Finding the Target

Finding and centering an object on a small CCD chip can be a tricky business if you do not have a telescope with a computerised high precision GOTO function like the Meade LX200.  Even with such a scope it is helpful to visually check that the GOTO successfully centered the object before taking the image.  What is needed is a device that allows you to look through the telescope while the camera is attached and focused.     A flip mirror finder is just such a device.  The finder is placed in the optical path of the OTA, directly in front of the CCD camera. The front of the finder attaches to either the telescope visual back, focal reducer or telenegative and the camera attaches to the rear of the finder.  The finder contains a small mirror which can be rotated from a horizontal to 45 degree position.  When the mirror is horizontal, light passes over it and into the camera.  When in the 45 degree position, light is reflected up through a standard eyepiece allowing you to see what is centered in the telescope FOV and thus on camera.  Finding and centering objects for the camera is now easy.  Just set the finder mirror to the 45 degree position and place the object in the finder eyepiece FOV using your normal methods (star hopping, finderscope, analogue setting circles, digital setting circles, GOTO etc..).  Now adjust the RA and DEC of the OTA to center the object in the finder eyepiece, move the mirror to the horizontal position and you are ready to take your image!  I have successfully used flip mirrors from Meade and True Technology

Taking the Image

Once you have focused your camera and centered an object there are two key things left to do to get that image.  The first is deciding how long an exposure to take and the second is to keep the object steady on the CCD Camera by guiding the telescope for the duration of the image.

Exposure Duration

With all exposures, the aim is not to expose the image to the point where any pixels become saturated.  This can be checked by looking at the histogram of the image at the end of the exposure.  If there are fully saturated pixels then reduce the exposure time.  You also need to check the histogram to ensure that you have used a long enough exposure to have captured an image with enough information for image processing to yield good results.  The histogram should start off at a high peak (which is the sky background noise) and then tail off in an exponential curve towards 0.   If you just have the high peak and little or no exponential curve then you need to increase your exposure duration.  Typical exposure times are:

•  Less than 1/1000th of a second for solar and lunar images

• A few tenths of a second for planetary images

• 5-10 mins for each sub exposure for deep sky objects.  This can seem contradictory as to see faint details around nebulae and galaxies you need a total exposure time extending into multiple hours.  However, it is impractical to take a single image that long as it could be ruined by a guide error of just a few seconds and some areas of the image would be over exposed (e.g. bright stars or the core of a galaxy).  The best technique is to take multiple exposures or subframes (say 24-36) of 5-10 min duration and then average them in image processing software.  This increases the signal to noise ratio by the square root of the number of subframes averaged.  The resulting image is now equivalent to a single exposure of multiple hours but avoids overexposing the bright areas and any guide errors will not result in all the image data being affected.

Guiding

Errors in polar alignment or in the RA tracking of the telescope drive can cause the object to shift position on the CCD chip during a an exposure lasting several minutes.  If these errors go uncorrected the result is a blurred image and star trails.

To correct the errors as they occur you first need to be able to see them happen and then have a means of altering the position of the OTA to compensate for the tracking error without touching the telescope.  Touching the telescope to operate the manual RA or DEC controls would introduce a lot of vibration in the image and ruin it.  Compensating for tracking errors in this way is known as guiding and there are two methods of doing it.

Manual Guiding

On the LX10 the handcontroller and DEC motor allow for small corrections in RA and DEC to be made without touching the telescope.   On the LX200 or RCX400 you can use its hand controller to make small corrections in RA and DEC without touching the telescope.  So now we have a method of correcting the errors, how do we see them happen?   For this we can use an off-axis guider (OAG) or a separate guidescope. 

An off-axis guider is a unit placed in the optical path of the OTA.  It connects to the visual back of the OTA or the focal reducer.  The flip mirror and CCD camera are then connected to the rear of the OAG.  The OAG has a small prism which intercepts part of the light around the edge of the field of view of the telescope.   The prism is small enough so that it does not intercept any of the light destined for the CCD camera.  The intercepted light is diverted into an eyepiece.  This allows the user to view a small portion of the sky at the same time as the camera is taking an image.   The eyepiece used with the OAG needs to have an illuminated scale and is known as a reticle.  To guide the telescope with the OAG you need to locate a star in the reticle eyepiece and watch its movement against the scale.  If it moves in RA (left/right) then adjust the speed of the RA drive with the handcontoller until the star returns to its original position.  If it moves in DEC (up down) then adjust the DEC of the OTA using the handcontroller.  Most of the corrections should be in RA as precise polar alignment should have eliminated nearly all the DEC tracking error.  Most OAGs will let you rotate the prism around the edge of the field of view of the imaging chip to make it easier to find a guidestar.  I have used the Celestron Radial Off-Axis Guider as it has several features which make it much easier to locate a star in the reticle. I have used the Celestron Illuminated Microguide Eyepiece as the reticle.

A guidescope is a separate telescope (usually a refractor) mounted on the main telescope OTA.  You use the reticle as the eyepiece for the guidescope.  As with the OAG, you locate a star in the reticle and correct any movement against  the scale using the hand controller. 

There are many pros and cons you should consider when deciding whether to use an OAG or guidescope for guiding.  These are discussed in depth in an article by Philip Perkins.   As a summary, the advantages of an OAG are that it does not suffer from flexure (movement of the guidescope against the the OTA) and it can correct for image shift (movement of the image due to movements in the OTA mirror as the telescope tracks across the sky).  The main disadvantage of the OAG is the difficulty in locating a guidestar in the small field of view afforded by the OAG prism.  The advantage of the guidescope is that it is relatively easy to find a bright guidestar and the disadvantages are the possibility of flexure and an inability to correct for image shift.

Autoguiding

With a telescope mount that has an autoguider port (such as the LX200, RCX400 and Takahashi EM-200 mount) it is possible to use a CCD camera to automatically make the guide corrections to the RA and DEC during an exposure.  Essentially, the CCD camera is put in the place of the reticle eyepiece on the OAG or guidescope. Using the camera control software the user will select the brightest star in its FOV as the guidestar and set the camera to take an image every few seconds.  The camera control software looks for movement in the guidestar from one image to the next and sends commands to the RA and DEC motors of the telescope to compensate for the movement.  Clearly the main advantage of this guiding technique is that you are freed up from the work of making the guide corrections yourself.  Manual guiding is very demanding physically and I was only ever able to take images of 20 minutes before I needed to rest.  Autoguiding lets you take as many images of as long a duration as you want.  The CCD camera never gets tired! The other advantage is the CCD camera can 'see' stars that are often too faint to be used for visual guiding.  Many of the SBIG cameras have two CCD chips, a main imaging chip and a smaller chip that can be used as the CCD camera for autoguiding.  As with everything there are pros and cons with this design.  Pros: You don't need two CCD cameras (one for taking the image and one for guiding).  You don't need a separate guidescope or need to use an OAG.  Cons:  If you are using color or narrowband filters then you will be imaging the guide star through the filter which means you need to select a brighter guide star and there may not be a bright enough star near your target.  Also, the guide chip is in a fixed position relative to the imaging chip and so the camera must be rotated and positioned such that a suitable guide star is on the guide chip.  This may mean the framing of the target in the imaging chip isn't exactly how you would like it.  I have always found that the pros of this design out way the cons, especially as you can use the guide chip in the SBIG cameras to control adaptive optics for guiding (see following discussion).

Autoguiding an LX200 or RCX400 can produce good results (round sharp stars) at lower focal lengths (e.g. 1000mm or below).  However, the accuracy of the mount on these telescopes even after PEC training does not yield very good results a the full focal length of the scope (2000mm or above).  This is because the mount position can only be corrected every few seconds at best and any error in the tracking of the mount during the correction periods shows up in the images taken at longer focal lengths as large of elongated stars.  It is not just the stars that are affected though.  The fine detail in the object being imaged is also lost as it is smeared out over the CCD chip during the exposure.  A very good solution to this problem is to use the AO-8 adaptive optics system from SBIG.  This unit attaches between the CCD camera and the telescope.  It has a lens which control where the light from the telescope falls on the CCD chip.  Very short exposures of a guide star (usually around 10 images per second) are taken using the guide chip in the camera and the AO-8 tilts its lens between each exposure to compensate for any movement caused by poor tracking or atmospheric disturbance.  If the guide star drifts too far over time for the lens tilt to compensate, the AO-8 sends a pulse to the mount to adjust the guide star position back towards the centre of the lens.  The frequent and very accurate correction of the tracking that is possible with the AO-8 means inaccuracies in the tracking of mount are corrected as well as some atmospheric disturbance effects.  You can see the difference between autoguiding with adaptive optics and just the guide chip in these images of M1.

Dithering

Once you have taken multiple 10 minute subframes of your target they will need to be combined into one long duration exposure using image processing software.  The normal way of doing that is to first register the images so that all the stars and the target are aligned in exactly the same pixel locations.  Then you average all the images to get a single image with a much higher signal to noise ratio than is in any one of the subframes.  In general the signal to noise ratio increases as the square root of the number of subframes averaged.  However, if we just average each pixel location using pixels from each of the subframes we are missing an important opportunity to remove artifacts from the image.  An artifact could be a satellite trail or a hot/cold pixel in the CCD camera.  Most image processing software will let you average only those subframe pixels that fall a within a certain range of the average value (the range is known as standard deviations from the mean or average).  Using this combine method will avoid including outlier pixel values like a satellite trail in the average.  Those pixels will be rejected allowing all the other good pixels in the image with the trail to be used.  However, this technique won't reject hot or cold pixels from the image unless we also use a technique called dithering.  Dithering simply means that the software controlling the CCD guide chip will randomly move the mount a few pixels up, down, left or right between each image.  That means that each image falls on a different set of pixels in the chip.  Once the images are registered all the stars and the target are aligned but the hot/cold pixels in each image will now be in slightly different locations due to the dithering.  Using the standard deviation rejection technique will now remove these hot/cold pixels from the average.

Calibration Frames

As well as taking images of the target you will need to take special images that characterize dark current, variations in the CCD chip sensitivity and dust in the optical system.  These images can then be used during image processing to correct for these artefacts.

Dark frames

Image exposures greater than a few seconds will be affected by noise, know as dark current, generated within the CCD camera itself.  To remove this noise from your final image you need to take what is called a dark frame. This is an image made at the same ambient temperature and exposure duration of the original image but with the telescope aperture covered or camera shutter closed.  You then subtract this dark frame from your image as part of the calibration step in image processing and the dark current noise will be removed.  A flip mirror finder is useful for taking dark frames as you can achieve the same effect as covering the telescope aperture or closing the camera shutter simply by rotating the finder mirror to the 45 degree position.   With the mirror in this position all light is blocked from entering the camera.

For cameras that can't be cooled to a set temperature you will need to take dark frames during each imaging session so they are taken as close as possible to the temperature the images were taken at.  Cameras like the SBIG ST range can be cooled to a set temperature and that allows you to take dark frames at any time which match the temperature the images were taken at.  You can build a library of dark frames for different temperatures, durations and binning that can be used to calibrate images across many sessions.  The library  should be refreshed this library at regular intervals (e.g. every 6 months) as the camera behavior can change over time.

To reduce the noise in your dark frame I recommend taking several dark images at the same temperature, duration and binning and then average them to create a single Master Dark you can use to calibrate your images.  I take 9 dark frames and average them to create a Master Dark.

Flat Fields

These are images taken against a field of even illumination (e.g. the evening or morning sky when it is light enough not to show stars but dark enough not to saturate the pixels quickly). These images can be used  to correct for variations in sensitivity across the CCD chip or for specs of dust on the chip, filters or optics that can appear as dark doughnut shaped shadows on the image after processing. 

 It is not always possible or easy to take a flat field of the evening or morning sky during an imaging session.  If that is the case it is best to use an artificial light source.  The easiest way to do this is build yourself a light box.   This is a device which fits over the end of your telescope and uses a standard lightbulb to provide the even field of illumination required for taking flat fields.  The advantage of using a lightbox is that you can take the flat fields at any time during the imaging session and you do not have to worry about stars appearing in the image and ruining your flat. I constructed a lightbox for the LX200 and your can see the plans here.

On my Takhashi TOA-130 I use a Flip Flat from Alnitak Astrosystems.  This is a circular paddle that contains an evenly illuminated light panel which covers the OTA opening in place of the dust cover.  It is motorized and so can be opened for imaging and closed for flat fielding under the control of automated imaging software (see below).

You will need to take a set of flat field frames during each imaging session.  A flat will need to be taken through each filter and at each camera angle you used.  To reduce the noise in your flat field frames I recommend taking several images through each filter and then averaging them to create a single Master Flat. I take 4 flat frames through each filter and then average them to create a Master Flat.

Automated Imaging

All the techniques described thus far can be put together in a manual imaging sequence:

1. Focus the telescope

2. Find the object

3. Program the camera control software to take a series of subframes

4. Refocus the telescope every hour to compensate for focus shifts due to temperature changes.

5. Take dark and flat field frames

What would be really great is if we could have software do all these steps for us so it can run all night and we can get some sleep!  Well there are several software packages available that will let us do just that.  I use CCD Autopilot from CCDware.  Here's how you can use CCD autopilot to automate your imaging session:

1. Point the telescope at a well known star using the Telrad.  Establish a link to the telescope in TheSky and sync on the star you pointed the telescope at.   This tells TheSky roughly where the telescope is pointed.

2. Use the field of view indicators in TheSky to select your target and camera position angle. TheSky can show you the actual field of view that will be captured by your imaging and guide chips .  This lets you select the camera position and angle to best frame the object and have a suitable guide star on the guide chip. Alternatively you could use CCD Navigator from CCDWare to select a target and camera angle which can then be saved and imported into Autopilot.

3. Import the target position and camera angle into Autopilot from the TheSky or one saved by CCD Navigator

4. Tell Autopilot how you want to image the target (e.g. number and duration of images to be taken through each filter), how often you want to focus etc... and then start the imaging session

5. Autopilot will now take a short image and  pass it through plate solving software to find the exact position of the telescope.  It will then select a close by star to focus on and move the telescope to that position.  It will then instruct FocusMax to autofocus on that star.

6. Once focused the telescope will be moved to the target and the camera rotated to the correct angle (if a rotator is installed).  It is likely that the telescope won't have moved to exactly the right position due to polar alignment inaccuracies.  So once at the target another short image will be taken and plate solved to find out precisely where the telescope is pointing.  The telescope position will then be corrected as necessary. 

7. Autopilot will now control the imaging and autoguiding using the camera's control software.

8. If you tell Autopilot to focus at regular intervals during the imaging session it will stop imaging at the specified time and move the telescope to a suitable focus star and autofocus again using FocusMax.  It will then move back to the target and resume the imaging sequence.

9. Once all the required images have been taken Autopilot can take dark frames and flat fields as required

This lets you image all night without any involvement.  It's the key to taking really deep exposures that may total 10 or more hours.