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

Quality Control and
Data Processing

CAL | HC | refs | QC
Trending & QC1
   fibre stability
QC links:
GIRAFFE: Fibre system stability
Fibre localization and availability | Fibre signal width | Relative fibre efficiency

Fibre availability:
Fibre signal width:
Relative fibre efficiency:
QC1 database (advanced users): browse | plot

The fibre system of FLAMES contains five slits for GIRAFFE and one for UVES-FIBRE. The GIRFFE fibre system has the two slits Medusa1 and Medusa2 with 135 freely allocatable fibres each, the IFU1 and IFU2 slits with 320 fibres each, organized in 15 subgroups of little deployable IFUs, and the Argus slit with a single IFU of again 320 fibres. In total, there are 1230 Giraffe fibres.

The calibration of fibre properties is done through the fibre flats. The fibre flats are obtained by a robotic calibration lamp which illuminates the fibres when they are configured in a spiral pattern. While this gives good results for the Medusa and IFU fibre systems, the Argus fibre set is not properly illuminated and is therefore calibrated in addition with nasmyth screen flats.

While all science setups taken during the night receive during the following day a set of at least three such fibre flats, there is also a set of dedicated health check (HC) flats which are taken at regular intervals and used to monitor the fibre performance. They always use the L543.1 setting. The Medusa HC flats are taken every 3rd day. The IFU and Argus HC flats have been added in 2006 and are taken once a week. These HC flats always come as a single raw file.

These data are quality-controlled and trended on the Health Check monitor to check

  • the fibre availability (to detect broken or unallocated fibres),
  • the quality of the pipeline localization solution (number of fibres found),
  • the fibre-to-fibre efficiency (stability of fibre transmission),
  • and the quality of illumination by the calibration lamp.

Find the performance monitoring of the flat calibration lamp here.

Fibre localization and availability
Fibre localization and availability | Fibre signal width | Relative fibre efficiency

The fibre availability plot monitors the number of illuminated fibres in the Health check setup ("Health") and, for comparison, also in all setups ("all"). This number is derived from the localization product of the pipeline. It is both a HC parameter (fibre availability) and a CAL quality parameter (if issues occur with the localization, this number is often signficantly lower than expected).


FITS key QC1 database: table, name definition class* HC_plot** more docu
[NPIX in localization product] giraffe_flat..num_fib number of fibres extracted by the pipeline KPI
*Class: KPI - instrument performance; HC - instrument health; CAL - calibration quality; ENG - engineering parameter
**There might be more than one, always check the link 'HC'.

For the robotic flats, giraffe_flat..flattype is always ROB.


The trending plot follows the number of fibres found by the pipeline. The "Health" plot has the HC data only, the "all" plot has all measured data (HC and calibration plan). Usually the "all" plot will have the same, or almost the same number of fibres and just show a denser time coverage. The signal from the last few fibres usually does not fall completely onto the CCD, and is truncated by the pipeline. Depending on wavelength, the index of the last complete fibre may vary.

Another reason for changes of this number may be a broken (or fixed) fibre.

The plot is particularly important for the Medusa fibre systems since they point the fibres to individual targets. A missing fibre means a missing target. For the IFU and the Argus system, a missing fibre results in a missing spatial pixel.

Number of fibres located in Medusa1 HC files.


This overview does not cover years 2003-2006.

Date event
2007-04-28 fibres 27, 79 no signal
2007-04-29 fibres 27, 84 no signal
2007-04-30 - 2007-05-14 fibres 27, 60, 84 no signal
since 2007-05-15 fibre 27 no signal
2011-09-04...2011-09-20 fibre 4 no signal
2006-09-20 fibre 109 no signal
2006-12-19 fibres 58, 109 no signal
2009-01-16 ... 01-20 same, plus fibre 107; #107 back again afterwards
2010-10-21 FPS58 enabled again
2011-04-22 FPS13 no signal
before 2008-02-19 fibres 192, 278 no signal
2008-02-19 fibres 192, 229,278 no signal
2010-07-20...2010-08-03 fibres FPS 86 and 107 missing
general fibres 102, 393 no signal
2010-06-03...2010-06-04 2 fibres disabled, PPRS34647
all fibres visible  

Check also out here.

Scoring&thresholds Fibre localization and availability

The scoring is tight. It is configured asymmetrically around the current value, plus/minus 1 or 2, in order to be sensitive in particular for broken fibres suddenly appearing.

Algorithm Fibre localization and availability

NPIX in localization product

Fibre signal width
Fibre localization and availability | Fibre signal width | Relative fibre efficiency

The average width of the fitted fibre signal in cross-dispersion direction is monitored in the fibre signal width plots. There are two of them, for the Health check setups ("Health") and for all setups ("all FLATs").

[click to enlarge] Cross-cut through a bundle of fibres to display the typical slope of the fibre signal in cross-dispersion direction. The pipeline detects and fits the individual fibre signals, their average width is monitored as mean_width.


FITS key QC1 database: table, name definition class* HC_plot** more docu
[NPIX in localization product] giraffe_flat..mean_width mean width of extracted signal (in pixels) CAL
*Class: KPI - instrument performance; HC - instrument health; CAL - calibration quality; ENG - engineering parameter
**There might be more than one, always check the link 'HC'.


The trending plot displays the average width of the fibre signal in pixels. This parameter describes the stability of the spectral format, in particular it is is sensitive to any failure of the pipeline solution. The "Health" plot has the HC data only (L543.1 monitoring data). Here the mean_width parameter should be very stable. The "all FLATs" plot has all measured data (HC and calibration plan). Their mean width (spectral format) depends on wavelength and therefore is less sensitive for trending.

Mean width of fibre signals, for the Medusa1 and Medusa2 HC data (L543.1, plot labelled "HC"). They are scored against the thresholds plotted as dotted lines.
Same, with all data from the respective slits (plot labelled "all FLATs"), and the HC data marked in the same color as above. They are scored against the thresholds plotted as dotted lines. These thresholds are relaxed in comparison to the HC plot.


2019-09-12: Thresholds changed for Medusa1/2 HC plots; was: 4.2/5.5; is: 4.6/5.8; parameter drifted slowly towards higher values, slow change of spectral format.

Scoring&thresholds Fibre signal width

The values are very stable. Sudden jumps are likely related to pipeline issues rather than indicative of instrument (fibre) issues. Slow drifts have been observed for Medusa1 and 2.

The entire range of possible (good) values of mean_width per fibre system) is best visible when the entire set of settings is taken once a year for the purpose of populating the calibration database. Find an example here.

Algorithm Fibre signal width

Averaged width is the mean of all fitted fibres (FWHM).

Relative fibre efficiency
Fibre localization and availability | Fibre signal width | Relative fibre efficiency

This is historical content.

The relative fibre efficiency has been monitored until 2020-03. It was switched off at the VLT shutdown and was not re-enabled afterwards, mainly due to maintenance difficulties. The historical content has been preserved, though. It is documented in the following.

The fibre efficiency is monitored with data from the pipeline product with pro.catg "FF_EXTSPECTRA". This has a binary table extension with columns, FPS and TRANSMISSION (check out here for a description of the pipeline-produced binary tables). The FPS is the fibre position index, a number which identifies uniquely each fibre per slit. TRANSMISSION is a number obtained by collapsing the chromatic efficiency per fibre and normalizing it to the brightest fibre (SIMCAL fibres are always set to one). The transmission numbers mostly are between 0.6 and 1.0.

The trending plot is a non-standard report that does not follow a QC1 parameter but displays all fibre efficiencies in one plot.


The fibre efficiency numbers are not available as QC1 parameters. If needed they can be downloaded from the trending plots.


There is one trending plot per fibre system. All extracted fibre efficiencies are normalized to the brightest fibre and diplayed as a vector "all mean efficiencies". Their rms differences are calculated and displayed in box 2 ("RMS").

Some care has to be taken with the interpretation of the RMS results. There might be differences induced by illumination changes (like in the example below). They can be recognized by coherent changes of a group of fibres. Then, there is real scatter since certain fibres show a larger instability than others.

Normalized fibre efficiency.

Top:all values plotted on top of each other
Bottom: related RMS values

The red lines mark the SIMCAL fibres which are always set to one. The efficiencies are normalized to the brightest (non-SIMCAL) fibre. Zero efficiency means fibre is broken.

The RMS plot color-codes the values below 3% green, the ones above red.

A second plot shows the same data, as 2D display. This plot is powerful to detect patterns, like the illumination defect for the first ~10 fibres for some time. Such effects corrupt the RMS plot in the sense that those fibres are not intrinsically unstable but have been illuminated improperly.

Normalized fibre efficiency, 2D plot.

All efficiency vectors are stacked, time increases vertically (tick marks for every 5th vector). The X axis is fibre index FPS, with tick marks for the SIMCAL fibres.

The trending pages include links to numerical data:

  • "details per fibre" with the mean numbers per fibre, including a flag marking SIMCAL, SKY fibres, as well as those with rms >1% and >3%, and those without a signal
  • "measurements" with a list of all entries (date of acquisition, file name, information about missing fibres).

Here is a list of unstable fibres for the Medusa slits (if possible these ones should be avoided for sky measurements):

Slit FPS of unstable fibres (>3% RMS)
Medusa1 26, 31, 56
Medusa2 101

Most fibres are stable at a level much better than 1%.

The IFU slits show some bundles of fibres with lower-than-average (but still stable) transmission.

The Argus slit shows a rather uneven illumination in the robotic flats (but a remarkable stability of fibre bundle transmission).


No specific remarks. The HC plot has been maintained until 2020-03.

Scoring&thresholds Relative fibre efficiency

There is no scoring.

Algorithm Relative fibre efficiency

Not applicable.

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