Dark Energy Camera (DECam)
==========================

The information was copied from various NOIRLab websites on October 31, 2023.
DECam is a high-performance, wide-field CCD imager mounted at the prime focus of
the Víctor M. Blanco 4-meter Telescope at Cerro Tololo Inter-American
Observatory, with 62 science CCDs (60.5 useful) with 520 megapixels and images 3
square degrees (2.2 degree wide) field at 0.263 arcsecond/pixel resolution.
DECam is a facility instrument, available to all users. DECam was built to carry
out the Dark Energy Survey (DES) Project by the DES Collaboration, which
finished observations in January 2019. Main Characteristics of Focal Plane
-------------------------------------------- Field of View 3 square degree (2.2
degree field of view) Pixel scale 0.2637 arcsec/pixel (center), 0.2626
arcsec/pixel (edge) Detector 62, 2048x4096 pixel red-sensitive science CCDs from
LBNL, 520 megapixels total (570 Mpix including guide and focus CCDs) Read-out
noise 7e- Read-out time 20 seconds Dark-current ~ QE 40% at 400 nm, 70% at 520
nm, then around 90% until 920 nm, falling to 40% at 1000 nm, 10% at 1050 nm
Dynamical range 16 bit Inter-CCD Gaps 3.0 mm (201 pixels) along long edge (e.g.,
between S4 and N4); 2.3 mm (153 pixels) along short edge (e.g., between N4 and
N5) Cosmetics Good to excellent. On average, each CCD has 0.05% bad pixels and
the worst CCD has 0.39% bad pixels. Filters 8 filters now available (ugrizY,
wide VR, narrow N964) Gain 4 e-/ADU (typica) Non linearity For normal observing,
keep level below 100,000 e-, 22,000 ADU Raw data format FITS (with extensions),1
GB/file, ~600 Mbyte/file compressed Instrument F ratio f/2.7 Status of DECam
CCDs ------------------------------ Nov. 16, 2018 (original), updated April 10
2019 and February 13 2020 (ARW). The following is a report on the current status
of DECam CCDs that have presented problems in the past: CCD 61 = N30 Fault: Very
low full well following an over-illumination event in November 2012. Comment:
This CCD is normally masked out in reduction pipelines. This failure prompted
extra “no bright light” operations procedures for DECam, observations closer
to 10 deg from the Moon are prevented, and we added extra protection photodiodes
to supplement those already on the FPA. CCD 31 = S7 Fault: Amplifier B has
unstable gain, from installation September 2012. Comment: Generally the whole
CCD is masked out in reduction pipelines. However Amplifier A does function
correctly. CCD 2 = S30 Fault: Stopped working November 2013, restarted 29 Dec
2016. Comment: It is postulated without any proof that a piece of electrically
conducting debris within the Imager fell on exposed serial clock lines for this
CCD, and then after three years fell off again. There are reports that the
amplifier(s) are unstable at the 0.3% level so caveat emptor for critical work.
CCD 46 = N15 Fault: Hot spot started 10 Nov 2017, intermittent, not seen since
January 2018. Comment: The hot spot was a charge defect and did not generate
light. Generally hot spots are associated with a pixel-level defect, e.g. a
short in the gate structure, and are not intermittent nor do they heal
themselves. So there is no explanation for why this very prominent defect
disappeared, with the affected pixel (800, 2630) now looking completely normal.
CCD 41 = N10 Fault: Poor serial charge transfer on amp B, started 31 August
2018, intermittent, most recent incident 20 October 2018 Comment: This was
diagnosed as a fault associated with a single serial clock line (H3). All
relevant external hardware (clock card, transition card, cables to VIB) has been
changed, but the fault kept reappearing. It was discovered that a script which
toggles vsub for all CCDs, and disables/enables the clock lines for the
associated backplane (#5, 18 CCDs), fixes the fault. This could be understood if
it was clearing a latched driver. CCD 44 = N13 Fault: From November 2018,
appears to be a thread or similar stuck to the surface of the detector near
x=1133, y=1980. FN2 and FS4 Fault: Half of each of these two F&A CCDs has low
full well (10%). Comment: Diagnosed as damage to output amplifier or signal
chain inside the Imager. Since installation, September 2012. FN2 Fault: Has
developed a super-hot pixel (~1645, 2096), since 9 April 2015 Comment: This is
on the very edge of the format.. It was previously known as a slightly hot
pixel, now it is very bright, clearly visible on bias exposures. General: Every
time the DECam CCDs are warmed up and cooled down again the conduction between
the CCDs and their mount can change very slightly, and with it the operating
temperature for the CCD relative to the average for all the others. This
temperature change means that the CCD response cutoff near 1.05 microns can
change slightly (warmer CCD implies redder cutoff), and this can be seen in the
relative flat-field response through the Y filter, which has detector-defined
red cutoff. Such changes have been seen now and then but they are always very
small (no more than a 1% change in Y band throughput), although a possible 2%
change for CCD 27 = S3 as a result of the 2018 March warm/cool cycle is under
investigation by DES. As expected, this is a Y-band effect only. DECam Shutter
-------------------- The DECam shutter is a large shutter made by Bonn Shutter
and consists of two accurately-controlled blades that sweep across the focal
plane, the first movement lets light to the focal plane, the second closes off
the light. The length of time it takes a shutter blade to sweep across the focal
plane is approximately one second. For any exposures longer than one second, the
movement of the two blades is sequential, but for shorter exposures the movement
of the two blades occurs simultaneously, and the two blades form a “slit” that
sweeps across the focal plane. Since the blades are accurately synchronized,
very short exposures (a few milliseconds} are possible. The longest exposures
regularly taken, through u band or narrowband filters normally do not exceed
1200 seconds although there is no technical limitation to going longer. The
number of cosmic ray events on the CCDs starts getting obnoxious for long
exposure times. There are several aspects to be considered if doing short
exposures of moving objects, or very rapid variables. Firstly, the the blades
move in the East-West direction. Blade “A” is on the west side and Blade “B” is
on the east side. So for an exposure assume Blade A covers the focal plane. It
then moves to the west position and so the exposure starts, then at the end of
the exposure Blade B moves from its east position to cover the focal plane. The
following exposure reverses the direction the blades move, and so on.
Unfortunately, which of the two situations (blades moving to the east or to the
west) applies to any given exposure is NOT written into the image header or to
the SISPI log. It is however in the low level PanView log. Since one only needs
to know which of the two situations you have for one “fiducial” exposure in a
given night, that information can easily be provided upon request. Secondly, the
actual epoch at which the shutter opens for a given CCD pixel depends on where
that pixel is situated. The time for the blade to sweep across the focal plan is
approximately one second, so this effect could be of importancce for short
exposure times. Thirdly, the shutter is sited, along with the filters, between
corrector elements C3 and C4, a long way from the focal plane. Photons falling
on a given pixel are contained in a cone that is some 20 cm diameter by the time
it passes through the shutter. And finally, the shutter control is not
hardware-synced to the CCD readout. The complications that might be relevant for
short exposures of moving objects, or rapid variables, are likely to be
irrelevant for the great majority of DECam programs. The following outstanding
specifications mean that very few people need worry about any shutter-related
effects on their photometry, and indeed the shutter performance far exceeds the
DECam design specifications in all respects. Minimum recommended shutter time:
100 msec (non-uniformity is then < 1% over the whole focal plane) Maximum
recommended shutter time: 1800 sec (but beware cosmic rays) Minimum allowed
shutter time: 10 msec Exposure time uniformity: < 1 msec Exposure time
repeatability: << 1msec Exposure time accuracy: << 1 msec Absolute timing: < 1
msec ( internal to the shutter, but the external error in the absolute timing is
very much larger due to external connections) Note these specifications are
those demonstrated at the factory at time of acceptance in 2011. Most of these
specifications we have no way of testing to the required accuracy with DECam on
the Blanco telescope.