Quality Control and
Data Processing

HC PLOTS
Bad pixels
CONAD
QC1 database (advanced users): browse | plot

With LED flat-fields, the spectrograph is not within the lightpath and the detector is completely illuminated. LED FFs are used to determine the number of bad pixels and the detector gain (conversion factor CONAD). They are measured on a regular basis for all read-out modes and detector binnings (1x1, 2x1, 4x2, and 8x4). A complete set consists of flat-fields with at least three different exposure times and at least five frames per exposure time.

Raw LED FF. Example frame for the blue detector. The gaps are due to the pre-/overscan of the 16 read-out ports

QC1_parameters

FITS key	QC1 database: table, name	definition	class*	HC_plot**	more docu
none	espresso_ledff..badpix_nb_tot	total number of bad pixels per detector	HC		[docuSys coming]
none	espresso_ledff..badpix_nb_rms	RMS bad pixel number across read-out ports	HC		[docuSys coming]
*Class: KPI - instrument performance; HC - instrument health; CAL - calibration quality; ENG - engineering parameter **There might be more than one.

ESPRESSO has two chips, one for the BLUE and one for the RED arm. Each chip has 16 read-out ports. LED-FF QC parameters are determined by the pipeline per read-out port which leads to 16 different values for each QC parameter per chip. They are written as header keys in the form of (for the bad pixel number):

QC.EXT<n>.ROX<x>.ROY<y>.BADPIX.NB

with <n> being the detector (0 for BLUE and 1 for RED), <x> the x position of the read-out on the chip (0, 1, to 7), and <Y> the y position (0 or 1). The corresponding names in the QC1 database follow the scheme

badpix_nb_<x>_<y> .

In addition to these individual values, the total number, the average, and the RMS per chip are calculated and written into the QC1 database.

Trending

The number of bad pixels is trended separately for the four different read-out modes: 1x1 binning (fast read-out), and 2x1, 4x2, and 8x4 binning (all slow read-out).

Scoring&thresholds Bad pixels

The total number of bad pixels is scored per detector. Thresholds have been set according to the measured values during commissioning.

History

None.

Algorithm Bad pixels

Bad pixels are detected based on their non-linear behaviour when dividing frames with different exposure times. The pipeline recipe determines the QC parameters per read-out port. The QC procedures calculate in addition the total number, the average, and the RMS of all 16 ports per detector.

QC1_parameters

FITS key	QC1 database: table, name	definition	class*	HC_plot**	more docu
none	espresso_ledff..gain_avg	average gain across read-out ports [e-/ADU]	HC		[docuSys coming]
none	espresso_ledff..gain_rms	RMS of gain numbers across read-out ports [e-/ADU]	HC		[docuSys coming]
*Class: KPI - instrument performance; HC - instrument health; CAL - calibration quality; ENG - engineering parameter **There might be more than one.

ESPRESSO has two chips, one for the BLUE and one for the RED arm. Each chip has 16 read-out ports. LED-FF QC parameters are determined by the pipeline per read-out port which leads to 16 different values for each QC parameter per chip. They are written as header keys in the form of (for the CONAD):

QC.EXT<n>.ROX<x>.ROY<y>.CONAD

with <n> being the detector (0 for BLUE and 1 for RED), <x> the x position of the read-out on the chip (0, 1, to 7), and <Y> the y position (0 or 1). The corresponding names in the QC1 database follow the scheme

gain_<x>_<y> .

In addition to these individual values, the average and the RMS per chip are calculated and written into the QC1 database.

Trending

CONAD is trended separately for the four different read-out modes: 1x1 binning (fast read-out), and 2x1, 4x2, and 8x4 binning (all slow read-out).

Scoring&thresholds Conversion factor

Scoring thresholds have been set close to the nominal value of 1.1 e-/ADU.

History

First pipeline versions used "GAIN" instead of "CONAD" for the names of the header keywords. The parameters in the QC1 database still have "gain" in their names; this will be changed eventually.

Algorithm Conversion factor

CONAD gives the conversion from ADU into electrons in e-/ADU.

The mean and difference frames of two exposures with the same exposure time are computed. Then, the standard deviation of the difference frame is calculated. CONAD is measured from the relation between mean flux level and standard deviation.