Linearity

The plot below shows the ADU per second and per pixel (normalized) as a function of the median value of counts per pixel in the detector. Non-linearity is ~2% over the range 0-12000 ADU (in the post-pre). This implies an uncertainty of about 0.02mag in your data if you are observing within that range of counts.

Currently no linearity correction is applied to the INGRID data. Therefore, select carefully your exposure time so that the level of counts of the sky and your sources is below ~12000 ADU.

NOTE: Each point in the figure above was obtained by averaging a set of seven dome flats, obtained via the 'multrun' command, each set having a fixed exposure time. Two series of such sets were taken so that the exposure time of each multrun was increasing through the first, decreasing through the second. Moreover, the two series were mixed up (1s,16s,3s,14s ... 4s,13s,2s,15s) in order to average out any possible influences of the lamp's intensity variations on the global trend of the curve. However, each single data point could still be affected by these variations, and the peaks in the normalized flux at low counts might be due to this.
The linearity has not changed since November 2001 (last measurement done in November 2005).