 degrees when the scope was re-wired, and I fiddled with the guide settings and the guide camera focus.  In the end, the yield for this entire data set turned out to be around 25 percent, which probably isn't all that bad considering all the stuff that went on while I gathered the data.  Since the QHY600 doesn't have a shutter, I had to gather darks by pointing at the floor under a dark sky and exposing through the S2 filter to minimize stray light getting to the sensor.  The guys at the observatory are finishing up building a "dark panel" for me for future use.  It will simply be a large, flat black panel positioned about 2" from the end of the scope where I can point to minimize light from getting into the scope at night for taking darks.  For flats, I had to use the data that I gathered in my shop back in Bend.   Since the camera was rotated, the flats didn't perfectly align but they were usable mainly because I had cleaned the system so well that there is no dust motes in the system.  In this case, the flats are only correcting for radiometric fall off and vignetting.  Since the images were rotated, the depth of the image stack varies with position in the image but I was able process and crop the image so that it's not very noticeable.

Guiding and Focusing with SkyGuard
With this image, I'll describe in some detail how guiding and focusing with SkyGuard (SG) works. Although I had used SG for guiding back in Bend, this was my first really detailed long term experience with it.  For any of you not familiar, SG is a software product developed and sold by Innovations Foresight (IFI) that provides both guiding and focusing capability when used with an IFI ONAG guider.  I've already written about the design of the custom ONAG on this scope, which at this point is one of a kind.  SkyGuard uses full frame guiding by computing the covariance function between a reference frame and subsequent frames from the guide camera.  This approach means that the shape of the stars don't matter and everything, including the object itself, contributes to the guide signal.  A key advantage of full frame guiding is that the guide signal does not come from the iso-kinetic patch surrounding a single guide-star within the field of view.  The covariance guide signal represents the "best" average guide signal over the entire frame to minimize chasing seeing induced tilt within a single iso-kinetic patch.  The other advantage of this system is that since it uses the entire frame for guiding, there is no need to select a guide star (either manually or by software).  It is a true point and shoot system for guiding.  I want to also be clear that SG will work just fine for guiding with any method--you don't need an ONAG unit to use SG for guiding.

To implement autofocusing with SG, you do need an ONAG guider.  The astigmatic focusing system on this scope is very similar to the auto-focusing system used in DVD players.  The converging beam from the telescope passes through a tilted beamsplitter, which generates astigmatism. By carefully co-focusing the guide-camera with the imaging camera, best focus is achieved when astigmatism is minimized (or when reaching a predetermined value.)  In that case, star roundness becomes an indicator of the focus position.  The beauty of this system is that the roundness of the star encodes both direction and magnitude of defocus--and the sensitivity of the system remains constant through focus.  The sensitivity of the system is in the range of 1/8 - 1/10 wave in the wavefront.  Of course seeing plays a significant role in the results but we handle that by time-averaging the response of the system to create a long time constant so that focus changes are done slowly over time as the system runs.  Auto-focus provides rapid initial focus and continuous correction while the shutter is open.  This handles focus shift due to thermal variations and mechanical flexure due to gravity at different pointing angles.  

A common notion about Planewave scopes is that because the truss structure is made of carbon fiber, they are immune from thermal variations.  It turns out that carbon-fiber is tricky stuff and if it's done just right, a carbon-fiber structure can indeed have a very low CTE over a finite temperature range; however, that's not always the case.   One thing that helps in this particular location is that the temperature variation appears to be pretty stable  throughout the night (within 2-3 deg C).  Still, carbon-fiber structures have mechanical flexure and respond to thermal variations so when you look closely at focus variations through the night, focus on my my scope does indeed drift.  The nice thing about focusing continuously is that I never have to stop to run a V-curve and that results in a significant gain in throughput.  The sensor is always held at seeing-limited focus point.

SG derives the focus signal from the guide camera image by computing the auto-covariance of the entire frame (which is just the covariance of the image with itself.)  This is a very efficient way to extract the focus signal from all of the stars (and objects) in the field with a single global computation.  There is no need to identify individual stars and fit stellar profiles to compute the focus signal.  This 20" CDK scope has a large central obscuration so the image of a guide-star is not very good.  Remember that the tilted beamsplitter not only introduces astigmatism, it also introduces spherical aberration as well as chromatic aberration.  So, guide stars do not appear as tiny pinpoints of light, but that doesn't matter.  By computing the autocovariance of the field, we get a very precise measure of the roundness of the stars in the field--independent of their shape and without having to do a lot of pre-processing on the signal to figure it out.  Remember that "roundness" is the direct indicator of the focus signal with this optical configuration.

Overall SG works extremely well, but that doesn't mean that I haven't come up with a number of suggestions for Gaston and his team for minor improvements here and there.  Still SG is an amazing piece of work and I am very impressed with how well thought out it is.  I give it my highest recommendation for anyone using ONAG--and it works well for guiding even when using OAG.

Processing:  Getting the Colors "Right"
As with many galaxy processing projects, I struggled to get the colors right and I wound up taking two runs at it before I was satisfied with the result.  There are a couple of ways to tell if the color balance is at least reasonable.  The first is to look at the distribution of colors in the galaxy based on expectations of the stellar population distribution.  Most galaxies have a creamy yellow/orange core and young, blue stars in the outer arms.  Emission nebula within the outer arms will show up as magenta since hydrogen emits most strongly in the red (for H-alpha) with a bit of blue from the H-beta emission line.  Next, it's important to look at the surrounding background star colors.  For most regions of the sky where we aren't looking through great clouds of interstellar dust, we should see a relatively even distribution of colors in the background stars--ranging from deep blue to deep orange-red.  I don't know how correct it is in an absolute sense, but this color balance meets all of my criteria for "reasonable-ness".

I processed this version at the full resolution out of the camera and only downsampled the result to 2x2 bin using the IntegerSample tool in PI in order to upload it to Astrobin.  Zoom in and look around.  There is a lot going on in this galaxy!  As usual, C&C are always welcome.

John