Quality Control and
Data Processing

HC PLOTS
dark current
ricc event rate
QC1 database (advanced users): browse | plot

The darkcurrent calibration is a technical calibration. It is not used to calibrate other calibrations or science data. The darkcurrent template acquires a sequence of dark frames with different DITs. The pipeline recipe fits the relation darkcurrent( DIT ). While the dark level retrieved from the dark recipe gives the dark counts in ADU for a master dark of a given DIT, the darkcurrent gives the current in ADU/sec in a more accurate manner.

QC1_parameters

FITS key	QC1 database: table, name	definition	class*	HC_plot**	more docu
QC.DARKCURRENT	vircam_current..qc_darkcurrent	median dark current [ADU/s]	CAL		[docuSys coming]
*Class: KPI - instrument performance; HC - instrument health; CAL - calibration quality; ENG - engineering parameter **There might be more than one.

The QC.DARKCURRENT gives the dark current (the dark flux) in ADU/sec.

QC report of a dark current technical calibration with the linear fit to the dark counts versus exposure time relaetion. click in the image to see dark current fit of all detectors.

Each detector has its own dark current. There following map shows the values in ADU/sec (June 2010):

chip13	chip 14	chip 15	chip 16
0.405	0.221	0.030	0.052
chip 9	chip 10	chip 11	chip 12
0.065	0.038	0.059	0.053
chip 5	chip 6	chip 7	chip 8
0.066	0.079	0.090	0.097
chip 1	chip 2	chip 3	chip 4
0.296	0.300	0.476	0.355

red labels mark detectors with relatively high dark current, green labels mark detectors with relatively low dark current.

In 2022-08, the current calibrations were acquired with science filters and the following dark current values (here for chip #5 only) have been recorded:

filter	ADU/sec
SUNBL	0.0839
NB118	5.75
NB980	2.93
Z	3.67
Y	3.89
J	7.94
H	14.0
Ks	27.2
SUNBL	0.0833

Trending

Upper left plot shows the average over 16 detector-specific values, the lower left box shows the rms over 16 detectors, the upper right plot shows 16 individual detector values, and the lower right plots shows the detector specific scores.

Scoring&thresholds Dark current

According to the 2018 version of the VIRCAM key performance indicator (KPI) document, the dark current must be lower than 2 ADU/sec for an individual chip.

History

• 2018-06-01: the detector temperature was modified from 75 K to 72 K. As a result the dark current decreased.

Algorithm Dark current

The dark counts versus exposure time is fit by a linear function. The gradient of the fit is the darkcurrent QC parameter.

QC1_parameters

FITS key	QC1 database: table, name	definition	class*	HC_plot**	more docu
QC.DARKCURRENT	vircam_current..eventrateall	events/sec average all frames			[docuSys coming]
*Class: KPI - instrument performance; HC - instrument health; CAL - calibration quality; ENG - engineering parameter **There might be more than one.

The eventrateall gives the rate of radiation induced charge collection events in events/sec. For an introduction of this topic, we refer to this link and to the HAWKI detector features. The motivation to monitor this detector feature is:

• the effect cannot be calibrated
• the effect is intrinsically variable (both spatially and in the number of events) and is so far unpredictable
• science observations are impact and may produce fake source identifications in particular for setups with NDIT=1, large DIT and low number of exposures.

Difference between two consecutive DIT=600 dark frames acquired with the dark current template. Chip #5 is shown.

Median collapsed (in horizontal direction) and re-expanded difference frame consisting only of horizontal stripes, to be subtracted as background. Chip #5 is shown.

Difference frame. The stripe pattern frame from above has been subtracted to obtain a stripe-free difference frame. Chip #5 is shown.

best fit ( f(x) = b * x ) to the event numbers in each of the dark current frames. The slope 'b' is stored as eventrateall. Chip #5 is shown.

destriped difference image of two DIT=600sec dark frames from the dark_current sequence. Left: example from 2009-07 with 0.08 events per sec. Right: example from 2022-08 with 0.13 events per sec.

Trending

The event rates are not trended. They are available in the QC1 DB as eventrate600 and eventrateall per chip. e.g. CHIP #5. This plot shows, that the number of events in chip #5 is increasing with time from about 0.08 events per second (= 4.8 events per minute) in 2009-07 to 0.13 e/s (= 7.8 events per minute) in 2022-07.

For the trending of RICC of other detectors, change filter_extname=DET1.CHIP5 in the address widget of the browser or use the QC1 DB plotting interface.

Scoring&thresholds radiation induced charge collection

The values are not scored

History

• 2020: eventrate600 is generated, results are saved on a test server.
• 2022: eventrateall is generated and put in operation, reprocessing is started.
• In 2022-08 dark current calibrations have been acquired with science filters to test if RICC shows up with the same number of recorded events as with the SUNBLIND filter. The dark current is much higher in science filters than with the SUNBLIND filter (see above) and hence the (photon) noise is larger as well, which makes it more difficult to retrieve the number of events. The visual inspection shows that the number of events is slightly lower in the NB980 filter (with RMS=30 in the destriped 600sec difference image of chip #5) with respect to the SUNBLIND filter (with RMS=10) consistent with the higher photon noise in the NB980 dark frames. Darks of other filters are event more contaminated by the higher photon noise and the resulting lower number of detected events are no more comparable due to the large dark current.
• 2022-09. Instead of analyzing the weekly acquired CURRENT difference frames, it is also possible to analyse the daily acquired DARK DIT=120 RMS frames (deriving for each pixel of the detector the standar deviation over the five exposures (five raw frames)) with appropriate scaling in the number of events (equivalent to 600 sec = five times 120 sec).

Algorithm radiation induced charge collection

• eventrate600: For each detector: The two DIT=600 dark frames (= highest DIT) are subtracted from each other and the median collapsed background is subtracted (see images above). The command /scisoft/bin/peak -F -m clip -k 5.0 is applied to the massaged difference frame (and to its negative compound) to retrieve the number of sources [1]. The average number from both difference images divided by DIT ('imageA minus imageB' and 'imageB minus imageA" since the peak command detects only positive sources) is saved as eventrate600.
• eventrateall: For each detector: All other frame pairs of the same DIT are subtracted from each other and the median collapsed background is subtracted (see images above). The command /scisoft/bin/peak -F -m clip -k 5.0 is applied to the massaged difference frame (and to its negative compound) to retrieve the number of sources. The average number from both difference images ('imageA minus imageB' and 'imageB minus imageA" since the peak command detects only positive sources) is stored for every DIT as in the case of the DIT=600sec frames. The relation between the DIT and number of events is fit by f(x)=a*x, (see plot above) where 'a', the slope is saved as eventrateall.

[1] The number of detected sources when counting the sources by visusal inspection is about 20% larger than what is recorded using /scisoft/bin/peak -F -m clip -k 5.0 . A smaller clipping value of e.q. k = 3.0 would deliver a higher number of events with the higher risk of false spurious detections. We use 5.0 as a trade-off (compromise) value.