ING Banner
Home > Astronomy > INT CCD Quality Control

INT CCD Quality Control


  1. Introduction
  2. Off-sky quality control of CCD characteristics
  3. CCD linearity tests
  4. WFC 0.5s timing offset
  5. Acknowledgements

1. Introduction

Any first S/D night of a new WFC or IDS instrument run, the ING students should complete the following quality control (QC) tests for the available camera (WFC or IDS), then measure the noise, gain and bias, then send the data to the corresponding instrument specialists.

2. Off-sky quality control of CCD characteristics

The quality control tests comprises two steps: image acquisition and measurement. We offer two options for each step: automated PyRAF scripts or manual acquisition and IRAF measurement. The automated option requires measurement right after acquisition (during the day before support or D-night). The manual option allows for measurement at later stage after acqusition (could be even next day).

The basic procedure usually tests slow readout and 1x1 binning which must be tested before any D-night. The results should be sent to the instrument specialist (WFC: or IDS:

2.1 Automated PyRAF option

The automated option consists of two PyRAF scripts.

The first script,, is for acquisition and needs to be run on the ICS environment (to be used under SYS> prompt). The acquisition script works for any speed and binning.

The second script,, should be run on the data processing console (to be used on the intdrpc1 computer). The measurement script works for any readout speed but only for binning 1x1.

Follow thes followingsteps for the acquisition and measurements:

2.1.1 Acquisition

Select the desired readout speed (slow or fast) and binning (1x1 or 2x2).

Before running the script, just run a glance to get the right exposure time for flats. Alternatively, for WFC read the exposure time for dome flats from the page WFC Dome Flat Assistant. The script should be available under the normal working folder /home/intobs/.
Running the script without parameters will show possible options and ask them one by one. Alternatively, the options can also be supplied together with the initial command, using the syntax:
   SYS> WFC 2

2.1.2 Measurements

Now on the data PC (intdrpc1), run the script to measure the bias, readout noise and gain:
   ecl>  ! INSTRUMENT
   ecl>  ! WFC
The script will ask which is the run number of the first bias and then the run number of the first dome flat. Alternatively, the prompt can be left empty and press enter. In this case, the script analyses the last two biases and two flats obtained with the script The output will specify bias (ADU), readout noise (ADU) and gain (e-/ADU) which can be compared with and should be included later in the CCD Quality Control page.

2.2 Manual IRAF option:

This option is available should the automated version fail for whatever reason.

2.2.1 Acquisition

Please follow the next steps, after selecting the desired readout (slow or fast) and binning (1x1 or 2x2) for the given camera (WFC or IDS):
    1. WFC: use any broadband filter (or no filter), prepare dome lamps for flats
        IDS: prepare for lamp flats
    2. check texp to get ~40 Kcounts max
    3. take two identic flats and check on the obslog their observing time to avoid the known random WFC 0.5s timing offset
    4. shut down the flat light and take a pair of biases;
    5. write down the run numbers for each dataset.

2.2.2 Measurements

    1. launch IRAF in the INT data reduction computer intdrpc1) and load noao.obsutil package
    2. cd to the working day directory (ex /obsdata/inta/20141102)
    3. run findgain for each pair of flats and biases (WFC for each CCD; IDS for the only CCD)

    For the WFC:

    4. for example, to analyze slow mode 1x1 images, r815593 and r815594 (pair of biases, taken last) and r815591 and r815592 (pair of flats, taken first), run:

    - findgain r815591[1] r815592[1] r815593[1] r815594[1] section="[700:1500,100:4000]"
    - findgain r815591[2] r815592[2] r815593[2] r815594[2] section="[700:1500,100:4000]"
    - findgain r815591[3] r815592[3] r815593[3] r815594[3] section="[700:1500,100:4000]"
    - findgain r815591[4] r815592[4] r815593[4] r815594[4] section="[700:1500,100:4000]"

    5. write down the noise and the gain;

    6. imstat r815593[1],r815594[1] also for [2], [3] and [4] CCDs
    7. write down the averaged median (midpt or mean column);

    8. transform the noise from electrons to ADU (by dividing the findgain's noise by the gain, which is needed to report in the table (by WFC manager).

    For the IDS (any of the two cameras, Red+2 or EEV10):

    4. findgain r815591[1] r815592[1] r815593[1] r815594[1] section="[785:1150,500:3700]"
    5. write down the noise and the gain;

    6. imstat r815593[1],r815594[1] also for [2], [3] and [4] CCDs
    7. write down the averaged (eyeball) median (midpt or mean column);
    8. transform the noise from electrons to ADU (by dividing the findgain's noise by the gain, which is needed to report in the table (by WFC manager).

3. CCD linearity test

For the IDS:

Setup: any. Use the following steps:

1. switch on the W lamp. Bin 1x1, slow mode
2. set up ND filters for getting around (or less) 30000 ADUs in a 30sec exposure.
3. take W flats with increasing exposure times from t=0.1s, 0.2, 0.5, 1, 2, 3....(steps of 1-5 sec) until saturation.
4. if you have time, repeat steps 1-3 for fast mode, and for binning 2x1, 1x2, and 2x2

5. Send the results to the instrument specialist (

For the WFC:

We run the test in Summer 2011 and do not expect any change soon. The results are here.

4. WFC 0.5s timing offset

There is an old known small defect with the WFC exposure times due to communication between the ICS and the old MCA controler of the WFC. This problems dates since at least 2004 (acc to Robert Greimel who first studied it) and it is present still (acc to Ovidiu Vaduvescu who also checked it in 2012). Once every ~7-10 WFC taken images and apparently randomly, the exposed time and consequently the EXPTIME value in the FITS images and the observing logs are showing exposures about 0.5s shorter than the commanded values (taken via run, multrun, flat or multflat commands).

This problem could affect the your above QA images, thus before working with them, please rapidly check the exposure time in the observing log (example: if you took 5s flats, none of the images should show 4.5s or 4.6s values in the log). In case any of the taken flat shows this problem, then reject it and take another pair.

5. Acknowledgements

Credits should be given to the ING student Pieter van Oers who contributed in 2013 with the Phyton scripts to acquire and reduce the QC data.

Top | Back

INT Manager
Last modified: 4 January 2022 (minor)