This manual provides a description of the advanced preparations, setup, and operation of the WFCCD for multislit spectroscopy.
Preparation of the metal masks is covered in a separate document, which is still in preparation. In the meantime, users should contact Andy McWilliam (andy@ociw.edu) for information at least 2 months before their observing run.
The metal masks are fabricated on Las Campanas, and will be handed to you upon arrival. It is suggested that before mounting any mask, it be inspected at low power with the binocular microscope in the library. With a used engraving tool (kept in the aperture box), small remaining threads of brass or other junk can be easily removed manually. Place the blanks in envelopes until you are ready to mount them.
To mount a mask, orient the mask frame holder so that the curved
piece is away from you and the 8 small screws holding the 4 holddown bars are
up. There is then a single pin on the right border of the frame and two pins
on the lower border. (These pins define the position of the metal masks
precisely). Using a 3/32" Allen wrench, remove the 4 holddown bars.
The metal slit mask is should be oriented so the single notch bears against
the pin on the right side of the holder, and the two lower notches bear
against the two pins along the bottom side of the holder. Place the
4 holddown bars over the mask and screw them in. If the 8 holes
in the slit mask interfere with the screws, you may need to slightly enlarge
them with a small rattail file.
The PC which controls the aperture and grism wheels runs under
Windows 3.1. When the PC is brought up it will display a panel with "Monitor
& Control" as one button, and "Configuration" as the other. When clicking
on "Monitor and Control", the application window will overfill the screen.
Click and hold the left mouse button on the upper blue border and drag
to the left so you can see all 4 buttons for both the aperture wheel and
the filter wheel. Clicking on any of the 4 aperture wheel or filter wheel
buttons will move the wheel. The wheels can also be controlled from the
CCD data taking window (see section 7).
A higher-resolution grism sometimes referred to as the "HK grism" is also available. This grism has a resolving power of R~1,400 (roughly 1.3 Å/pix), corresponding to a FWHM of about 220 km/s. The undeviated wavelength is near 3700 Å. Spectral coverage depends on the location of the slit; however, wavelengths larger than ~5300 Å are inaccessible.
An echellette grism is now available. It provides a resolving power of R~3,000 (FWHM~100 Km/s), with spectral coverage from 3700 to 9000 Å, in three orders. An inter-order gap is present in the interval 6200-6700 Å. As with the HK grism spectral coverage depends upon location in the field. Presently, there is no cross-dispersion available, for the echellette so orders, or parts of orders, must be isolated with filters. To fit the echellette grism into the available space it was necessary to vingnette the beam; as a consequence its transmission is low.
See Appendix B for the further information on
the performance of these grisms.
Let the LCO staff know well in advance of your run (via the WFCCD instrument setup form) what items you wish mounted in the filter/grism wheel. The grisms have been carefully mounted in their cells; alignment of the grisms in their cells and the cells in the filter wheel is described in a separate technical document. It should only be done by one of the LCO staff who know how do to this.
If you are not planning to use more than one grism during your run, and have need for at most one of the 3" circular filters, it is suggested that you use the "standard filter wheel" complement. If you elect to do this, you can use the cookbook focus procedure for the spectrograph, focus using the two Hartmann masks. This complement has the filters arranged as follows:
In the ccd command window are menus for the Aperture wheel and Filter wheel. Selecting one of the 4 positions in each will move the wheel(s) to the correct position.
CAUTION: There may still be occasional problems with either the aperture or filter/grism wheels not properly seating in their detents. Whenever the position of either of these wheels is changed, it is a good practice to check the screen of the PC wheel controller program. A fault will be indicated by the particular button in question remaining "red". See section 4 for instructions on how to correct this problem.
To start up an "ximtool" window for displaying your data, use the right mouse button to select "IRAF" and then "ximtool". Then within the IRAF window once again, set the ximtool window to the correct size by typing:
Finally, to load the IRAF tasks that will be used at night for aligning the field and slitlet masks, type:
Next, the exposure time in seconds should be entered in the "extime" field. The object name may be entered in the "object" field, and the "imtype" button can be used to set the image type. Clicking on the "start" button will then begin the exposure.
Note: There is a new feature of the ccd control system which displays a small green window which says "abort" , during the readout of the chip. Do not be alarmed! this simply gives you the opportunity to abort the readout by clicking on this button during the readout. During readout, the CCD window is "locked" and no operations in it or changes in it can be made. For gain 3 and the default overscan of 16 lines, the total readout and storage time is ~70 seconds for the full chip.
b) A window will appear which lists column labeled x1 x2 y1 y2 for up to 8 subrasters, which are the boundaries for each subraster. Note that the y-values are not allowed to overlap for any of the subrasters.
c) Enter the name of the previously stored subraster file and click LOAD to load it.
d) Now click APPLY to enable the subrastering for subsequent readouts with the subraster button on.
e) Click DONE to close the window.
f) The files will be stored as small images with suffixes _nn where nn is the number of the subraster, in order of increasing y; e.g., if disk file 105 is being taken, and there are 4 subrasters then ccd105_01.fits... ccd105_04.fits will be stored.
The spectrograph will normally be focused by the day crew at the start of
your run. Should this not be the case, or should you feel that a refocus
is required, refer to section 2 of the Aligning
and Focusing the WFCCD Optics manual.
Alignment of the dewar rotation should also be carried out by the day
crew, but is also covered in the Aligning
and Focusing the WFCCD Optics manual.
For both grisms, wavelengths increase with decreasing column
numbers. The wavelength solution is not well approximated
by a linear fit. Use both HeAr and Neon lamps to get a good solution.
Not that if the ring angle is substantially changed during the night, this procedure should be repeated.
The telescope should be focused in the normal way by using the "focus"
function in the ccd command window. Enter some number (e.g., 7 or 9) in the
loop field, set the exposure time, and then use the focus button in the ccd
command window (see the
CCD manual for more details).
a) Make sure you bring finding charts with you in order to identify the alignment stars!
b) During the day you may wish to take an undispersed image of each of the masks that you plan to use that evening. You can do this at the zenith, but if you are well enough organized to know the approximate HA when you will observe your field, go to that target HA and Dec. Now run the IRAF task
You will be presented with a parameter file which should be edited to look something like the following:
mask = "ccd004" Image of slitlets sub_file = "ton180.sub" Output file of subraster coords (xsz = 15.) Full size of boxes in x (pixels) (ysz = 15.) Full size of boxes in y (pixels) (cursor = "") (line = "") (mode = "ql")The mask parameter corresponds to the names of the slitlet mask image, and sub_file is the name that you wish to call the subraster file. The hidden parameters xsz and ysz refer to the sizes in x and y of the alignment boxes; it is important that these be set to the correct values for this program to function properly.
Once the parameters for mksubr have been set, the task should be executed. The program first displays the slitlet mask image mask and then asks the user to mark the approximate locations of the alignment boxes by centering the image cursor on them and hitting "space bar". The intensity mapping of the image display can be changed manually by hitting the "d" key. The "r" key should be hit if a mistake is made and the user wishes to remark the alignment boxes; to abort from the program completely, hit "I" (not "control-c"!). Once the last box has been identified, hit "q". Precise coordinates for the box centers will be written to the screen, and the subraster coordinate file subr_file generated. Each line of this file corresponds to a subraster; the format for each line is "x1 x2 y1 y2", where x1,y1 is the lower lefthand corner of the subraster and x2,y2 is the upper righthand corner. The subraster coordinates are ordered in increasing y values, as required by the CCD data acquisition program.
Note that this step can be skipped if one is willing to put up with the time required (~70 seconds) to readout the entire CCD during the alignment procedure.
c) Select the appropriate mask in the CCD command window to put the mask for the field in question in the beam.
d) Slew to the position of your field and go about 1/2 hour further west (adjust this when you have some experience of how long it takes you to set up).
e) Have the dome turned so you can take a dispersed image off the dome with the flat field continuum lamp rheostat turned to max (500). Try about 90 seconds for the exposure time.
f) Remove the grism from the beam and take a short (~30 seconds) undispersed mask frame.
g) Acquire your field and find a guide star. The parameters on the Shectman TV guider, some of which are specific to use of the guide probe (probe #1) (not 'center field' , which is really guide probe #2) for the WFCCD are: pf=inv, xf=nrm, yf=nrm, px=222, av=1, gn =4, and bx=15. The "pa" parameter should be set as follows: pa = cass ring angle - 180 degrees. These are the correct parameters; some entries in the guider manual are not correct. Look for a guide star, moving the X,Y directions with the guide probe box. Do not increase the Y value above about 2100, else the probe will run into the edge of the WFCCD mounting base. A limit switch has yet to be installed. Try to find a guide star with X between about 4500 and 6500 and Y between about 1500 and 2000-- this range puts the guide probe as near to the optical axis of the telscope as it can get. If you go to X near 0 or near 11000 and Y near 0 the images get very bad because you are so far off axis. Focus the guider lens using the in/out buttons on the guider chassis and selecting "1" with the toggle switch. On the guider PC keyboard, F3 starts the guiding, F1 turns off the signals to the telescope, while F2 lets you see how well it thinks it can guide. See the Guider manual for further details.
h) With the guider guiding, no grism in the beam, no mask in the beam and pointed at your field, take another short exposure (~30 seconds; adjust this depending upon how bright the 4 or 5 alignment stars are). Once you have a good exposure, type:
You will be presented with a parameter file which should be edited to look something like the following:
mask = "ccd005" Image of slitlets field = "ccd006" Image of field box_file = "ton180.box" Output file of box positions ring = 180. Current Cass ring angle xtvold = 132. Current x-coord of TV guider ytvold = 123. Current y-coord of TV guider subr_file = "" Input file of subraster coords answer = no Skip procedure to find box coords? (xsz = 15.) Full size of boxes in x (pixels) (ysz = 15.) Full size of boxes in y (pixels) (xrot = 915.) x-coord of rotation center on CCD (yrot = 1005.) y-coord of rotation center on CCD (chippa = 0.) dewar alignment angle (camscl = 0.77) CCD scale (arcsec/pixel) (tvscl = 0.227) TV guider scale (arcsec/unit) (fwhm = 2.) star FWHM (pixels) (ncols = 2064) Number of columns in full image (nrows = 2048) Number of rows in full image (cursor = "") (fbox = "") (line = "") (mode = "ql")The mask and field parameters correspond to the names of the slitlet mask and field images, respectively; box_file is the name you wish to call a file of precise positions of the slitlet mask alignment boxes which will be output by the task; ring is the current Cass ring angle; xtvold and ytvold are the current x and y coordinates of the TV guider. The parameter answer should be set to "no" the first time through this task.
The align task compares the slitlet mask and direct field images to measure the correction to the Cass ring angle and telescope offsets required to align the boxes of the slitlet mask image with the corresponding alignment stars in the direct field image. The program first checks to see if the file box_file, which contains precise coordinates of the alignment boxes as measured from the slitlet image, exists. If box_file does exist, then the program will ask whether you wish to use this file and to skip the procedure to measure coordinates for the alignment boxes. If the answer is "no", or if box_file does not exist, the program displays the slitlet mask image, and then will ask you to mark the approximate locations of the alignment boxes by centering the image cursor on them and hitting "space bar". The intensity mapping of the image display can be changed manually by hitting the "d" key. The "r" key should be hit if a mistake is made and the user wishes to remark the alignment boxes; to abort from the program completely, hit "I" (not "control-c"!). Once the last box has been identified, hit "q" to move on to the next part of the program. The findbox task is called to measure precise centers for the alignment boxes, and these are written to the file box_file. (See the IRAF help page for findbox for details on the algorithm employed.)
At this point, the field image is displayed. The positions of the alignment boxes are displayed as yellow squares as an aid to locating the alignment stars. You will be asked to mark the position of the first star by hitting "space bar". This causes orange circles to be displayed at the predicted positions of each of the alignment stars. Finally, you will be asked to mark each alignment star, one by one. Here there are two options: If "space bar" is hit, the imcntr task is called to automatically calculate the stellar center. This is generally the preferred option, except in cases of crowded fields, in which case the "m" key can be used to manually mark the center of the star. As before, the "d" key allows the user to change the intensity mapping of the image display, the "r" key can be hit to remark the positions of the alignment stars (beginning with star 1), and the "I" key can be used to abort the program.
After measuring the positions of the alignment stars, the cstar task is called to solve for the differential rotation and shift required to bring the slitlet mask and field images into alignment. This task first checks the relative scales of the stars and boxes and prints this information to the screen. Ideally, the scale ratio should be nearly exactly equal to one. Next, the program solves for the differential rotation and x-y shift required to center the stars in the alignment boxes. The rotation is specified in terms of the Cass ring angle, and the shift is given both as NS-EW telescope offsets, and as new x-y coordinates of the TV guider window. (See the IRAF help page for the cstar task for further details.) If the offset is fairly large, it will be necessary to move the telescope and find another guide star; if it is small, the offset can be made by changing the x-y coordinates of the present guide star. Note that the required adjustment of the Cass ring angle is given in degrees, minutes, and seconds for use with the differential encoder box.
i) With the guider guiding, take another short exposure of the field and then run the align task again, updating the field parameter to the name of this new image. Since this time box_file should exist, answer "yes" when asked if you want to use this file. The new field image will be displayed; follow the instructions given above for marking the alignment stars. Offset the TV guide probe and Cass ring angle as required.
j) If all has gone well, the alignment stars and boxes should now be sufficiently well aligned that the slitlet mask can be put in, a new exposure (with the grism still out of the beam) taken, and the alignment stars can be observed through the alignment boxes of the mask. At this point, the falign task should be run. Type
and you will be presented with the following parameter file:
mask = "ccd008" Image of slitlets ring = 180. Current Cass ring angle xtvold = 164. Current x-coord of TV guider ytvold = 129. Current y-coord of TV guider box_file = "ton180.box" Input file of box positions subr_file = "" Input file of subraster coords (xsz = 15.) Full size of boxes in x (pixels) (ysz = 15.) Full size of boxes in y (pixels) (xrot = 915.) x-coord of rotation center on CCD (yrot = 1005.) y-coord of rotation center on CCD (chippa = 0.) dewar alignment angle (camscl = 0.77) CCD scale (arcsec/pixel) (tvscl = 0.227) TV guider scale (arcsec/unit) (fwhm = 2.) star FWHM (pixels) (ncols = 2064) Number of columns in full image (nrows = 2048) Number of rows in full image (line = "") (fsub = "") (mode = "ql")This task first measures precise centers for the alignment boxes using the same algorithm used in findbox (see the IRAF help page of findbox for details). The previously-generated input file "ton180.sub" listing positions of the alignment boxes is required. Next, the coordinates of the stars are measured via marginal distributions. You will then be presented with the marginal distribution plots in x and y for the first alignment star. Centroids and the sky level calculated from these distributions are displayed in the same graphs. You may modify these if desired -- type "?" in the graphics window to list the possible commands. To move on to the next star, type "q" in the graphics window.
After measuring the positions of all of the alignment stars, the task solves for the differential rotation and x-y shift required to center the stars in the alignment boxes, and prints this information to the screen as described above for the align task. At this point, the rotation should normally be sufficiently close as to not require adjustment; the offsets should also be small, and often will not require further adjustment. If you decide to make further adjustments, run the cstar task again to check the alignment.
k) Once the slitlet mask is well aligned with the field, insert the grism in the beam. You are now ready to make an observation. If you are going to stay on the field for an hour or more, you should periodically check the alignment of the mask by removing the grism from the beam, taking a direct exposure through the mask, and running the falign task.
l) Note that in steps "f" and "j" (and perhaps also in "i"), considerable time can be saved by taking subraster images instead of a single full format image. The subraster images can be input directly into the align and falign tasks by using file templates (e.g., set the mask parameter in align to ccd005*. You will be asked to give the name of the subraster file when running either of these tasks on subraster images. (See the IRAF help pages for align and falign for further details.)
m) The WFCCD has flexure which we have not yet tracked down and corrected, nor mapped completely. Until this is has been done you should:
a) Try to limit your exposures to 40 minutes or less, otherwise you will degrade the spectral resolution somewhat.
b) Take the mask alignment exposures at about the same position in the sky where you will begin the series of (40 minute) exposures on the field in question.
c) After each 40 minute exposure, you might want to check the alignment of the stars and boxes by taking a short, undispered image without removing the slitlet mask. This check can be done rapidly by taking subraster images, but don't forget to turn the subraster function off before starting the next 40 minute exposure!
Slit masks are fabricated from brass, or aluminium, plates with a computer-controlled milling machine at Las Campanas. The milling machine code is generated from sky positions by use of program "maskgen.e", which is available on the Carnegie Observatories ftp site.
Input to "maskgen" includes id number, X (arc sec), Y (arc sec), priority, magnitude, Position angle. Input is in free format; data beyond the sixth column do not affect operation of the program. Note that positions are relative to the field center, with positive X to the East and positive Y to the North. The slit PA, measured in the usual sense clockwise from North, is only used if the user requests tilted slits in the interactive session. An example input file is shown below.
Sculptor 0, mask 5 ID X(arsc) Y(arsc) Priority Mag. P.A. 9028 -487.03 175.15 -1.00 13.42 0.00 9055 -544.44 -309.28 -1.00 14.68 0.00 9031 382.77 137.40 -1.00 13.61 0.00 9025 275.20 -270.04 -1.00 13.30 0.00 10001 0.0 100.0 0.0 13.0 0.00 10002 260.0 100.0 0.0 13.0 0.00 10003 520.0 100.0 0.0 13.0 0.00 44 -650.83 11.46 16.90 16.90 0.00 43 607.59 -10.60 16.95 16.95 0.00 56 519.86 -152.24 17.11 17.11 0.00 64 462.51 -384.99 17.14 17.14 0.00 100 -94.16 159.28 17.26 17.26 0.00 98 378.38 262.68 17.31 17.31 0.00 111 37.34 -625.07 17.32 17.32 0.00 79 -537.72 -238.84 17.33 17.33 0.00 96 89.10 292.48 17.35 17.35 0.00 134 -279.86 -190.27 17.37 17.37 0.00 84 49.52 44.83 17.37 17.37 0.00 144 277.17 -67.92 17.39 17.39 0.00 82 -609.85 38.28 17.39 17.39 0.00 110 -282.72 360.20 17.40 17.40 0.00 129 -277.62 412.20 17.43 17.43 0.00 135 97.93 -506.50 17.45 17.45 0.00 161 22.00 184.44 17.46 17.46 0.00 127 341.55 -40.79 17.50 17.50 0.00 119 83.25 -83.31 17.51 17.51 0.00
An arbitrary number of comment lines can be placed at the top of this file; the user is queried for the number. Notice the negative priority of the first four objects, which indicates alignment stars; they take priority over all porogram objects.
The next three objects with priority 0.0 are optional, but recommended if holes are to be used rather than slits; these objects are part of a grid of sky holes.
Finally the list of program objects. Notice that the priority
and magnitude are the same. Since smaller numbers have higher priority
the program will choose the brighter object when clashes occur.