Tuesday, 5 August 2014

Master Dark and Bias Frames

In order to get a good image back from a CCD on the back of a telescope, there are a number of processes you have to put the image through to account for discrepancies and errors. These include applying bias and dark frames as well as flat fielding. We handle these for you automatically, but for those that like to do things by hand we are now producing master dark and bias frames for more advanced users to use.
If you're still reading then I'm assuming you're someone who already knows what bias and dark frames are so I won't bother explaining the theory in this post. Maybe I can do another post if there is interest.

The telescope now uses time at the start of the night while the telescope is cooling to take some dark frames.It also takes some more throughout the night, roughly once an hour. This takes about 1 minute each time.

The frames are captured and calculated as follows:

  1. The telescope captures a 20 second and 40 second exposure with the shutter closed
  2. Many 20s and 40s exposures are median'd together to get a good representative 20s and 40s exposure
  3. The bias frames is obtained by performing the calculation: (2 * 20_median - 40_median)
  4. The dark frame is obtained by performing the calculation: (20_median + 40_median - 2 * bias)
(The latter calculation being equivalently 3 * (40_median - 20_median))

This gives a bias frame for the readout of a theoretical zero-length exposure and dark frame containing the dark current for 60 seconds of exposure, with readout noise removed. A frame to subtract from an image of exposure time T minutes can then be calculated in the usual way:

D(T) = bias + T*dark

Of course, this must be applied to an image not ordered with automatic dark frames and downloaded without flat-field applied. (So you'll also need to apply a flat-field yourself).

So far (only a few days in), around 100 frames have been gathered per exposure time per camera. The results are reasonable. They do show that the CCD on galaxy camera seems to have a much poorer level of dark current than either of the other two cameras. This could be because light is leaking in somewhere, although the illumination would have to be suspiciously uniform. If it is the camera, then it will be swapped at the next servicing trip.

The final files can be downloaded on this page

I will add newer ones as we gather more exposures and build better calibration frames. If you'd like to be notified when new frames become available, email me or leave a comment.

The eventual goal is to have these downloadable from the website proper, and possibly to do away with the current instantaneous darks in favour of file darks.

If anyone with greater expertise than I has any suggestions for how this process might be improved, please email me.