mirror sites: PL (internal link) HQ     [?]

Quality Control and
Data Processing

CAL | HC | refs | QC
Trending & QC1
   lamp flat
QC links:
SINFONI: lamp flat
Lamp Efficiency
number of bad pixels
Lamp OFF counts
Lamp Fix pattern noise (region#1)
Lamp Fix pattern noise (region#2)
Vignetting Check (slit#1)
Vignetting Check (slit#32)
QC1 database (advanced users): browse | plot

Flat frames are used to monitor the lamp efficiency in each of the available gratings/filters (J, H, K, and H+K) and pixel scales (25, 100, and 250 mas), the number of bad pixels for each grating/filter, the number of counts in the lamp-off frames, as well as the fixed-pattern-noise. For data processing flat frames are used to correct the pixel-to-pixel variations of the detector. Flat frames are routinely obtained during day-time calibrations during SINFONI operations. They are usually obtained in sets of 5 lamp-on and 5 lamp-off frames, and generally mirror the gratings and pixel scales used during the night's science observations. The SINFONI pipeline processes the lamp-on minus lamp-off frames into a master lamp flat and a bad pixel map.

A SINFONI master flat as computed from 5 (lamp-on minus lamp-off) input raw frames (here the master flat is for the J-band and 25 mas pixel scale). The 32 slitlets are evident in this master flat.
A SINFONI master bad pixel map. This map is composed of hot and non-linear pixels. A value of 0 indicates the hot pixels; otherwise 1..

~ ~

Parameters trended


FITS key QC1 database: table, name definition class* HC_plot** more docu
QC.SPECFLAT.NCNTSAVG sinfoni_flat..qc_specflux_cavg average counts HC [docuSys coming]
QC.SPECFLAT.OFFFLUX sinfoni_flat..qc_specflux_coff off lamp counts HC [docuSys coming]
QC.LFLAT.FPN1 sinfoni_flat..qc_lflat_fpn1 fpn region #1HC [docuSys coming]
QC.LFLAT.FPN2 sinfoni_flat..qc_lflat_fpn2 fpn region #2HC [docuSys coming]
QC.BP-MAP.NBADPIX sinfoni_flat..qc_bpm_nbadpix number of hot pixelHC [docuSys coming]
[calculated by QC procedure] sinfoni_flat..qc_rf1 relative flux in slitlet 1 HC [docuSys coming]
[calculated by QC procedure] sinfoni_flat..qc_rf32 relative flux in slitlet 32 HC [docuSys coming]
QC.SPECFLAT.NCNTSSTD sinfoni_flat..qc_specflux_cstd scatter of counts CAL [docuSys coming]
QC.BP-MAP.METHOD sinfoni_flat..qc_bpm_method bad-pix method CAL [docuSys coming]
*Class: KPI - instrument performance; HC - instrument health; CAL - calibration quality; ENG - engineering parameter
**There might be more than one.


The QC1 plots for the SINFONI FLATs are divided into 6 html pages, each of which shows the trending for each of the grating/filter configurations (J, H, K, and H+K) and for each of the pixel scales (25, 100, and 250 mas). The QC1 FLAT trending pages are:
QC report for FLAT frames.

  • UL: the mean collapsed columns of the current master flat (blue) and the single columns at pixels 500, 1000, and 1500 (dark grey, medium grey, and light grey, respectively). The reference master flat mean column is shown in red, and the difference between the reference and current master flats is given in black. Ideally, the difference should be near the zero line.
  • UC: as above, but with the mean collapsed rows of the current master flat and the reference master flat.
  • UR: a logarithmic histogram of the current master flat (blue), the reference master flat (red), and the difference between the two (black). The difference histogram should be approximately symmetric about the zero axis.
  • LL: an image of the current master flat.
  • LC: an image of the difference between the current master flat and the reference master flat.
  • LR: logarithmic histograms of the difference between adjacent raw, lamp-on an lamp-off flat frames. The first raw lamp-on flat minus the second raw lamp-on flat is shown in purple, while the lamp-off first raw flat minus the second lamp-off raw flat is show in turquois. The quality control relevance of this plot lies in the fact that the former histogram (purple) is dominated by photon noise, while the latter (turquois) is dominated by read-out noise.

    Scoring&thresholds Parameters trended

    Lamp counts should be set such that they are above 4000 counts (for adequate S/N in the flats), and below about 10, 000 counts (to avoid any persistence effects). These limits are shown on the median lamp count trending plots as dotted lines.


    SINFONI FLATs have been very stable over the current lifetime of the instrument. This can best be seen the the full history FLAT trending plots shown here: FULL FLAT history.

    Algorithm Parameters trended

    median lamp counts: The halogen lamp flux depends on the intrinsic brightness of the lamp, as well as the chosen gratings and optical elements. The lamp-off frame is subtracted from the lamp-on frame, and the median value of the level in these 5 frames is monitored.

    number of bad pixels: The number of pixels in the master lamp flat having an intensity greater than a pipeline-defined threshold (default threshold is 10s). median lamp-off counts sinfoni_flat: qc.specflux.coff

    The lamp-off flux monitors the light and heat contamination within the optical path. The lamp-off flux levels are usually a few counts above the reset anomaly (dark) at the same DIT. This is due to the fact that the dark frames are obtained using two excluding narrow-band filters, whereas the off-lamp frames are made through a broad-band filter. The monitored flux is the median level of the 5 lamp-off frames.

    fixed pattern noise region 1: The fixed pattern noise is computed by fitting a Gaussian to the histogram to the master flat. The FPN is simply the standard deviation (s) of this Gaussian fit. The first FPN is computed over the central quarter of the detector [512, 512: 1536, 1536].

    vignetting check: Slitlets #1 and #32 correspond to the upper-most and the lower-most slitlets in the field-of-view. As such, these slitlets can be subject to vignetting to the extent that their overall count rates can be a few 10% lower than for other slitlets. Following image reconstruction, these slitlets are located on the central part of the detector. A possible vignetting is monitored by comparing the relative count rates (i.e. the median counts in the slitlet divided by the median count rate of the entire frame) in slitlets #1 and #32 of the master flat frame. This QC1 parameter is most relevant following an instrument intervention. ~ ~

  • Send comments to <qc_sinfoni@eso.org> powered by QC [webCMS v1.0.1]