DECam Hardware

2. The DECam Hardware

2.1 The CCDs

The DECam has 62 2048x4096 pixel CCD detectors for science images and 12 2048x2048 pixel CCD detectors for guiding, alignment and focus.  The CCD detectors are 250 µm thick, which is about 10 times thicker than conventional CCDs.  This thickness greatly improves their quantum efficiency in the near infrared (800-1100 nm).

As seen below, the number of bad pixels varies from CCD to CCD, with the worst CCD having ~0.4% bad pixels.  The average percentage of bad pixels per CCD is ~0.05% and the median is 0.02%.

Figure 2.1.1  Left: The DES 2k x 4k CCD module.  Right: A histogram of the percentage of bad pixels in the 62 2kx4k DECam CCDs (Flaugher et al. 2012, SPIE).

Figure 2.1.2 shows the layout of the CCDs on the focal plane as well as a typical flat field image.
 

Figure 2.1.2.  Right: The DECam focal plane populated with 72 CCDs.  Left: A flat field image from the DECam.

The schematic layout and dimensions are detailed in Figure 1.2.1 which also gives the dimensions of the gaps between the CCDs.

2.2 The Cooling System

The DECam CCD array is cooled autonomously and can operate for extended time periods (months) without human intervention.  This is achieved by using a liquid nitrogen (LN2) closed loop, two phase circulation system.  The LN2 process tank contains approximately 200 liters of LN2 and this LN2 is circulated to the imager vessel heat exchanger and back.  As the LN2 cools the imager, some evaporates in the process as it absorbs the camera's heat.  Therefore, inside the imager vessel, the LN2 is separated into liquid and gaseous phases.  The liquid phase continues to circulate through the cooling system whereas the gaseous phase is condensed using two cryocoolers, and returned to its cold liquid state.

The cool down time is 4 hours before the CCDs can be read out, and another few hours before the temperature is stable within the requirements of -100°C +/- 0.25°C.

[1]
Figure 2.2.1.  The main components of the DECam vessel, showing (left) the vessel, the readout crates, the vacuum interface boards (which transfer the CCDs’s signals from inside the imager to the readout crates), vacuum pumps, and liquid nitrogen ports.  Also shown (right) is the internal heat exchanger cooling ring used to cool the CCD array.  The heat exchanger is not normally visible due to the protective vessel casing.

The temperatures of the CCS are displayed in the GUI (fill in here).....  The CCDs should be between -95°C and -105°C.

More information on the DECam cooling system can be found in SPIE 2012 Commissioning the LN2 system at CTIO [2], SPIE 2010 Cooling the Dark Energy Camera CCD Array [3] and SPIE 2008 The Dark Energy Survey CCD Imager Design [4].

2.3 The Readout Controllers

The DECam CCDs are read out using the Monsoon [5] system, an open-source image acquisition system designed by the National Optical Astronomy Observatory (NOAO). The electronics consists mainly of three types of modules, the Master Control Board, the Acquistion Board and the Clock Board.  The Clock Board generates the clock rails for the readout of 9 CCDs (there are 13 clocks per CCD and these clocks move charge either vertically or horizontally).  The Acquisition Board provides bias voltage to the CCD and digitalizes CCD output signals. The Master Control Board is designed to control the clock and video boards and communicates the Monsoon crate with higher level communication modules.  The DECam Monsoon system provides a total of 132 video channels, 396 bias levels and about 1000 clock channels in order to readout the full mosaic at 250 kpixel/s speed with 10 e- noise performance.

More information on the DECam Monsoon controller can be found in Castilla et al. 2010 [6].

2.4 The CCD Shutter

The shutter for the DECam consists of stepper motors that drive the blades via belt drives.  This synchronized motion of the shutter blades is advantageous for fast exposures.  The absolute timing of an exposure is measured to a precision of 10 millisecond.  The shortest exposure time is limited to ~5 milliseconds and the exposure duration accuracy is ~1 millisecond.  The exposure start precision is <50 microseconds. (doc-db-750 and 2071-v6).  The DECam shutter is about 2 m long, and 0.76 m wide.

[7]
Figure 2.4.1.  The DECam shutter, pictured with its Bonn University (Germany) manufacturers.

2.5 The Filter Changer

The filter changer provides positions for eight DECam filters.  These filters are housed in four stacked cassette mechanism sub-assemblies, and each cassette houses two filters (see Figure 2.5.1).  Compressed gas moves the filters into place and back out from the aperture. The air cylinders producing the force of the compressed gas have integral air cushions at the end of travel to absorb energy of motion and integral needle valves for safety and speed control.

It takes less than 10 seconds to exchange a filter for another and the DECam filters can be changed while readout is happening.  The filters are always within 0.125 mm of the optimal position.

Figure 2.5.1 Left: The filter changer mechanism Right: An individual filter and filter cell

The current position (as of Jan 23, 2013) of the filters is as follows:

even side                       odd side position

c1f2: g-band in cell 7    c1f1: i-band in cell 3
c2f2: r-band in cell 6    c2f1: z-band in cell 4
c3f2: Y-band in cell 8   c3f1: u-band in cell 5
c4f2: pinhole                c4f1: block

c1f1 is slot 1
c1f2 is slot 2
c2f1 is slot 3
c2f2 is slot 4
c3f1 is slot 5
c3f2 is slot 6
c4f1 is slot 7
c4f2 is slot 8
 

More information on the DECam filter changer can be found in SPIE 2010 Large Format Filter Changer Mechanism [8].

2.6 The Guider CCDs

Guiding with the DECam is accomplished using four 2kx2k CCDs on the north and south sides of the science field.  The field of view of each CCD is about 9.2 arcmin on a side, and at a given location, suitable guide stars are almost always available without moving the telescope from the desired position. 

The acquisition of a guide star is based on stars that (1) have few saturated pixels (2) are far way from the border of the image (3) are without neighbors and (4) are point sources.  SEXtractor is used to facilitate this process.  In each of the four guide CCDs the best reference guide star is chosen and an tracking correction is calculated.  An example plot showing the output offset (error signal to the telescope control system) is shown in Figure 2.6.1.  Therefore it is straightforward to visually see how well the guide CCDs are adjusting the telescope tracking and also to discard or adjust any problematic guide CCD.

Figure 2.6.1. The output offset calculated from each guide star in each guide CCD.  This offset is used to correct the telescope control system and ensure precision telescope tracking. 

The first time slewing to a new field, the process time to find a guide star is about 1.3 seconds.  After this, a region of interest much smaller than the full CCD (the default is 50 x 50 pixels), is saved as a region of interest fits file, and the guider can analyze images in 0.06 seconds.  The precision of the tracking depends on the quality of the guide stars, as well as the number of guide CCDs used.  Table 2.7.1 shows the measurement of the tracking precision using the DECam Guider Software.

Star Magnitude	1 Guide CCD	4 Guide CCDs
15	0.044 pixels	0.038 pixels
16	0.074 pixels	0.049 pixels
17	0.103 pixels	0.056 pixels
18	0.143 pixels	0.082 pixels
19	0.220 pixels	0.095 pixels

Table 2.7.1 Precision of the tracking error as a function of the brightness of the guide stars and the number of guide CCDs used.

There are three configurations that can be used for guiding:

1.  Auto -- The guider continuously reads new images, tracks guide stars and sends correction signals until asked to stop guiding.
2.  Self -- The guider analyzes the next (first) image and finds the best guide star for each of the four guide CCDs. 
3.  User -- The guider presents the image to the observer, who then selects the guide stars.

The Auto guider is the fastest and easiest, and this configuration is recommended.

The DECam guider manual can be found here (there is a draft of this manual, I will add the link when a more robust version is ready). 

2.7 The Focus and Alignment CCDs

The focus and alignment system of DECam is automatic--the camera actively controls the focus and alignment so the user can sit back and enjoy in focus images.  The adjustments made to ensure precision focus are hexapod tilt (2 axes), hexapod translation (2 axes), hexapod piston (focus) (1 axis) and primary mirror support system astigmatism (2 axes).  These adjustments are made using eight 2K by 2K CCDs, which are defocused by either 1.5 mm extra- or 1.5 mm intra-focally.  As exposures are taken, the donuts (out of focus star images) on each 2k x 2k CCD are analyzed for coma, astigmatism and focus (see Figure 2.7.1).  A Zernike analysis of the wavefront error at the focal plane measures the Zernike parameters, which are then turned into hexapod and primary mirror corrections.  “Time is precious,” Physicist Aaron Roodman of SLAC National Accelerator Laboratory says. “The nice thing about this system is that it will focus automatically every minute."

Figure 2.7.1  Simulated out-of-focus star images (donut) images, resulting from imposing aberrations on the primary mirror. Left donut shows the comatic aberration resulting from the optical axis being not perfectly aligned.  Right donut shows the astigmation resulting by moving the primary mirror closer to the focal plane by 300 microns.  The middle donut shows a simulated DECam donut, smeared to simulate 0.75 arc-second seeing and pixelated to simulate the CCD pixel scale of 0.27"/pixel.

[9]
Figure 2.7.2  The corner of a section of a DECam image, showing the 2k x 2k focus and alignment CCD with out of focus stars.  The out of focus stars are used to ensure that science frame CCDs has stars which are perfectly in focus.

More information on the focus and alignment CCDs and Zernike analysis can be found in Roodman (2012) [10].  Note that the highest consumption per capita of donuts is in Canada [11].
 

2.8 The Hexapod

The DECam hexapod consists of six jacks to finely tune the x,y,z (lateral, longitudinal and vertical) linear movements as well as the three rotations pitch, roll and yaw.  Because of temperature variations and telescope movement, the DECam longitudinal position, transverse positions and tip-tilt angles change slightly relative to the primary mirror between exposures.  The DECam hexapod is driven by the wavefront sensor CCDs and from laser alignment sensors that provide active control of focus and alignment.  The position of the camera barrel and focal plane is adjusted, positioning the DECam in an accurate and stable optical alignment and ensuring very precise and repeatable focus, tip-tilt, and transverse movements.  Generally the hexapod makes very small (~50 microns and 5 arcseconds) adjustments in transverse motion, in focus, and in tip-tilt, after every exposure to keep the camera aligned with the primary mirror.  (doc-db-1213-v7)

Figure 2.8.1  The DECam Hexapod

2.9 Correctors

There is no Atmospheric Dispersion Corrector (ADC) available with the DECam.  The Earth's atmosphere disperses the light from stars significantly when observing away from zenith. The effect is greatest and similar at U and B where the stellar image is stretched, for example, ~0.5" at a zenith distance of 45° (1.4 airmasses), and 0.9" at 60° (2 airmasses). We should test the blurring for long exposures at high zenith distances during commissioning.