1 Introduction

1.0 Status

The report is complete. Automatic QC to be commissioned.

1.1 Purpose

This document describes the ING test requirements for the commissioning and acceptance of all new CCDs. As the generic test plan, some tests may not be relevant to a new detector on a particular telescope or focal station. This may be noted in any such test.

1.2 Scope

The tests can be divided into two broad categories:

1. Functional tests - designed to verify and validate the functional behaviour of the CCD within the Data Acquisition System.

2. Acceptance tests - tests designed to verify that the package is acceptable to the ING commissioning astronomer.

The functional and acceptance test will be carried out at the direction of the ING commissioning astronomer, although it is not expected that they will themselves carry out all the tests.

The test results will be fully documented, both in the condensed form of the CCD characteristics data sheet, and in the full form of the completed test plan to be held by the detector group.

Both the functional tests and the acceptance tests might be applicable to the testing of any subsequent modifications to the CCD and it's software - when only the relevant sections need to be carried out.
 

1.3 References

1. TEK4 Commissioning tests, RGMR.

2. TEK on INT/JKT - Testing, and Acceptance, PMF.

3. CCD Commissioning, DRM.

1.4 Overview

In the following sections the functional and acceptance tests are described in detail. Please see the above links (under the title) where the master document, figures, plots and all FITS data reside.

When carrying out the tests in the following, sections, the run numbers of any frames (and dates) evaluated should be recorded for future reference in the ING archive and printouts of relevant images stored with the commissioning file.
 
 

2 Functional Tests

2.1 Mechanical arrangement
 
 

MechU001 The cryostat should accept filler tubes of different lengths and be filled with liquid nitrogen and cooled.
 

Notes: The cryostat will accept both a half length and full length tube and will correctly seal with a teflon washer.

 

MechU002 LN2 filler tube location should allow LN2 to remain in the cryostat at all telescope positions. Put appropriate filler tube in and attach CCD to its nominal focus on the telescope.

MechU003 The filler tube length needed for each of the various instruments that the device will be used on, should be recorded here.

MechU004 Capstans should have clamps on and with these clamps tight, the capstans should have no movement - both rotationally, and in slop.
 

Notes: There is the tiniest of slop on capstan C but this does not cause cryostat movement.

 

MechU005 Establish the position of the CCD surface relative to the window front face. and relative to the cryostat (metal) front face, and record them here.
 

Notes: 11 mm measured optically

 

MechU006 The cryostat should physically fit all the instruments it could ever be used with. Note down here the instruments that have been tested.

MechU007 The flatness of the chip itself should be known. This should have been measured before delivery to ING. Check that there is no gross curvature across the chip by either FWHM measurements of stellar profiles ( if chip is to be used for imaging ) or through Hartmann shifts (or FWHM estimates of arc lines if chip is to be primarily used for spectroscopy) Note down the method used and the approximate shape of any curvature.

Notes: This is a metal buttable device with a peak to valley of 20um as measured by E2V.

 

MechU008 Flexure of the CCD relative to the sky should be noted for the most common focus for the detector. A full flexure test is not necessary for full commissioning, but the flexure between zenith and ~60 degrees from zenith should be noted and compared with what is expected. Advice from the appropriate instrument specialist can be sought.
 

Notes: Five arc images were taken (by Amy Tyndall on 14 Apr 2010) with same settings (low resolution gratings R300V) and telescope pointing 90 deg, 70, 50, 30 and about 15 deg in altitude (starting from the Zenith and lowering only the DEC, telescope no tracking). Same three arc lines were measured in all images. A vertical shift of about 1 pixel (1.87 A, acc to R300V used gratings) for each 20 deg telescope altitude shift could be clearly measured, although FWHM of the lines remained the same for whole interval, thus some flexure is present. Given this, we recommend users to take one arch for each ~10 deg change in altitude from their science to ensure about 1 A precision of lines along the whole lambda interval. Data in /data/ovidiuv/IDS/MechU008 folder, while measurements in my (Ovidiu's) handwritten IDS folder notes.

 

MechU009 It should not be possible for the radiation shield inside the cryostat to impinge on the light path, for any, of the likely foci that the detector may be used. Indicate here that this has been checked and is acceptable.

Notes: The cryostat window is very close to vignetting the corners of the device. This is however well outside the window used on IDS.

 

2.2 CCD performance

PerfU001 The number of the controller used for the performance tests should be noted.

 

PerfU002 The .lod file used for this detector should be recorded here.
 

AstrospectU003 Confirm that the IRAF routines in the ING package work with this CCD and that the advice on direction and magnitude of capstan movements are correct. Confirm also that the collimator movement advice is correct and that the CCD can be focussed within the acceptable non-astigmatic region of the collimator. Contact the instrument specialist from implementing these these routines and any other queries.
 

Notes: During the setup process, I fixed the IRAF ING scripts "ids_rotation" and "ids_tilt" (in fact their associated two files database) to work with REDPLUS2 camera. Two lines defining the REDPLUS2 camera orientation and some plate scale factors (related to micrometer and capstan readings) were added at the end of the files in the two database, namely (see copy here): ids_rotation.dat and ids_tilt.dat. New data was found based on trial and error starting from previous data for EEV10 CCD, because I could not find exactly the meaning and calculation mode for each constant (nobody knew to tell me, incl Pablo former INT and IDS specialist).

 

AstrospectU004 Take exposures of a spectroscopic flux standard under the usual conditions for measuring throughput. This generally means a wide slit, lowest spectral resolution possible plus no filters. Contact the instrument specialist in advance for the required set-up. Either reduce the data and note the results here or inform the instrument specialist of the run numbers and the setup used.
 

Notes: Using R300V lower resolution gratings centered at lambda 7800A, Annemieke observed one standard (SP1036+433) later during the night. I present her reduced spectra (by bias and flat field) here: SP1036+433. Before Annemieke, I also observed two other standards (SP0934+554 and SP0946+139) using the higher resolution gratings H1800V centered at 5000A (run numbers r786151-..52 and r786153, respectively), taking also arcs with 5" and 1" slit width (run numbers r786154-..159). Data in /data/ovidiuv/IDS/AstrospectU004 folder. On 10 May 2011, Lindsay observed the standard star Feige 34 close to Zenith. After comparing the three datasets, I decided to use the last one to derive the throughput of RED+2 with IDS. Final plot included on the public IDS website here and comparison plot with older cameras is here. RED+2 is responding as expected, being more sensitive in the red (above 5000 A compared with EEV10 and 7500 compared with other older CCDs).

 

AstrospectU005 Calculate the minimum usable exposure time on the focal station on which the device is to be used most commonly. This can be done by carrying out a linearity test at low shutter speeds. Write down the method and the results. As this is generally set by the shutter in the instrument itself it will not usually need to be recalculated for every CCD. The commissioning astronomer and the instrument specialist should satisfy themselves that this has been done for the station in question.
 

Notes: I used lamp flats taken with 4s, 3s, 2s, 1s, 0.5s, 0.2s and 0.1s. I divided 4s to the faster ones and studied the result (image and plot column cut) to look for any shutter effect. Clear level changes (drop by about 5%) at the top and bottom of the CCD were seen for the exposures shorter than 1s, and less changes for those longer than 1s (which could be associated with vignetting, I think). Given these, we recommend 1 sec as a limit of visibility of the shutter effect. Data in /data/ovidiuv/IDS/PerfU009 folder, and resulting divided flat images in REDPLUS2-SHUTTER folder (ex: FLAT4DIV0p1 means the resulting flat 4s divided by flat 0.1s image).

 

AstrospectU006 Confirm the pixel size on sky by observing a pair of stars of known separation on the slit.
 

Notes:   I observed a double Tycho star, namely TDSC25498 (found by Vizier and Tycho Double Stars Catalog), carefully placing both stars in the slit (PA = 258.7 deg). I measured the X1 and X2 and calculated the REPLUS2 pixel scale by knowing the separation of the stars from the catalog (79.41"), namely 0.437"/pix (matching the calculated 0.44"/pix taking into account the 13.5 micron pixel size and comparing with the old camera EEV10 pixel scale 0.40"/pix and 15 micron pixel size. Data in /data/ovidiuv/IDS/AstrospectU006 folder.

 

AstrospectU007 Define the direction of movement along the slit viewed on the TV acq. system with respect to the CCD columns or rows e.g.. a movement to the right of the TV results in the object moving which way on the CCD.
 

Notes:   Center the star (run r786201), then move telescope to the South (or star to North) take another image (run r786202), spectra went to the right on REDPLUS2 image. Thus, REDPLUS2 orientation is normal for the INT, with North to the left. Data in /data/ovidiuv/IDS/AstrospectU007 folder.

 

AstrospectU008 Confirm that the slit-to-detector reduction factor expected is reproduced with this CCD ie calculate what width of the slit should project to in terms of the FWHM of the arc lines and measure this at best focus position.
 

Notes: Using H1800V high resolution gratings with 0.6" slit, I obtained FWHM 1.5-1.6 pix = 0.7" for most CuNe+CuAr arc lines across the whole wavelength interval. Former tests using lower resolution R300V gratings gave similar FWHM values.

 

2.3.3 Other device specific tests or comments.
 

Notes: Other tests suggested by Chris and Ian: Fringing: First, I worked on Annemieke's lamp flat using her low resolution setup (R300V, central lambda 7800, coverage 3400-12000A). Not much fringing visible after displaying the whole (zoomed out) image: REDPLUS2-FLAT-R300V-5400A-10000A and column cut REDPLUS2-FLAT-R300V-5400A-10000A-COLCUT and bigger REDPLUS2-FLAT-R300V-5400A-10000A-COLCUT1 and top red end here REDPLUS2-FLAT-R300V-5400A-10000A-TOPRED. Nevertheless, some fringing visible while displaying non zoomed central image with convenient contrast/brightness: REDPLUS2-FLAT-R300V-7200A-8500A-rotated270deg (limits: 7200A at left and 8500A at right). There are different algorithms to plot the fringing. I used 3 approaches and 3 datasets to quantify and plot the fringing: In a first approach I combined 5 lamp flats taken by Amy Tyndall and Lindsay Magill on 14 April 2011 with gratings R300V, lambda central 5510A. I fitted the combined image with FIT1D (spline3, order=20) and divided the two, in order to quatify the fringing, obtaining the following plot: RED+2-FRINGING-MODULATION (not much fringing visible). A second approach to quantify the fringing: I inverted in IRAF the lines in the original flat image (creating a new image) which I then divided by the original flat. Taking a column cut I obtained the following fringing modulation plot: REDPLUS2-FRINGING-R300V-5400A-10000A corresponding to limits between 5400A and 10000A (not much fringing visible) In the third approach I averaged the central columns 201-204 producing a one-dimensional image along the whole wavelength which holds the signature of the fringing, then I averaged the central columns 100-300 producing a similar one-dimensional image which averaged out the fringing. Then, I divided one by the other to plot the fringing in RED+2-FRINGING-MODULATION. I used the reduced lamp flat taken by Annemieke with R300V gratings, central lambda 7790A and cut the plot between 6,000A and 10,000A. I addopt the above third plot as the best, which show some 1% RED+2 fringing above 600nm growing up to 2% close to 1000nm. All approaches show similar results. The following plot FRINGING-REDPLUS2-EEV10 compares fringing of IDS RED+2 (plot in red) with that of the old camera IDS EEV10 (plot in blue) both taken with the same R300V gratings and similar central wavelenghts. While RED+2 remains very low fringing (bellow 0.2%) up to 1000nm, EEV10 frining grows up to 25% at the very red regime. The following mosaic image shows the two reduced fringing with EEV10 (left) and RED+2 (right), both cut between 600nm and 1000nm: FRINGING-REDPLUS2-EEV10-IMAGES Working fringing data in /data/ovidiuv/IDS/Fringing folder. Flexure: Evidenced - please see MechU008 test above. Ghosts: no ghost visible other than a few (~10 pixels) brighter lines at the bottom on the bias. Direction of red/blue: Red is increasing upward (in normal displayed image).

 

2.4 Software tests

SoftU001 The software should detect illegal windows and implement legal ones when requested. Try defining windows outside of the physical range of the device and note the software response and whether it is expected. Check that no windows are saved in the controller particularly window 0.
 

Notes:   It works for illegal windows, so I recommend some tech fix in the near future.

 

SoftU002 The software should detect illegal attempts to bin and implement legal ones when requested. Try to bin in the following, way and note the software response and whether it is as expected: BIN 2x2, BIN 1000x1, BIN, 4x1 BIN 2000x2, BIN 1x4
 

Notes:   It works for illegal binning, so I recommend some tech fix in the near future here.

 

SoftU003 Check that all the information held in the FITS header is correct for this CCD. On the WHT the detectors definition file will have to have this detector added, along with the gains and readout noises for all readout speeds.
 

Notes:   Checked, all data OK in headers, e.g., DETECTOR=REDPLUS2 GAIN=0.91 READNOIS=3.48

 
 
 

3 Acceptance Tests

3.1 Delivery

3.2 Performance

The commissioning astronomer should accept that the chip is configured in an acceptable way

• The chip gain and readout noise is acceptable at the readout speeds available.

• The results from all the performance tests are acceptable (any deficiencies should be noted);

• The results from the astronomical tests are acceptable (any deficiencies should be noted);

• The results from the software tests are acceptable (any deficiencies should be noted, and improvements requested here),

3.3 Acceptance

If improvements have been agreed, or deliveries are yet to be received, this section should not be completed. If the CCD is acceptable in its current state, this document should be signed off by the people below.

Commissioning Astronomer (ING): Ovidiu Vaduvescu, 25th November 2011

Commissioning Engineer (ING): (sign and date) Andy Ridings 28th November 2011

Head of Engineering (ING): (sign and date)
 
 
 
 

When this document has been completed. it must be held by the detector section.