Checking the linearity of the CCD

The data reduction process will require the linearity of the pixels of the CCD. This means that the outreaded pixel intensity of a pixel is proportional to the number of photons in the pixel is implemented. The curve obtained under the same lighting conditions, when plotting the average of all pixels of the CCD to the exposure time must be a straight line. Is the curve shallowed with increasing exposure time the exposure is beyond the linear range of the counts of the CCD.

The linear range can be easily checked experimentally. The CCD camera is installed in a darkened room without optics. Ensure that each pixel receives the same (constant) amount of light. It is advantageous to install a tube in front of the camera (the tube should have a black inner surface), mounted at the other end of a white, but translucent paper as a diffuser is (you can of course take a white sanded glass or similar). With the aperture and the length of the tube you can simulate the proper focal ratio of the starlight beam from the used telescope.

Then you take a number of shots at different exposure times. It is not performed the usual dark correction of the frames. The exposure times chosen so that is the saturation of pixels between 0 and 100% of its capacity.

The following is an example. It is the Fischer-CCD camera, which in the campaign 140-WR spectroscopy was be used. The straight line extrapolates to zero exposure resulting the bias (1225). It is obvious that virtually the entire range of counts (ADU) from 0 to 65000 (16bit) is linear and thus suitable for quantitative measurements to be evaluated. The slope of the line (1410.4 / s) is of course dependent on the light flux that prevailed during the recordings (and therefore not transferable to other recording series).