Back

Audine with Ethernaude = characterization of the offset frames

---

I take 50 offsets (with closed shutter, 0 sec exposure) with Prism v7. The CCD camera is stabilized at -3.2 °C. I will now analyse them with a Pyfits/Python script.

Read-out noise is 5.20 ADU ; the mean level is 2719.1 ADU.

First offset picture


Histogram of first offset picture

A close inspection shows that the histograms of all the offset pictures are exactly similar. In particular, the observed "hole" in the histogram for the pixel value 2719 appears in the same way on all histograms ! I think this has nothing to do with read-out noise, but most likely it is an artefact in the transfer curve of the ADC, such that the bin 2719 is under-represented for an input signal covering continuously the range 2700-2740.



Average of the 50 offset pictures
Some diagonal lines do appear, but they are very faint and probably play no role. The faint vertical lines are due to hot pixels, which contaminates the other pixel that must pass through them for the reading. This effect appears clearly with the LPT1 interface, due to the long read-out time.


Picture of the sigma values found pixel by pixel on the 50 offset frames
The sigma values are absolutely uniform, no periodic noise (as with the LPT1 interface) is observed.



Here is the Pyfits/Python script that was used for the analysis: