[ ESO ]
VIMOS Quality Control:
screen flats

CAL | HC | refs | QC
Trending & QC1
Data Package
Data Management
QC links:
lamp performance | CONAD | noise
lamp performance
noise (fpn, photon)
QC1 database (advanced users): browse | plot
   Click on CURRENT to see the current trending (Health Check).
   Click on HISTORY to see the historical evolution of the trending.

Screen flats are taken in imaging mode (low gain, 1x1 binning, 2148x2440 window) with a flat-field lamp: the QTH10 is used for VRIz filters, QTH50 for screen flats with U and B.

top Lamp performance

This is mainly an indicator of the flat-field lamp efficiency, but contains also information on the optical components and CCDs. Any lamp flux anomaly should be detected simultaneously in the four quadrants. An efficiency variation in a single quadrant and filter should be related to the filter, while a variation in all filters of a quadrant may be related to the CCD. The efficiency is measured in ADU/sec.

QC1 parameters

parameter QC1 database: table, name procedure
lamp efficiency vimos_scrflat, efficiency median signal divided by exposure time


QC1 parameters

parameter QC1 database: table, name procedure
CONAD vimos_scrflat, conad_avg conversion factor of ADU into electrons (e-/ADU). The difference frame of the frist two raw frames is computed. Then, CONAD is measured in 16x18 windows of 100x100 pixels (in the central 1600x1800 region) by deviding the median signal level from the first raw frame by the variance in the difference frame scaled by 2. The final CONAD is the average of the 100 individual measurement.


July 2010: new detectors
since December 2010: first two flat fields of the exposure sequence often differ by about 500-1000 ADU
2011-03-01: video board of FIERA A exchanged; gain changed for Q2 and Q3
2011-04-18: gain values for all quadrants adjusted
2011-06-01: video board of FIERA A exchanged again

top Noise

QC1 parameters

parameter QC1 database: table, name procedure
photon noise vimos_scrflat, photon_noise The standard deviation of the 1600x1800 central pixels of the difference of the first two raw screen flats is computed and scaled by sqrt(2).
fixed-pattern noise vimos_scrflat, fixp_noise The difference between the 1600x1800 central pixels of the first frame and the same region shifted by 10x10 pixels in the second frame is computed. The standard deviation of the signal is the combination of fixed pattern and photon noise (scaled by sqrt 2). The photon noise component is quadratically subtracted.


Screen flats taken with different filters vary in average level. For trending, noise parameters are plotted against average level. The relation between fixed-pattern noise and level should be linear. There are, however, non-linear deviations between spectral bands: flats taken with the I band filter have remarkably less fixed-pattern noise. Photon noise should depend on the square root of the level.