Overview of the Imaging Procedure

The image of NGC253 was captured using an SBIG ST-10XME and Celestron 14" SCT operating at f/7 (with a Reducer/Corrector in place).  The Reducer was used in order to provide a wide enough field for the object and to allow shorter exposures.  The image is an LRGB composite.  This requires shooting a set of images through the clear filter to create a black-and-white luminance image, then shooting a set of images through red, green, and blue filters to create a color image.  The color image is then combined with the luminance image to create a final image.  The reason for shooting an LRGB instead of a plain RGB is that it actually allows for a shorter overall exposure time.  This is because the RGB images are shot with the CCD camera binned.  This increases the sensitivity but lowers the resolution.  By then shooting the luminance image at full resolution then combining it with the low-resolution color image produces a high-resolution color image that actually requires less total exposure time.  This is discussed in more detail below.  A total of 15 images were captured, plus 6 dark frames.  The camera was self-guided by the built-in autoguider of the ST-10XME.

Basic Steps

• Polar align telescope

• Attach CCD camera

• Set CCD camera temperature

• Balance telescope

• Align computerized mount

• Focus CCD camera

• Find target

• Focus telescope

• Find guide star

• Calibrate autoguider

• Begin guiding

• Start Luminance exposure sequence

• Stop autoguider

• Select Red filter

• Focus telescope

• Begin guiding

• Start Red exposure sequence

• Repeat previous 5 steps for Green and Blue

• Take dark frames

Initial Setup

Basic setup includes balancing the telescope, drift polar alignment, and aligning the computerized mount, each detailed in their respective sections on this site.  This may be unnecessary for a permanently-mounted telescope, but this setup was a portable instrument transported to a dark site for imaging.  For this setup, the telescope was initially roughly polar aligned and balanced for visual use.  Drift polar alignment was done visually with a crosshair eyepiece.  The mount was then powered off, the eyepiece removed and the focal reducer and CCD camera installed, the telescope re-balanced, the mount powered up again, the CCD turned on and started cooling, and the mount computer aligned using the CCD.  For aligning the mount the CCD was only roughly focused as accurate focus was achieved at a later step.

The CCD was cooled to -5°C.  There are nice things about living in Arizona, but one drawback is that cooling CCDs is difficult.  The ambient temperature while imaging NGC253 was around 70°F.  Imaging was done from about 11:00PM to 1:00AM in mid-September.

Above:  C14, AP900, and ST-10 used for imaging NGC253

Focusing

The telescope was roughly focused for the purpose of aligning the computerized mount.  The scope was then pointed to the target, NGC253, and a more precise focus was achieved.  Focusing was done manually using a Feathertouch SCT MicroFocuser and the Focus mode of MaxIm DL.

A 1-second exposure was used for focusing.  Initial focusing was done with the camera Binning set to 3.  Then a subframe was selected and the Binning set to 1.  Focus was then achieved using the Inspect mode of MaxIm DL by minimizing the FWHM value.  Manual focusing is not as fast as using an autofocuser, but in this case it took less than 5 minutes.

Above:  Initial focus settings.  For fine focusing Binning was set to 1.

Above:  Minimizing the FWHM value in the Inspect window

Finding a Guide Star

Finding a guide star is sometimes a bit tricky, although if you ever tried finding a guide star for manual guiding using an off-axis guider during the days of film astrophotography, you know it is infinitely easier than it used to be!  Software is very helpful here.  For this image, Desktop Universe (now incorporated into Starry Night Pro Plus) was used to determine the best position for the camera in order to frame the galaxy properly and to locate a guide star.

NGC253 is large enough that it only fits into the field of view of this telescope/CCD combination diagonally.  On a German equatorial mount, such as was used here, the default position of the guide chip in an SBIG self-guiding camera (with the cables hanging down from the camera) is to the east of the main CCD chip.  This is shown below.

Above:  Initial orientation of the CCD and guide chip

Any star between the two dotted circles is accessible to the guide chip simply by rotating the camera.  There are three relatively bright choices.  The bright star at the top is the best choice, in terms of brightness, but the target galaxy would not fit in the field with the camera in that orientation.  The same is true of the potential guide star directly below the galaxy.  However, by convenient coincidence, the guide star to the lower left of the galaxy (at about the 8 o'clock position) is perfect.  When the camera is rotated to place that star on the guide chip, the galaxy is framed perfectly.

Above:  Selecting a guide star and framing the galaxy

According to Desktop Universe this is 160° from the initial position.  So the camera is rotated slightly less than halfway around.  The cables are now pointed up away from the dovetail plate attaching the scope to the mount and slightly to the south, or right side if looking at the back of the scope.  If this seems complicated, that's fine:  I agree.  So I made a cheat sheet for determining CCD orientation with a German equatorial mount.  I carry this with me into the field for imaging.

After finding the guide star, be sure to go back and double check that your main target is framed properly.

Autoguiding

Once the guide star is located, the autoguider needs to be calibrated.  This tells the software how to adjust the mount to correct for any tracking errors.  The guide rate (the speed at which the mount moves to correct errors) and the calibration time (how long the mount is moved during the test exposures to calibrate the guider) will vary from scope to scope.  I usually use a guide rate of 0.5x with a long-focal-length scope such as the C14, and 1x with short-focal-length scopes like small refractors.  An appropriate calibration time is then about 5 to 10 seconds.  I try to get the guide star to move around 20-30 pixels.  See the Software Instructions section on using MaxIm DL for autoguiding for more details on the exact procedure used.

Once the calibration is successful, autoguiding can begin.  A typical guiding exposure is 2-3 seconds long unless the star is unusually faint (or unusually bright, which is uncommon).  For this setup, only LRGB filters were used, so the guiding exposures were short.  In situations where narrowband filters such as Hydrogen-Alpha are used, guiding exposures can be 10-20 seconds.  I let the autoguider run for a full minute or so (20-30 guiding corrections) to be sure everything is working fine and to let the system settle in.  You will see autoguider errors listed.  The maximum allowable value depends primarily on the focal length of the scope and the pixel size of the CCD.  An error of 2 pixels is only 1 arcsecond on a C14 at f/7 with an ST-10XME.  The same 1 arcsecond error on a 500mm-focal-length refractor is only 0.36 pixels.  Once I see that the guiding errors are not too large, I begin the exposures.

Luminance Exposure

Once the software is autoguiding the mount, the exposures can begin.  Luminance exposures are binned 1x1; in other words they are shot at full resolution.  Multiple images are acquired and then stacked together to produce a final image.  It is preferable, for example, to take three 5-minute exposures versus one 15-minute exposure.  So two things need to be determined: the length of each exposure and the total number of exposures.  The Imaging Theory section has a discussion on ideal exposure times and subframes, full of all sorts of delightful equations.  On the other hand, you could do what I do: pick based on attention span.  For this exposure sequence I took six 10-minute images.  10 minutes is about how long I can go without checking in on the computer to make sure everything is working right.  Also, if things go bad during an exposure, I'm only out 10 minutes.  It turns out that mathematically this is not too far off from the ideal exposure time (it's actually a bit longer than ideal).  Six total exposures is a good number for stacking.  More is better but there is a point of diminishing returns.  Also, as will be seen below, this produced a total imaging time of two hours for both luminance and RGB, and that's about how long I can stand to image one part of the sky.  Someone with more patience could likely have imaged longer and captured even more detail.

In MaxIm DL a Sequence was setup to capture six 600-second exposures.  The sequence was begun and each image was briefly checked after downloading to make sure everything looked okay.

Color Images

After the luminance sequence was finished, the autoguider was stopped.  This was done to prevent losing the guide star while the filters were changed.  The Red filter was selected and the guide star reacquired.  Calibration does not need to be done after changing filters unless a longer exposure is required.  If the guide star is reasonably bright during the luminance exposure, there will be no need to lengthen the exposure for the color filters.

It is best to check focus when changing to the red filter (and again for each other color).  The lenses in the optical system may cause each color to focus slightly differently.  Also, after an hour of luminance imaging, focus may have changed due to the temperature dropping.

Selecting RGB Exposure Times

The color images were binned 2x2, or shot at half resolution.  This increases the sensitivity of the camera and allows shorter exposures.  Since the final resolution of the image will come from the luminance image, shooting low-resolution color frames is fine.  The trick is to get approximately the same amount of data in both the luminance and RGB images.  This is simply a matter of determining equivalent exposure times, which is easy enough to do.

An exposure through a red, green, or blue filter will normally be about 1/3 of the same exposure through a clear filter.  In other words, a 10-minute exposure through the clear filter would require a 30-minute exposure through the red filter for equivalent depth.  But, since the camera is binned 2x2 for the color images, the sensitivity has increased fourfold.  This means the equivalent exposure is 30/4 or 7.5 minutes, or 450 seconds.  In the interest of my short attention span, I rounded down and took 400-second exposures for the RGB images.

Since most of the detail will be obtained from the luminance exposure, it is not necessary to take so many individual frames for the RGB images.  3 each is sufficient to remove most of the noise from the images.  Any remaining noise can be eliminated during post-processing.

The total exposure sequence for the NGC253 image was then:

• 6 at 600 seconds Luminance

• 3 at 400 seconds Red

• 3 at 400 seconds Green

• 3 at 400 seconds Blue

• Total Exposure Time = 7200 seconds (2 hours)

After refocusing, the autoguider was restarted and the Red sequence was captured.  The autoguider was again stopped afterwards and the procedure repeated for the Green and Blue exposure sequences.

Dark Frames

The final step was to take dark frames.  Flat frames are also a common calibration image to take, but I have never used them.  The main reason is that they require either an observatory (which I don't have), being setup before dark (which never happens), or staying up until sunrise (which I don't do too often since I have to come into work and write this webpage).  Also, I've never looked at one of my images and said, "You know what this needs?  A flat field."  As long as you have darks, you are good to go.  If you want to take flats, they can't hurt.  And for imaging systems with a fair amount of vignetting, flats are highly recommended.  If, for example, I was using an STL-11000 on the C14, there would be some noticeable vignetting in the corners and flats would be helpful.  With the smaller chip of the ST-10XME, flats are less critical.

It is best to take a sequence of dark frames and combine them to create a master dark frame.  This will eliminate any extraneous noise.  More is better, as always.  If I had a permanent observatory and could roll the roof closed and go to bed, I would probably take ten dark frames.  Since I don't have an observatory, I limit myself to three frames in the interest of sleep.  Remember that you need darks for both the luminance exposure and the color exposures since differ in length and were taken at different resolutions.  For this image, three 600-second darks binned 1x1 and three 400-second darks binned 2x2 were captured using the Sequence mode of MaxIm DL (allowing me to go eat snack mix and put my greasy fingers all over my friends' telescopes).  This increases the total time spent imaging to 2 hours 50 minutes.

(An image of NGC7331 was captured the same night using the exact same exposure times, so the dark frames were used for two images.  This helps make 50 minutes of dark frames seem worth the trouble.)

Summary

At the end of the night, a total of 21 images (including dark frames) had been captured for NGC253.  Image all night, sleep all morning, process images all afternoon--sounds like a fun day to me!  Continue to the next section for details on how the images were processed.

Processing the Image

Return to Imaging NGC253 Main Page