Return to Previous Section

4. Strategy for Obtaining Lamp-Flat Frames


The tests described above indicate that lamp-flats obtained at detector temperatures within +/-10 degrees C of the data frames should be adequate for determining the proper pixel-to-pixel sensitivity map. However, lamp-flats should be obtained on every observing night, and at detector temperatures as close to those of the actual data frames as possible. A minimum of three lamp-flats is needed for good statistics. Observers are encouraged to obtain as many lamp-flats as is reasonable, given time constraints. It is best if the total number of frames is an odd number.

The exposure time should be at least a few seconds, and the maximum number of counts in the frames should be around the linearity threshold (20,000 ADU or so). This requires adjustment of both the integration time and the LED gain. On nights when the ambient temperature is above 0 degrees F, one can probably determine a good exposure time and then use it for all lamp-flats obtained that night. On nights with ambient temperature below 0 degrees F, the battery signal degrades rapidly with use, and higher gain or longer integration time will be required as the lamp-flat series goes on.

Observers should also make sure they obtain bias frames at the same detector temperature as the lamp-flats. The total number of such frames should be odd, and there should be more of them than there are lamp-flats. Dark frames are also needed, mainly to correct for hot pixels. Again, one wants dark frames at the same detector temperature as the lamp-flats, and one should have at least three of them. If one is taking special darks for this purpose, they can have the same integration time as the lamp-flats. It is, of course, most simple to obtain the lamp-flats at the same detector temperature as the bias and dark frames that one must obtain for the reduction of one's science frames. In this case, one should scale the dark correction frame to match the exposure time of the lamp-flats.

After the lamp-flats have been bias- and dark-corrected, they should be cropped to (1:1530,21:660), as will be discussed in Section 4.2, below. Then the frames should be median-combined and normalized. The combined frame should then be processed with a low-pass median filter. The frame should then be divided by the filtered frame to produce a final pixel-to-pixel sensitivity map, ready to be combined with the illumination-correction frame.


Return to Main Lamp-Flat Page

Return to Main Flat Page

Return to Outline


Updated: 2009 August 20 [pbe]