The Boller and Chivens Spectrograph
Mark Phillips, Ian Thompson, Bill Kunkel, Nidia Morrell January 2002, March 2006; June 2010
Table of Contents
This document describes the operating procedures for the Boller and Chivens (B&C) spectrograph. This instrument was originally designed for use on the du Pont 100-inch telescope with an image tube system and photographic plates. Later, the spectrograph was modified for use with an intensified Reticon scanner. In the mid-1980s, the spectrograph was equipped with a 2D-Frutti photon counting detector and was used with a plug-board fiber optic system for multi-object spectroscopy. Later, a bench-mounted spectrograph was built for the fiber system which employed the semi-solid Bowen-Schmidt camera from the B&C. In the 1990s the B&C was put back into use on the du Pont telescope with a Nikon lens serving as the camera and a CCD as the detector. In 2000, the Nikon lens was replaced by the original Bowen-Schmidt camera mated to a 1024x1024 pixel CCD with 24 micron pixels (Tek#1). This camera/CCD combination was briefly used with either the B&C or Fiber spectrographs. In December 2001 the Tek#1 CCD was replaced by a 2048x515 pixel CCD with 13.5 micron pixels (Marconi#1) which has resulted in a significant (~50%) improvement in efficiency.
In February, 2001 the B&C was commissioned for operation at the f/11 Nasmyth focus of the Baade 6.5-m telescope (a.k.a. Magellan I). This involved replacing the f/7.5 collimator used on the du Pont telescope with an f/10 collimator borrowed from the CTIO 1-m telescope B&C spectrograph, which is a near-twin of the LCO B&C. In March 2006 the original f/7.5 collimator was reinstalled and the spectrograph moved back to the du Pont telescope. This manual is intended for users of the B&C on the du Pont telescope. Please pass along any comments and suggestions for improving this manual to Mark Phillips (mmp at lco.cl).
The B&C spectrograph is a modification of the standard Boller and Chivens Model 31523 Cassegrain spectrograph built in the mid-1970s to meet special Carnegie specifications. Two of the distinctive characteristics of the spectrograph were the absence of slit viewing optics (these are built into the Instrument Mounting Base (IMB) of the du Pont telescope) and the capability for remote control of the decker mask, spacing of the comparison prisms, comparison lamps, and the shutter.
In this section, we describe the optical layout of the spectrograph, following the path that a photon would travel from the telescope (or comparison lamps) to the Marconi#1 CCD. In reading this section, the reader should refer to Figure 1 which shows an optical diagram of the spectrograph. (Note that this drawing of the spectrograph includes components such as the field lens, spectrum widener, and exposure meter which are no longer used.)
Three comparison sources are provided: an Fe-Ar hollow cathode lamp, a He-Ar glow discharge lamp, and a Ne glow discharge lamp. Samples of their spectra are shown in Appendix A. All lamps illuminate the same diffuser which acts as the source for the comparison system. The three lamps are completely independent and can be used individually or in combination.
The slit assembly consists of two polished and aluminized jaws forming a bi-parting aperture adjustable over a range of widths between 5 and 1200 microns. To a width of approximately 500 microns, the jaws remain parallel; for larger widths some tapering of the aperture will be seen. This is due to the design of the mechanism and does not offer problems under normal circumstances. The knurled ring at the base of the micrometer clamps its motion; it should be fully counter-clockwise before adjusting the micrometer. At the du Pont telescope a 1 arcsec slit corresponds to 91 microns.
The decker mask is located immediately in front of the slit. It is motor driven in a direction perpendicular to the slit.
A removable filter holder for blocking filters is located just behind the slit and is manually positioned by a push rod on the side of the spectrograph opposite the slit and grating controls. Nominal filter size is 1.25 x 0.75 x 0.125 inches (32 x 19 x 3 mm). Data for available filters is given in Appendix B.
The shutter is a simple flag type for controlling the exposure. It is motor driven, and its position is sensed by microswitches. The shutter can be controlled either manually or via the CCD data acquistion program. Note that the shutter opens and closes fairly slowly, and is unsuitable for accurately timing exposures less than approximately 20 seconds (see Appendix C)
The collimator is located in the tube at the bottom of the spectrograph. For the du Pont telescope, the collimator is an f/7.5 off-axis paraboloid. This mirror has a clear aperture of 100 mm.
Motion of the collimator is used to focus the spectrograph through a range of +/-0.75 inches from its nominal position. A dial indicator on the side of the collimator housing reads the position in increments of 0.001 inches; the extreme position farthest from the slit is 0.210, and the nearest to the slit at 1.740. Attempts to exceed these limits can damage the focus mechanism. There is no collimator motion clamp.
The nominal auto-collimation position of the collimator was measured in March 2006 to correspond to a value of 0.965.
Several Bausch and Lomb replica diffraction gratings are available with this spectrograph, each mounted in its own cell and labelled as to grooves per millimeter, Littrow blaze wavelength, and mounting direction. The grating to be used is bolted into the grating turret which in turn is bolted into the spectrograph body. Do not attempt to change a grating unless you have been trained in the proper procedure.
The grating characteristics are summarized in Table 1. Because the spectrograph is not a Littrow design, the actual blaze wavelength is shifted (by a factor cos(1/2*49)=0.91) shortward of the value marked on the grating cell and used for identification. The linear reciprocal dispersion is given in Angstroms/pixel in Table 1, and the total spectral coverage with the Marconi#1 CCD is given in Angstroms. The spectral resolution will depend on the width of the slit employed: 2-pixel resolution corresponds to an ~2.1 arcsec slit on the du Pont telescope, with the precise value depending on the amount of anamorphic magnification (see Appendix D).
|600||6250(2)||0.80||1630||blaze = 1.25 µm|
|600||6133(3)||0.52||1060||blaze = 1.84 µm; vignetted!|
A two-leaf dark slide is located immediately in front of the camera corrector plate. Each leaf can be pulled separately, allowing light to enter half of the camera, thus permitting a type of Hartmann test for focus. In normal use, both leaves are pulled clear of the beam.
The camera is a semi-solid Bowen-Schmidt design with a focal length of 140 mm. The camera-collimator combination gives a demagnification of 4.79 on the du Pont telescope. All optical elements are fused silica, are factory aligned, and require no adjustment by the observer.
The detector currently mated to the Bowen-Schmidt camera is Marconi#1, a thinned, backside-illuminated Marconi 2048x515 CCD. The pixel size is 13.5 microns which gives a scale along the slit of 0.70 arcsec/pixel with the du Pont telescope. The total slit length is 270 arcsec. Peak Q.E. reaches 83% at 5000 Å, falling to 58% at 3500 Å and 21% at 9000 Å. (See Figure 2 for a plot of the Q.E.)
A characteristic of the Marconi#1 CCD is strong fringing longward of 7000 Å. This is illustrated in Figure 3 which shows a spectrum of the spectrophotometric standard LTT 2415 taken with the 600/7500 Å grating. This spectrum was taken through a 4 arcsec slit with the spectrograph somewhat out of focus; the resulting resolution is ~13 Å. It is recommended that observers working in this wavelength region take flats on the inside of the dome at each telescope pointing so as to be able to remove this strong fringing.
On the du Pont telescope, the spectrograph is controlled via either a Local Control Panel or a Microprocessor Controller. These interfaces are described in more detail in this section.
The Local Control Panel (see Figure 4) is mounted on the spectrograph body. Under normal circumstances it is used only for tests, and allows control of the shutter, the slit prisms, the decker mask, and the comparison sources.
The "Slit Length" control actually refers to position of the comparison prisms. The meter on the Local Control Panel shows the separation between prisms in millimeters. At maximum separation (25 mm), the prisms send no light down the spectrograph.
The decker switch moves the decker plate before the slit. It's position is indicated schematically on the meter, the pointer of which represents the slit. When the pointer is in zone A (see Figure 5), the obscuring mask blocks the central portion of the slit to a degree determined by the mask position. With the pointer in zone B, the mask has no effect and the slit length is determined by the separation of the comparison prisms. Most observers will want to leave the decker in this position for the during their entire run. Zone C was used as a moonlight eliminator when the spectrograph was used with photographic plates. In zones C1 and C2 the widths of the apertures are approximately 2.0 and 1.0 arcsec, respectively.
The total slit length is 270 arcsec.
The "shutter" switch on the control panel controls the shutter motor. The "open" and "closed" lights are activated by microswitches on the shutter mechanism and therefore correctly indicate the shutter position.
Three switches on the control panel supply power to each of the respective comparison lamps. The panel lamps show which sources have been selected; they do not indicate that the lamp is operating.
The Microprocessor Controller is operated from the data acquisition window. The Decker, Prisms, and Lamps controls are in the central section of the acquisition window, see Figure 7.
The "Decker Position" indicator controls the motion of the decker. For simplicity, the range of possible positions is represented by numbers in the range 0-100. The conversion between these numbers and the different decker zones shown in Figure 5 is given in Table 2.
|50||Tip of of Obscuring Mask|
|35||Middle of Open Position (Zone B)|
|14||Middle of Zone C1|
|0||Middle of Zone C2|
To move the decker to the desired position, click the left mouse button on the number in the command (left) box. Type in a new number, hit return, and the decker will start moving to the desired position. Alternatively, the decker may be moved to preset positions of either 0, 14, or 50 by clicking the left mouse button on the appropriate button to the right of the decker command window.
The comparison prisms ("Prisms") indicator functions in exactly the same manner as the "Decker Position" control. In this case, preset positions of 0, 25, and 100 may be entered by clicking the left mouse button on the buttons to the right of the command window. Note that a value of 100 corresponds to the maximum separation (25 mm) of the comparison prisms; minimum separation is a value of 0 (the normal setting when taking an arc exposure).
The comparison arc lamps may be turned on or off individually by clicking the left mouse button on the individual buttons. If desired, more than one lamp can be turned on at the same time. The "auto" button coordinates the turning on of the lamp(s) with the movement of the comparison prisms. When "auto" is set to "on" and a lamp is turned on, the comparison prisms will automatically be moved to position 0. When the lamp is turned off, the prisms will move to a position of 100.
At the du Pont telescope the slit viewing optics are built into the telescope Instrument Mounting Base (IMB). The spectrograph slit is inclined by 6 degrees to feed these optics. The pick-off prism can also be rotated to view the field directly. In slit-viewing mode the guider images 1.6 x 2.2 arcmin centered on the slit and in centerfield mode the guider images 2.3 x 3.1 arcmin.
Typing the command "bnc" on the du Pont data acquisition computers will start both the instrument Microprocessor Control window (see Section 3.2) and the CCD data acquisition window.
The "bnc" command first launches the CCD configuration window (see Figure 6). In most cases, the only thing the observer needs to change is the "Observer" parameter. It is possible to also select the number of horizontal and vertical overscan lines and columns that are stored for each CCD frame. The default (and recommended minimum) is 128 lines and columns. Click "Ok", this will start the CCD Camera command window (see Figure 7).
Figure 6 - CCD Startup Window
Loops: Setting up a loop can be a convenient way to take flat fields or make synoptic observations.
ExpType: The exposure type (saved in the fits header) can be set to "Object", "Bias", "Dark", or "Flat". Selecting "Bias" will automaticaly set the exposure time to 0 seconds. Selecting "Dark" will disable the shutter for the exposure.
Binning: The CCD can be binned by factors of 1, 2, 3, or 4 in rows and columns.
Subraster: The full slit length covers less than 300 lines of the CCD, so observers may want to read out only this region by defining a subraster. Clicking on the subraster button (labelled "Full" when the window is started) generates the CCD Subraster window. Each extracted subraster is defined in column number by the values of X1 and X2, and in row number by Y1 and Y2. Note that the values of X1, X2, Y1, and Y2 are in binned pixels if the CCD is read in a binned mode. The final subraster includes the full requested number of overscan columns. After entering the values for the subraster, the setup is entered by clicking the "APPLY" button and the window is closed by clicking on the "DONE" button. The setup can also be saved into a file and loaded from a file.
Start begins a science exposure. The exposure can be paused (e.g., for clouds) and resumed by toggling the "Pause" button, but observers should remember that cosmic rays are accumulating while the exposure is suspended. The progress of the exposure can be seen in the bar-graph immediately to the right of the ExpTime entry. This graph fills to the right during the exposure and as the CCD reads out the bar empties to the right.
The "Snap" button automatically sets the 4 x 4 mode, takes an exposure, and -- unlike the "start" command -- does not update the frame number at the end of the exposure.
For a "File" number of nnn the frame will be saved as file ccd0nnn.fits. The file number will be updated by one for all exposures except a "Snap". Note that if a frame with the same number as the current exposure already exists on the disk, it will be overwritten.
Abort: Exposures can be stopped or aborted. If the "Abort" button is clicked during an exposure the shutter will close and the frame read out. If the "Abort" button is clicked again during the readout during read out the data will be dumped and the file number not updated. A wipe of the chips follows, but since the user will not generally know when it is finished, it is recommended that a short "Snap" exposure be taken to insure a clean chip before the next science frame is taken.
Speed: The CCD camera has three readout speeds, Slow (the default), Fast, and Turbo. Readout times for a full frame plus 128 rows and columns are 21 seconds (Slow), 12 seconds (Fast), and 8.5 seconds (Turbo).
There are three function buttons at the top of the frame which open pull-down menus (click and hold the left mouse button). The "File" button has options to reset the CCD electronics (it is best to take a short "Snap" exposure after a reset) and to exit the window gracefully. The "Modules" button has options to start a large or small scale quick look tool that displays the image during readout. The "Options" button allows the observer to change the data path (which defaults at startup to the directory from which the program was launched initially). The "DataPath" window will create desired data directories if they do not exist. It also has options to start three engineering windows to monitor the CCD temperature, and to set the CCD voltage and readout times. These engineering features are password protected.
During normal night operations the Decker and Prism positions should be set to 35 and 100, respectively; and the Auto mode should be on (indicator will be green). This will automatically move the Prism to 0 position when the comparison lamps are turned on.
Auto mode should be off for focusing the spectrograph.
Gain and read noise values for the three read speeds are 1.0 e/DN and 3.1 e (Slow), 0.9 e/DN and 3.3 e (Fast), and 1.7 e/DN and 4.3 e (Turbo). The CCD is linear to 75,000 electrons. This corresponds to approximately 45,000DN in Turbo readout mode and to above digital saturation in the Slow and Fast modes. The dark current is less than 1 electron per pixel per hour at the operating temperature of -100 C.
With the slit oriented East-West the spectrograph flexes 3 pixels in the dispersion direction when the telescope is moved from declination -70 degrees to declination +10 degrees on the meridian. Observers should take calibrating arcs at each telescope pointing, and take particular caution when working to the red of 7000 Å because of the strong fringing at these wavelengths (see Section 2.9 and Figure 3).
IRAF is used to display the data and for simple real-time data analysis. DS9 is used for image display. These programs can be started from the file icon in the toolbar at the bottom of the observer's monitor. An appropriate IRAF "stdimage" parameter can be set by typing:
- set stdimage=imt2048
Running the package "bnc" in IRAF will load routines which will help to measure the CCD gain and read noise ("ccd_gain") and to focus the spectrograph ("fbnc" and "fplot" or "focusplot", see Section 6).
The spectrograph is focused by adjusting the collimator position. Focusing will normally be done by LCO staff; nevertheless, the basic steps in this procedure are covered in this section.
- Turn on the comparison lamp(s).
- Set the spectrograph slit width to 100 microns.
- In the Microprocessor Controller window, set the Comparison Prisms to 25, and the Decker Position to 14.
- Set the collimator focus to 850.
- Take an exposure with the CCD. Make sure that the count level is reasonable. (An exposure time between 20 and 300 seconds will be required, depending on the lamp used and the wavelength region observed.) The spectrum should look something like the one in Figure 8 - i.e., with four narrow spectra visible along the slit.
- Repeat for collimator positions of 950 and 1050.
- In the IRAF window, type "bnc" to load some scripts ("fbnc" , "fplot" and "focusplot") that are useful for focusing. The support staff will show you how to run these tasks. The basic idea is to measure the widths in both the x (along the dispersion) and y (along the slit) directions of three unblended lines in the spectra (e.g., one near column 500, one near column 1000, and one near column 1500). Do this with the "fbnc" task for the images corresponding to all five collimator positions.
- From these measurements, you should be able to choose a collimator focus which will give the best overall image quality in both directions. (The "fplot" and "focusplot" tasks are useful for this.) Set the collimator to this value, and take a final exposure to check the results. The best focus should be near the auto-collimation focus of 965.
If necessary the camera and CCD dewar can be rotated over small angles to adjust the alignment of the spectrum on the CCD. For example, to improve sky subtraction for single star spectroscopy, the spectrum should be oriented so that the slit is as parallel as possible to CCD columns. This should already be done, but the instrument specialist can help you to adjust the alignment. The final alignment can be checked by taking a comparison spectrum with the prisms at the center ("Prisms" = 0), and using the "implot" task in IRAF to measure the positions along the slit of 2-3 spectral lines.
Dome flats can be obtained by pointing the telescope to the white screen located inside the dome, which is illuminated by 4 quartz lamps controlled through the switch and knob in the blue rack at the control room. The telescope operator or the day time technician will point the telescope.
Dome flats should be taken for every slit width to be used at night.
Note that there is significant fringing at red wavelengths, thus, if working in this spectral region, it may be wise to obtain dome flats at night right after or before the object observation, by closing the dome and turning on the quartz lamps.
At sunset, it is also possible to obtain sky flats, if illumination calibration is desired.
Click here for an atlas of HeAr spectra taken with the internal comparison lamp.
At the beginning of the night, the Telescope Operator will focus the telescope on the spectrograph slit. The slit viewing camera is itself focussed on the slit using various features (scratches) on the slit jaws. The telescoe operator will acquire a guide star with the off-axis guide probes. Once the telescope is guiding, and the object is centered in the slit, an exposure can be started.
For programs which do not require a specific position angle for the orientation of the spectrograph slit, light losses due to differential refraction can be minimized by keeping the slit oriented close to the perpendicular to the horizon. This is accomplished by adjusting the IMB rotation angle. The exact angle to use will depend on the position of the object on the sky and the of the exposure time that is planned. Just before moving to your next object, ask your telescope operator adjust the IMB rotation angle (Cassegrain Ring rotation angle) to the optimum value which can be calculated with the IRAF task "bpangle". As a quick reference, when the Cassegrain Ring angle is 180 degrees, the Boller and Chivens slit is oriented in the East-West direction.
For many programs, observations of spectrophotometric standard stars will be necessary. Observers are referred to the following two sites for coordinates and finder charts of suitable standards:
Observers are encouraged to take their data home on DVD discs. A supply of disks is avaiable in the observer's room at the du Pont, each disk holds about 4.4 GByte of data. Instructions to burn the DVD's on the Linux computers Clarity or Caballo can be found at the LCO web page under:
The following Schott glass blocking filters are available for use with the B&C spectrograph. Click on a particular filter to see representative trasmission curves (reproduced with permission from Schott Glass Technologies, Inc.):
|BG38||2 mm, 3 mm||1 mm|
|WG360||1 mm, 2 mm||2 mm|
|GG400||3 mm||3 mm|
|GG420||2 mm||3 mm|
|GG455||3 mm||3 mm|
|GG495||3 mm||3 mm|
|0G570||3 mm||3 mm|
|RG610||3 mm||3 mm|
Tests performed on July 3, 2002 gave a timing error of approximately 175 milliseconds for a 2-second exposure, with the error constant over the field of view two within 1-2%. From these results, the following graph may be plotted of the ratio of the actual exposure time to the requested exposure time as a function of the requested exposure time. As may be seen, for exposure times greater than or equal to 17 seconds, the shutter timing is good to 1% or better.
Shutter precision for the B&C spectrograph plotted in terms of the ratio of the actual to requested exposure time versus the requested exposure time in seconds.
As a result, to achieve a given resolution, the slit can be widened as the grating angle is increased. Note that the demagnification along the slit (i.e., perpendicular to the dispersion direction) is a factor of 1 / 4.79 independent of the grating tilt.
To calculate the value of r for a specific grating tilt and the slit width which projects to 2 pixels on the detector, a Java Script is also available here.
The Anamorphic demagnification factor (dashed blue line), r, is plotted versus grating angle for the B&C spectrograph on the du Pont telescope. Also shown (solid red line) is the slit width (in microns) which projects to a 2-pixel line width on the CCD.
In the following graph, central wavelength is plotted versus grating angle for the B&C spectrograph on the du Pont telescope. To more precisely calculate a specific tilt (along with the anamorphic demagnification), please use the available Java script