CAPSCam Users Manual
Alan Boss / Alycia Weinberger
Last Modified: 31 July 2012
Table of Contents
This manual describes how to take direct imaging data using the CAPSCam camera (Carnegie Astrometric Planet Search Camera) on the 100-inch du Pont telescope.
Please pass along any comments or suggestions for improving the acquisition program or this manual to Alan Boss.
Table 1 describes the 100-inch du Pont telescope and Table 2 describes the performance characteristics of the CAPSCam detector.
|Table 1. du Pont 100-inch Direct Imaging Mode|
|Type:||Ritchey-Crétien optimized at f/7.5|
|Diameter of Primary:||100″ (2.540 m)|
|Primary Focal Length:||300.0″ (7.620 m)|
|Cassegrain Focal Length:||750.0″ (19.050 m)|
|Nominal Plate Scale:||10.8 arcsec/mm|
|Field of View:||718 mm (~2.1 degrees) in diameter|
|Sky Shields:||2-element provides complete shielding over 10 arcmin diameter field and blocks 3% of beam|
|3-element provides complete shielding over 18 arcmin diameter field and blocks 11% of beam|
|Table 2. Characteristics of CAPSCam|
|Detector:||Teledyne Hawaii-2RG HyViSi Array|
|Pixels:||2048 x 2048 x 18 microns (2040 x 2040 imaging pixels)|
|Plate Scale and Field of View:||0.194 arcsec/pixel, 6.63 x 6.63 arcmin|
|Detector Gain:||2.1 electrons/ADU|
|Detector Read Noise:||12.5 electrons|
|Detector Linear Full Well:||130,000 electrons|
|Dewar Window:||Astrometric quality filter (Barr Associates lambda/30), bandpass 8100 – 9100 A|
The day crew will have set up the camera and the control computers. The camera is controlled by a PC located in a rack in the telescope control room. The control GUI runs on the observer’s Mac mini computer in the control room. Normally the day crew will start the program and make sure that everything is running; the start-up procedure is given here in case you have to do it. The GUI should be left running at all times, as it is needed to maintain the proper temperature in CAPSCam — do NOT exit the GUI when finished with a night’s observations.
To get started from scratch on the 100-inch, login to the Linux workstation with username obs1 or obs2. The correct user name and password will be given to you by the instrument specialist. Open a terminal window.
In the terminal window type caps. This will start the CAPSCam configuration window shown in Figure 1.
In the CCD/DSP line, type in
- caps.s – the full guide window shutter mode, in which the shutter closes during every GW readout. This is the standard mode for the CAPS program.
- caps-nogws.s – the no guide window shutter mode in which the shutter closes during the initial and final FF read but remains open during all intermediate GW reads. The latter is more efficient, minimizes wear on the shutter, and eliminates the star-shaped shutter pattern on the field, but it also results in slightly different exposure times on the FF and GW and results in slightly lower astrometric precision.
For regular observing CCD: and Telescope: are set to online, and the CCD-Host is usually ccd11 (or ccd10).
Click on OK to start the CAPSCam-GUI tool.
IRAF can be used to display the data and to measure simple diagnostics from the images. To start IRAF and the display tool DS9, go to the Dock and clik on the red iraf button icon. Alternatively, type in goiraf in an xterm. At the IRAF prompt, the ds9 window can be set to the correct size by typing at the ecl> set stdimage=imt2048.
To display and examine an image, at the ecl> prompt, type:
imexam pxxxx 1 (displays and then examines an image)
Inside imexam, useful commands are “r” to measure the FWHM in pixels of a star (1 pixel = 0.194″), “e” to show a contour plot of a region, and “h” to get a histogram of pixel values.
The CAPSCam command window controls all aspects of taking data. Figure 2 shows the layout of the command window.
The window is divided into several parts.
Pull Down Menus
The top part contains three pull down menus, and a bar graph showing disk capacity. The first pull down menu, labeled File, allows the user to reload the camera DSP code and to exit gracefully from the program. The second menu, labeled Mod, allows the user to start a large or small format quicklook tool (see Section 4 Quick Look Tool). There is also the option to start a Dewar Status GUI (see Appendix A Camera Cool Down), and to start a GUI to modify the camera operating voltages (this GUI is for engineering purposes only and is password protected). The Dewar Status GUI should be clicked on and monitored frequently, as if CAPSCam should run low on liquid nitrogen coolant, this GUI will be the first warning to stop taking data and get CAPSCam refilled with LN2. The third menu, labeled Options, allows the user to select the data directory where the observations will be located.
Data taking is described in the following sections.
Unlike a CCD, the HyViSi Hawaii-2RG arrays are made with a light sensing layer (silicon in this case) bonded to a multiplexer. This allows each pixel to be addressed individually. The result is that a single sub-array, known as the Guide Window (GW) can be read out multiple times while the rest of the array, i.e. the Full Field (FF), is integrating. For the purposes of the CAPS program, the GW will be centered on a bright program star. Multiple images of this star will be taken while the array integrates on the surrounding background field containing (faint) astrometric reference stars. Thus the end result is a camera with very large dynamic range. Data are taken on the very bright program star and the faint reference stars simultaneously on the same detector. Binned images are not possible, nor is any charge shuffling on the detector.
The detector has 2040 x 2040 light sensitive pixels. The array is surrounded on all sides by four pixels which are not illuminated, and which can be used to monitor “bias” fluctuations during the camera readout.
- ExpTime: This sets the exposure time (in seconds, minimum is 1 second) for the Full Frame images. When the exposure time is changed this panel is highlighted in red, and the exposure time is recorded by hitting Enter. This is written to the FITS header as keyword EXPTIME.
- Loop: This sets the number of exposures to repeat in a loop. This is written to the FITS header as keyword NLOOPS. It may be changed at any time during an exposure sequence.
- ReadMode: This selects the Read Mode for the camera. Data can be taken in Full Frame (FF) only (no Guide Window), in Full Frame plus Guide Window (FF+GW), or in Guide Window mode only (GW). This is written to the FITS header as keyword READMODE.
- EType: This selects the Exposure Type for the image. For Exposure Type Object and Flat, this is just a flag that is written to the fits file header keyword EXPTYPE. For exposure type Dark then the shutter is not opened for the exposure.
- GuideExp: This panel sets the exposure time (in seconds, minimum is 0.2 seconds) for the Guide Window exposures. There is some overhead in the operation of the array for both the Full Frame and the Guide Window exposures, and the program calculates the number of Guide Windowed exposures that will fit within the exposure time of the Full Frame exposure. This number is displayed in the panel immediately to the right of the GuideExp panel. The active Guide Window exposure number is displayed in the next panel to the right during an exposure sequence. This entire process is repeated for each exposure in a loop. The GW exposure time is written to the FITS header as keyword GW-EXPT and the number of GW frames as keyword GW-NUM.
- GuideGeo: This sets the geometry of the Guide Window within the Full Frame. The first two numbers define the central pixel of the Guide Window in the array, and the next two numbers define the width and height of the Guide Window. Only one Guide Window can be defined. This is written to the FITS header as keyword GW-GEO. The GW can be forced to fall on the target star by moving the green box in the Overview (QL-Tool) Window (using the mouse or up and down arrows on the keyboard) to be centered on the target star and then hitting <ALT> <G>. The coordinates for the GW on that star will then appear in the GUI box. Note: CAPS targets should generally be taken with 1024 1024 64 64.
The exposure or loop of exposures is started by clicking on the Start button. Any exposure or loop of exposures can be aborted by clicking on the Abort button. To stop after the current exposure, set Loops to value of current loop.
The File# window defines the file number for an observation. In Full Frame mode the chip is read out in four stripes of 2048 rows x 512 columns through four amplifiers. The Full Frame is stored as a single stitched-together image. In Guide Window mode the selected subraster is read out through a single amplifier. All four readout channels are nominally identical, but there may be some offsets in the image data (see Section 5 for a description of the required calibration frames).
The data are written to disk as 32-bit integers. As of 7 April 2011, the Full Frame and Guide Window observations are written as independent fits images with names pNNNN.fits and pNNNNg.fits, where NNNN is the file number. In addition, the sum of all the GW images is inserted into the saved FF file. Prior to April 2011, each GW image was saved separately in a file of the form pNNNNgMMM.fits, where MMM is the sequence number of the GW image.
The location of the Guide Window in the Full Frame is stored in the fits headers as the fits keyword GW-GEO with the usual IRAF format for an image section: [X1:X2,Y1:Y2].
The standard size of the GW is 64 x 64 pixels.
Note that if a frame with the same number as the current exposure already exists on the disk, it will be overwritten.
In order to achieve the best astrometric accuracy, each CAPSCam target should be dithered around in a square during any set of exposures.
To execute a macro, type the full path of the macro file in the field next to the Execute button and then hit the Execute button.
Dithering is best done using the macro.dither file stored in /Users/obs1/CAPSCam or /Users/obs2/CAPSCam. Before executing the macro, position the target to the upper left of center in the GW. The macro starts taking data at the current location and then moves dx=2.0, dy=2.0, and dx=-2.0, so that the target is dithered around the corners of a square with a side 2 arcseconds long. After the target is in place, make sure the operator knows to start guiding using the F3 command. The macro makes coordinated telescope/guider offsets using the “coord” command. Thus the operator does not have to be involved in the dithering, as long as the F3 command is used to start the guider.
Dithers may also be done using the TelOff dx and dy fields on the CAPSCam GUI and then pressing the Move button. TelOff moves only the telescope; it does not do a coordinated telescope/guider offset. Tell the telescope operator to stop guiding (F1), move, and then ask the operator to begin guiding on the new position by centering the guide box on the star before starting the guider (F5).
dx >0 moves the telescope East (i.e. the star right to higher x pixel values on the detector)
dy >0 moves the telescope North (i.e. the star down to lower y pixel values on the detector)
Macros are also stored there for the dome flats and darks: macro.flats and macro.darks. They are also invoked by entering the macro name with its full path inside the GUI window next to the EXECUTE button, and then clicking on the EXECUTE button. Unlike the macro.dither, these set the exposure times and loops values.
To recenter the target within the GW, place the green cursor box in the QL-Tool magnifier window on the location where you wish to center the GW, and then press ALT-C to enter the necessary coordinate shifts in the TelOff boxes in the GUI. Then tell the operator you wish to do a Move so he can stop guiding, and hit the Move button.
Saturation occurs at ~44,000 DN. The A/D max is 65,536 but the image bias level takes up some of this dynamic range. By 50,000 DN, horizontal trailing streaks appear in the images. These levels apply for data taken after 16 February 2009. Prior to that, the offset voltages were different and saturation occurred at ~25,000 DN.
Short Guide Window times result in a large overhead and a loss of observing efficiency. For example, for a Full Frame exposure time of 120 seconds, the overhead time required for a FF+GW sequence varies from 18 seconds for a GW exposure of 30 seconds, which is the same as if the GW was off, to 200 seconds for a GW exposure of 0.2 seconds. The observing efficiency is even worse for shorter FF exposures, so it is best to use a relatively long FF exposure and the longest GW exposure possible that does not lead to saturation of the target star in the GW.
Bright stars should be avoided as they lead to charge persistence on the array that can last for hours. Bright stars should not be used to do the initial “C-Set” of the du Pont when starting a night’s observations. Instead, go to your first target field and use your finding chart to do the C-Set.
Charge persistence is largely avoided when exposures are chosen to keep data numbers less than about 20,000.
A QuickLookTool — launched from the ‘Options’ menu of the control GUI — will display the data in real-time at the end of a readout after the data are stored to disk. The user can select a large display (filling the display monitor) or small display from Options menu. The tool comes with three windows, a main display window, a magnifier window, and a control window.
The main display window is shown in Figure 3. This window has a pull down menu to set various automatic and manual image display scaling options, and has a “compass” indicator of the orientation of the field on the sky. The window contains a yellow outline square indicating the location of the magnifier window. During each full frame and windowed exposure the magnifier window defaults to the position of the window in the array.
The magnifier window is shown in Figure 4. This window allows the user to zoom in at 1X, 2X, 4X, or 8X to inspect the data at the location of the yellow box in the main display. The magnifier window can be stretched to a two-times-larger box size.
Mag4 is the best size to use in general, as the entire GW is then visible when the GW size is 64 x 64 pixels. Set the Radius to 24, so that the peak counts encountered by the target star in the GW are displayed in the min/max box. Counts above about 35,000 result in saturation and should be avoided by using a shorter GW exposure time. The Grey map color will show Red when saturation occurs in the relevant pixels.
The zoom is controled by the Quick Look Tool control window, shown in Figure 5. Placing the cursor in the main window and hitting the left mouse button will sample the exposure level at a given pixel position. Statistical information is given by hitting “return” or the space bar when the cursor is positioned appropriately. The Radius feature controls the area of the fit for that exposure. This rms diameter is not closely equivalent to what is calculated in IRAF imexamine (which is more equivalent to the image FWHM) but is still useful for relative measurements.
Set the Radius to 32 in order to sample the entire inner region of the GW with the mag4 option for the 64 x 64 GW size.
This section describes the basic images you will need to correctly reduce, calibrate, and analyze your data. Use the macros macro.flats and macro.darks for the standard CAPSCam calibration images.
The array shows significant dark current and in addition has additive instrumental signatures that depend on exposure time. There are also instrumental signatures that depend on the size and location of the window if this mode of observing is used. It is essential that a frequent series of dark frames be taken with the same exposure times for the full frames to be used on the target stars. These frames can be median filtered and subtracted from the raw data before processing begins. Typically, 10 darks for each exposure time will suffice.
No bias frames are necessary, as any additive structure is removed when the appropriate dark frames are subtracted.
Flat field frames can be taken with either the flat field screen on the dome in the afternoon, or during evening or morning twilight.
For dome flats, typical exposure times are 10 seconds with the flat field lamps set to 260 units, resulting in peak counts of about 3000 data numbers. The control box for the flat field lamps is at the bottom of the rack containing the UT/ST clock to the left of the observer’s desk in the observing room. Ask the instrument specialists to set the telescope to look at the flat field screen for dome flats.
Dome flats are also useful for determining if the optical path of the du Pont has been set-up correctly. If the dome flats show obscuration, the optical path is probably blocked. The correct settings for the CAPSCam configuration on the Blue Offset Guider Box above the observer’s workstation are Diagonal Mirror Position A (the A light is burned out though) and the Prism Turret set to 1/Park.
Fringing in the CAPSCam images is essentially non-existent, and so fringe frames are not needed.
The dark contribution is negligible, even to 600 sec exposures. However there is some significant (but stable) bias component, especially using the Guide Window.
For Full Frames (no GW), flat fields can be taken in the standard way, e.g., 10 sec exposures with the lamp at 300 V yields about 10000 counts per pixel.
Dome Flats and the Shutter
If Dome Flats are taken using the fast GW reading modes (exptime < 1 s), a significant structure is apparent in the images due to the finite shutter open/close time. This generates a six-petal flat-like structure that can be calibrated out as a flat effect, but it strongly depends on the GW exposure time. The six-petal shape can also be seen in some images on the sky. Fully removing this structure requires a number of additional calibration frames.
Center the GW on the target star at the same position at the center of the image and with the same size, i.e., at 1024 1024 64 64. If this is not possible (e.g., you need a larger GW, or have to move very bright object out of the FOV), you will need Dark Frames for each GW configuration. Putting the GW at the central position minimizes the astrometric effect of the petal like structure, even if it cannot be totally removed.
The recommended set of calibrations per night of observation would then be:
Full Frame mode:
Dark Frame FF: 10 x 10 sec exposure
Dome Flat FF: 10 x 10 sec exposure (lamp @ 300 V)
Guide Window mode:
Dome Flat FF+GW: 10 x 10sec FF/ 0.2 sec GW (lamp @ 300 V)
Dome Flat FF+GW: 10 x 10sec FF/ 1.0 sec GW (lamp @ 300 V)
Guide Window mode with all different GW positions used during the night**
Dark Frame FF+GW: 10 x 10sec FF/ 0.2 sec GW
* Can be done at the end of the night
** Required if the GW exposure time is < 1 sec (under investigation). The GW can be positioned at any place, but the best location is 1024 1024 64 64. IF THE PETAL-LIKE STRUCTURE DOES NOT APPEAR, IT PROBABLY MEANS THAT THE SHUTTER IS NOT WORKING.
By far the easiest way to find guide stars for the du Pont telescope is to use the GMap Guidestar Tool. GMap actually consists of two different tools: Skymap which displays guidestars available in the SAO, PPM, or HST Guide Star catalogs for any position on the sky, and Airmass which plots a graph of airmass vs. U.T., again for any sky position. See the manual for further details on the setup and use of GMap on the du Pont telescope.
The output of GMap is the offset guider xy coordinates for the guide star that has been selected. The guider/finder assembly for the 100-inch is presently separate from the data systems. See the Instrument Mounting Base manual in the 100-inch observing room.
The night assistants will handle the choice of the guide star and seldom have problems finding one.
Better yet, copy the data to your laptop and/or external disk. Use on your laptop the command:
> sftp firstname.lastname@example.org
to copy over the data files stored on /Volumes/Data_Clarity/obs2/xxx on clarity.
CAPSCam has three vertical stripes on each image that delineate the domains of the four read-out amplifiers. Avoid placing the target star close enough to these stripes that the star might fall close to one during the dithering process. The shutter will fail after some number (~million) of operations. In caps.s, the shutter “petal” pattern should be visible in FF+GW flat fields with small GW exposure times (e.g. 0.2 s). Also, when listening to the shutter operate, it should have regular single clicks. Some times the guider diagonal is left in by mistake, and this causes vignetting at the bottom corners of the camera. Check the blue guider box to see that Diagonal Mirror is at “A” and Prism Turret is at “1 PARK” Double check that the camera is not rotated too much from N up / E left. You can see this easily by pointing to a bright star and examining the secondary spider diffraction spikes. The spikes should be well aligned with the X/Y directions. When all else fails, tune the radio to the good La Serena 80s station FM 90.1
Once the telescope has been focused, the focus can change as a function of airmass and temperature. These variations are tracked at the 100-inch by the telescope control program over a reasonable range of airmasses. Note that the focus is azimuthally dependent at airmasses of greater than about 1.7. This amounts to some 20 units, and is not included in the control program. See the TCS manual in the 100-inch observing room for more information. For precision astrometry, of course, high airmasses should be avoided – ideally targets should be observed as they pass over the zenith.
Focus: The Guide Window presents a handy means for focusing the du Pont. Put a suitably bright star in the GW and set the ExpTime to at least 100 seconds, the GuideExp to 0.5 seconds or so, and the ReadMode to GW-only, hit Start, and watch the maximum number of counts of the star in the Magnifier QL Tool — note the highest values achieved as the telescope operator slowly changes the focus. Ask for an “F-Set” on the focus that gives the highest number of counts, then do a 10 second exposure to check to see if the du Pont is indeed focused by using iraf to plot the psf of the stars in the image. The psf has very little scatter about the best fit profile when the du Pont is properly focused. The du Pont typically holds an excellent focus all night long, though it can appear to be degraded at certain inclinations and high airmasses.
TeleGUI: Start the TeleGUI application from the Dock on clarity. Load a target list (CMD-O). The CAPS master list is in /Users/obs1/CAPSCam/CAPS_master_list.objects and the typical FF/GW integration times are listed under the “Comment” field. Select the appropriate target and hit “Send to TCS.”
Unlike CCD’s, it is very important to control the cool down and warm up rates for the Hawaii-2RG arrays. If the temperature of the array changes faster than one degree Celsius per minute then thermal stresses in the array and bonds between the silicon and the multiplexer can destroy the device. As a result, it is essential to cool the array from room temperature only when the camera is on the telescope, powered up, and with the CAPSCam GUI control program running.
The temperature of the array is controlled by heater resistors in the header mount of the detector in the dewar. A section of the control program adjusts these heater resistors and the temperature set point during cool down to keep the rate of change of temperature to an acceptable level. This cool down sequencing is run from the Dewar Status GUI started from the Options pull down menu in the CAPSCam Control GUI. This Dewar Status GUI is shown in Figure 6.
This window displays the temperatures and heater currents for the two temperature sensors on the detector. In normal operation these will update every minute of the TC-Loop control is turned on (the default setting). The target operating temperature of the array is -150 C. The present status of the temperature sensors and heater current can be updated at any time with the Update button. The following is the ONLY accepted sequence for preparing CAPSCam for an observing run:
- Pump the dewar while at room temperature.
- Mount the dewar on the telescope. The side that is to point north is noted with the letter “N” on the housing.
- Attach the communications optical fiber to the saddlebag and connect the dewar to the 48-volt power supply. Power up the dewar and start the CAPSCam control GUI.
- Launch the Dewar Status GUI from the Options pull-down menu in the control GUI.
- Set the function to on and fill the dewar with liquid nitrogen in the normal way. The dewar will take approximately five hours to cool to the final operating temperature of -150 C.
- If the camera has to be warmed up then the nitrogen can be dumped from the dewar, but the dewar cannot be warmed by blowing dry air or nitrogen into the dewar. The dewar MUST warm up passively.
Figure 7 shows a plot of the combined window transmission curve and the detector quantum efficiency as a function of wavelength.
|Table 5. People to Contact|
|Person||Phone Number||Phone Location|
|Alan Boss||boss at dtm.ciw.edu||1-202-478-8858||office|
|Ian Thompson||ian at ociw.edu||1-626-304-0225||office|
|Greg Burley||burley at ociw.edu||1-626-304-0261||office|