Skip to content. | Skip to navigation

Personal tools
You are here: Home Telescopes Irenee du Pont Instruments Website ModSpec Manuals

MODULAR SPECTROGRAPH

 

INTRODUCTION to OPERATING the MODULAR SPECTROGRAPH

 

2.02 nd Edition
October 1997
William E. Kunkel

 

Editor's Note

 

This is a first step in preparing a WEB RESOURCE collecting together the operating manuals of Las Campanas instrumentation, as well as the reports from observers describing their experiences, that might prove relevant to future users. The multi-dimensional stucture of WEB links permits adding cross referencing simplifying user access to information on related topics, such as test results on detector quantum efficiency, curvature of detector surfaces, reports on measured throughput, variation of mechanical flexure with hour angle and declination, data processing methodologies, and whatever other information that the user community deems of interest.
The success in compiling this information depends on the collaboration of observers in the reports they provide. Contributing information should be e-mailed to the editor (currently: kunkel@ociw.edu, flagged with "-s manuals"), giving references to the instrument configuration, dates of observation, reduction procedures (when relevant), and the like. The multi-dimensional structure made possible by WEB links will allow for the inclusion of a far greater variety of information than has been possible in the past. Conceivably even descriptions of repair and maintenance activity may be included, in the sense of appendices in the form of a logbook. Suggestions on how to make this database more useful will be warmly welcomed.
It is suggested that users make no attempt to modify the manual files by themselves. Depending of the success achieved, the web-structure will acquire some considerable complexity, and excercising adequate control over this would become more problematic if organization were haphazardly upgraded. Cooperation on this point is important, and will be much appreciated.


 

TABLE of CONTENTS

  1. Introduction
  2. Data Acquisition
  3. "Quick Look" data processing
    • Setup
    • Usage
      • For bright objects
      • For faint objects
  4. Focusing the Spectrometer
  5. Focusing the Telescope
  6. Flat Field procedures
  7. Precautions worth considering

 

APPENDICES

  1. Rebooting ...
  2. Grating Wavelength Tables
    • B.1 The 85mm camera
      • SITe2: 150, 300, 600, 831, 1200 l/mm gratings
      • TEK1: 150, 300, 600, 831, 1200 l/mm gratings
    • B.2 The 200mm camera
      • SITe2: 150, 300, 600, 831, 1200 l/mm gratings
      • TEK1: 150, 300, 600, 831, 1200 l/mm gratings
  3. Logsheets describing previsouly used SETUPS
  4. Descriptive Portions of the Original (1988) Manual



    1 INTRODUCTION

    The MODular SPECtrograph (MODspec) was designed and built in 1987-88 for use with thinned the Texas Instruments 800 x 800 CCD (using 15u pixels) on the duPont 100-inch telescope, though it is currently also available at the Swope 40-inch telescope. Its design and fabrication centers on an organization into distinct modules that permit operation in a broad variety of configurations. Some of the possible configurations, such as the multi-slit mode, have not been extensively used. This manual incorporates extended information based on the experience of previous users, upgrades implemented since its conception, and includes the replacement of the TI CCD with a thick 1024 x 1024 CRAF CCD (using 12u pixels), and more recently the 1024 x 1024 TEK1 and the 2048 x 2048 TEK5 CCD (using 24u pixels). Since early 1997 observers have also had a new spectrographic format 532 x 1752 SITe2 available, (using 15u pixels), which currently offers the lowest readnoise (near 3.2e). This chip has a slitlength of 189 arc-sec on the plane of the sky. Its 0.35 arc-sec pixels have led some observers to bin their data by 2 along the slit.

    The spectrograph has two camera ports, which in the current configuration hold a CANON 85mm f/1.2 lens inclined at 31 degrees to the collimator axis, and a CANON 200mm f/2.8 lens inclined at 27 degrees to the opposite side of the collimator axis. The 15u pixels of the SITe2 chip gives scales of 0.83 and 0.34 arc-seconds per pixel, respectively, at the duPont telescope and 1.75 and 0.75 arc-seconds, respectively, at the Swope. The TEK1 chip, with its 24u pixels, gives scales of 1.32 and 0.56 arc-sec per pixel at the duPont telescope, and 3.5 and 1.5 arc-sec per pixel, respectively, at the Swope.

     

    The gratings available with MODspec are the same ones offered with the Boller and Chivens spectrometer. These are





    Gratings available with the Modular Spectrograph
            Dispersion -- A/pixel
    grating groove blaze TEK1 & TEK5 SITe2 Comments
    l/mm angle (order) 85mm 200mm 85mm 200mm
    _____________________________________________________________________
    Reflection Gratings
    150 2.15 5000 18.8 8.0 9.4 4.0
    300 4.3 5000 9.4 4.0 4.7 2.0
    600 8.6 5000 4.7 2.0 2.35 1.0
    600 13.0 7500 4.7 2.0 2.35 1.0
    600 22.0 6250(2) 2.35 1.0 1.2 0.5 blaze = 1.25u
    600 34.0 6133(3) 1.6 0.67 0.8 0.33 vignetted!
    832 19.7 8000 3.4 1.44 1.7 0.72
    1200 17.5 5000 2.35 1.0 1.2 0.5
    1200 26.7 7500 2.35 1.0 1.2 0.5
    100 26.7 8800(10) 2.8 1.2 1.4 0.6 Echellette
        5500(16) 1.76 0.75 0.88 0.38 Echellette
    150 37.9 8200(15) 1.25 0.53 0.63 0.27 Echellette
        6168(20) 0.94 0.4 0.47 0.2 Echellette
    Transmission Gratings (Mirror in grating rotator)
    300 17.45 5200 8.5 3.6 4.2 1.8 center 6500
    200 15.0 6730 13.3 5.7 6.7 2.8 center 7000

    The collimator is a 429mm f4.3 UK50-fluorite cemented triplet (designed by Jim McCarthy) that produces a 61.3mm collimated beam. It has very little chromatic aberration, while the camera focal lengths vary by as much as 1mm between 4600 and 10000 Angstroms. Perhaps half a man-year of effort would be required to design and oversee the fabrication of comparable camera lenses incorporating FK52 or FK54 glass, which might be expected to reduce this chromatic aberration by a factor approaching five.

     

    The linear portion of the chromatic aberration is compensated by mounting the CCD detectors on plates that can be tilted (see Section 4). While cumbersome to manage initially, adjusting the plate inclinations is quite straightforward, though observers are advised to allow several hours at the start of observing runs to perform this adjustment, particularly when contemplating configurations that have not been used before and for which no data were recorded in the logging section of this manual (Appendix C). Grating changes during the night, requiring a re-tilting to a previously known setting and consequent re-focusing of the spectrograph, are somewhat time consuming; at least an hour must be allowed for this operation. When repeatable flat-field exposures must be relied on, such changes are not recommended. The spectrograph can be used as a "focal-reducing" camera, replacing the gratings with a plane mirror. Other utilization modes offer objective prism modes via zero deviation "grisms" (of 200 and 300 lines per mm) installed in the parallel beam; a multi-order "echellette" mode relying on two high order immersion gratings cross-dispersed with one of these grisms; and last, the slit mechanism may be replaced by a multi-aperture mask giving (almost) unvignetted coverage over a 35mm diameter field, with a useful maximum field approaching 50mm diameter. Appendix D contains portions of the original 1988 manual describing these options.

     

    The gratings utilized with the spectrograph are those employed in the Boller and Chivens spectrograph and the Multi-Fibre spectrograph designed for a 90mm light bundle.

     

    Control of the spectrograph is divided between two locations (see Section 2). Adjustments done at setup: slit width and length (decker), as well as CCD tilts and focus, are manual, and are carried out directly at the spectrograph. Control during the observing process: of the comparison lamps and the comparison mirror are via a PC clone computer, while the data acquisition is controlled via the data acquisition Sparc II computer, and is essentially the same as that employed for direct CCD imaging. See the CCD imaging manual for fuller details.

     

    Quick look reduction relies on the IRAF packages in the Sparc II (see Section 3). These procedures are functionally somewhat different from those used with direct imaging. Data are currently recorded on Exabyte casettes (March 1996).

     

    2 DATA ACQUISITION

     

    Time and thought are needed to set up the instrument in a configuration that will yield the most useful possible data. The evaluation of alternative setups depends on controlling the computer-operated spectrometer functions, and so reading of this chapter on the data acquisition process, and the next on quick-look data evaluation is recommended.

     

    2.1 Comparison Lamps

     

    The control of comparison lines and the comparison mirror is via a PC Clone running a program written by Nick Shectman.

     

    After re-booting the PC clone, the program is initiated by typing the number corresponding to GOMOD; the user enters: 6<cr>. The program inquires if the collimator is connected. Answer: y<cr>, and enter any number for the collimator setting. The collimator portion of the program has not been used.

     

    The monitor now shows the operating screen, in which the top line shows the available menu options. The <F1> key will list all the available commands. The <esc> key returns to the operating screen. The principal commands used are:


    comparison mirror
    to turn function
    ON OFF Action
    =============================
    hc1 hc0 Fe hollow cathode lamp
    he1 he0 helium lamp
    he1 ne0 neon lamp
    in1 in0 incandescent lamp
    cm1 cm0
     

    The blocking filter inserted into the optical path below the slit is selected by typing fl1<cr>, fl2<cr>, ... fl6<cr>, for any of the six available filters.

     

    The communication lines to the spectrograph go via a RS 485 line that is on occasion subject to some minor noise. When this happends, the PC Clone may "beep" and show error messages at the input cursor. These are readily cleared by striking the <enter> key. The only message that may not clear this way is one informing that "power may have been lost". When this happens assistance from an observatory technician will be required.


    2.2 Image acquisition

    On the SPARC II computer, besides launching IRAF in an xterm window, the user should expect to find the CCD CONTROL window, which is divided into four regions organized vertically from top to bottom. The beta version of March 1996 has been replaced by the definitive fully implemented version (October 1996)..

     

    The top group contains eight buttons, which perform as described below:

     

    START: The uppermost button of the left column STARTs an exposure with a left click.

     

    RUN MACRO: the middle button of the upper row, allows running a sequence of operations from a script. There is as yet no experience with this feature.

     

    SNAP: at the top right, the SNAP takes an image binned to a 512 by 512 format to permit quick surveys with large format chips. The frame counter is not advanced during a SNAP. The user is cautioned that this feature may not perform properly when subrastering is set, unless the observer sets the row and column starts and ends to the properly reduced values. Read times with the MODspec's 1024 x 1024 CCD chip are so short that this feature offers little practical advantage.

     

    FOCUS: This offers no useful function for spectrographic operation. (The left-most button of the second row allows focusing a telescope for direct imaging. The number of exposures set in the LOOP counter may be combined on a frame, spaced by a partial read of several dozen rows.)(see, however, multiple and focusing the telescope below.

     

    MULTIPLE: (the middle button of the second row) allows putting several exposures on the same frame; the number of exposures is determined by the value put in the LOOP counter (see below). This feature may prove useful when focusing the telescope (see below).

     

    PAUSE: the right button of the second toggles a halt in an exposure (closes the shutter and stops the clock), and restarts the exposure and the clock.

     

    ABORT: The left button of the third row, the ABORT, stops an exposure IF the read has not yet begun. Once a read has begun, ABORT has no effect; the read must be allowed to complete. Note however the small ABORT window appearing in the lowest pannel of the acquisition window. Left clicking this button aborts a read. Experience of some observers suggests that taking a SNAP exposure, by reading the entire chip, wipes any residue left by the incomplete read.

     

    ABORT/READ: the middle button of the third row, ABORT/READ terminates an exposure and reads the frame.

     


    The second group of buttons consists of eleven fields.

     

    Disk file: At the top of the left column is a field to set the file count (click left in the field to erase and rewrite the new frame number). A frame of count NN is written as "ccdNNN.fits" to the disk, where NNN is the frame number. If the SAVE button in the third group of the window is set, the frame count advances by one after each read is completed. If instead the TEST button is set, the frame count does not advance on terminating a read.

     

    Loop: Below the frame count is the LOOP COUNT; a number entered in the leftmost of two fields sets the number of exposures to be taken in a sequence. The field to the right shows the number of exposures already completed. The primary use of this feature is to automate exposing a sequence of flat-field and bias frames.

     

    Object: The third long field allows entering a string into the OBJECT field of the IRAF header.

     

    Filter: The fourth field of the left column labeled Filter has no function in spectrographic operation.

     

    Imtype: The fifth field allows selecting the IRAF image type from a menu.

     

    UT: The last field at the left in this group is the UT CLOCK, which is read from the operating system.

     

    Extime: The upper field at the right allows entry of exposure times in seconds.

     

    Actime: The field immediately below shows the elapsed time portion of an exposure.

     

    R.A., Dec.: Below the object field, the two fields at the right take the RIGHT ASCENSION and DECLINATION, which must be entered from the keyboard. To be consistent with the IRAF format, these should be entered as HH:MM:SS.S and sDD:MM:SS, (s is the sign) respectively. Digits running off the left edge of the field are retained, even if not shown in this window.

     

    Date: The U.T.date field at the lower right must be entered from the keyboard.

     

    The third group of buttons allows opening and closing the shutter manually (toggled), starting and ending the auto-wipe (also toggled), to enable (SAVE) or disable (TEST) advancing the frame counter on completing a read, and to set the binning with which a frame is read. The SUBRASTER controls allows reducing the portion of a frame written to the disk. The user must enter the starting and ending values of the rows and columns. If rows or columns are not to be subrastered, the limiting values must be entered (1 and 1024). Read times are somewhat reduced only if subrastering reduces the number of rows. Alternatively an observer may save disk space with the IRAF imcopy package to strip off the unwanted portion of a frame. So, for example, the command: imcopy ccd176[396:512,*] ccd176<cr> discards the first 395 and the last 512 columns of frame ccd176.fits. Similarly: imcopy ccd218[*,482:581] ccd218<cr> discards the first 481 and the last 443 rows of frame ccd218.fits.

     

    BINNING: a frame may be binned during read by setting a 1 or a 2 in the XBIN field to bin along columns, and in the YBIN field to bin the rows.

     

    MACRO: a command script may be entered here. Currently there is no experience with this feature.

     

    GAIN: The field at the lower left allows selecting gains of 1, 2, and 3 from a menu. The actual gains on read are approximately 1.15, 1.78, and 3.6. The readnoise is least with a gain of 1, and with TEK1 is typically near 5. This should be preferred for most spectrographic work.

     

    The lowest portion of the window reports the status of operations currently active, or of errors encountered during the setup of the aquisition program.

     

    3 QUICK-LOOK DATA PROCESSING

     

    Observers familiar with the IRAF data processing program know a number of ways with which to examine image frames, all of which serve their purposes. The following section supposes a minimal familiarity with IRAF, and suggests several approaches with which to evaluate data and test images acquired by the spectrograph.

     

    Since the spectra imaged by the transmission optics of the spectrograph are subject to geometric distortions, the IRAF apall or apsum packages may be used to trace stellar spectra and produce sky-subtracted plots. The setup and operation of this feature described below uses the apall package.

     

    3.1 SETUP

     

    The spectrographic IRAF processing packages must be loaded; enter the three that are needed: onedspec<cr>, twodspec<cr>, and apextract<cr>. If spectra lie along rows enter: dispaxis=1<cr>, if along columns: dispaxis=2<cr>.

     

    The processing parameters are set editing the apall parameters. If there are doubts that a previous user may have left these in disarray, enter: unlearn apall<cr>. Then enter: epar apall<cr>, and with <down> (or <up>) arrow examine the parameters, resetting those indicated here. Parameters that are not mentioned here may be left as they are. To change a parameter, move to it with the up or down arrow, type in the new value, then move to the next parameter. The <cr> key is not used. The editing process is terminated with a <^d> (control-<d>).

     

    Parameters to set or modify:

     

    First group

    format: multispec
    set to YES: interactive, find, recenter, edit, trace, fittrace, extract, and review.
    set to NO: extras

    Default aperture parameters:

    lower: -2.5 ; use larger values if seeing is bad.
    upper: 2.5
     

    Default background parameters:

     

    b_find: chebyshev
    b_order: 1
    b_sample: -8:-4,4:8; the two stripes should clear aperture
    b_average: 1
    b_niter: 1

     

    Automatic finding and ordering parameters:

     

    nfind: 1
    minsep: 25

     

    Tracing parameters:

     

    t_nsum: 25
    t_step: 25
    t_nlost: 3
    t_niter: 1

     

    Extraction parameters:

     

    background: fit
    skybox: 1
    weights: variance
    pfit: fit2d
    clean: yes
    readnoise: see text
    gain: see tex

     

    For TEK1 the gains 1, 2, and 3 are 1.15, 1.76, or 3.6, and the corresponding readnoise for these gains are 5.1e, 6.5e, or 7.0e, respectively. For the CRAF chip the gain is 1.45 and its readnoise is 5.3e. For the SITe2 chip, where the preferred spectrographic gain setting is 1, the measured gain is 0.89 with a readnoise of 3.2e.

     

    3.2 USAGE

    The spectral reduction package cannot operate on fits frames. Fits frames are converted of IRAF imh frames by entering, by way of example for frame 117: rfits ccd117.fits<cr> then answer: ccd117.imh<cr> to the output file query. The extensions .fits and .imh must be typed in full. The offset and scaling parameters of the rfits package should be disabled, and value short for the pixel type parameter.

     

    The processing should be kept interactive. Cosmic rays in longer exposures of faint objects occasionally require manual interventions.

     

    For bright objects:

     

    After converting a fits frame to IRAF .imh format, start the reduction by typing: apall ccd###<cr>, where ### is the frame number. Three questions asked with the default response (yes) should be answered with a <cr>.

     

    The first graph shown is a section across the spectrum (along the slit). The find routine shows its selection. If the selection is correct, a <q> will continue to the next step. If the selection is not appropriate, striking <d> will remove the selection, after which placing the cursor on the appropriate feature, striking the <n> (for "new") will enable the new choice; now <q> will advance to the next step. Responding with a <cr> enabling default yes answers continues to the next step. If by chance apall was started on a comparison spectrum, substituting no to these questions allows aborting apall.

     

    The second graph shows the location on the chip where the spectrum was traced. Not infrequently a hot pixel or cosmic ray will introduce unacceptable deviations. When this happens, placing the cursor on the bad points and striking <d> will throw these out. The fit can then be improved by striking <f>. The routine is exited striking <q>. Now apall asks if this information should be written to the database, to which a <cr> response will answer with the default yes. The last default yes is to review the spectrum. If the spectrum is of a sufficiently bright star, the routine will advise completion by asking if the spectrum should be shown; a <cr> here will produce the desired graph. If the spectrum was too faint, the extraction routine may fail, so indicating by a variety of error messages. The spectrum may still be created as shown below for faint objects.

     

    The third graph is the raw extracted spectrum. If the object is faint, or the exposure long so that cosmic rays dominate the scaling, that scaling may show the ordinate unusably compressed. The plot should be exited with a <q> and a <cr> response to the question of whether to write the plot to disk. The spectrum may now be examined with IRAF's splot package by typing: splot ccd###.ms<cr>. One should note that the extracted one-dimensional spectrum file is specified by the .ms following the frame number in the file name. The splot package allows a very broad variety of manipulations that the user should peruse typing: help splot<cr>. For example, the compressed scaling of the ordinate is compensated by placing the cursor at the desired upper limit of a replot and typing: <w> <t>, and similarly, setting the cursor at the lower limit and typing: <w> <b>. In the presence of excessive noise boxcar smoothing is obtained by typing: <s> and entering the smoothing width in pixels followed by <cr>.

     

    For faint objects:

     

    The procedure for faint objects is almost the same as given above for bright ones; only two parameters must be modified. An object that is too faint can no longer be extracted without relying on the aperture found shortly beforehand for a bright star. Secondly, the fittrace parameter must be disabled. Both are achieved by calling apall with: apall ccd### refer=ccd%%% -fittrace<cr>, where ccd%%% is the reference spectrum of the bright object reduced beforehand, and the minus sign before fittrace disables that function. NOTE: especially when working in the blue, where atmospheric dispersion displaces different wavelengths to different locations along the slit, the reference spectrum curvature may vary by more than five pixels along dispersion. So it is desirable that a reference spectrum taken at comparable airmass and slit orientation be employed.

     

    4. FOCUSING THE SPECTROMETER

     

    This spectrgraph relies entirely on transmission optics for all but the dispersion of light. Consequently the orientation of the curved focal surface varies with each instrumental setup, and properly adjusting the instrument for minimal imaging errors involves not merely moving the detector or collimator lengthwise along the optic axis, but also finding the proper angle of the detector surface with respect to the optic axis, both along the dispersion direction and perpendicular to this dispersion.

     

    For this purpose the detectors are mounted on a sturdy platform (one for each camera) supported along the optic axis at six points, with two sets of push-pull screws arranged in parallel to the dispersion direction, one set of three at each end of the platform. Three dial gauges allow recording the orientation of the detector platform to facilitate repeat set-ups on some future occasion. At first, the task of adjusting these six screws (together with the collimator position) may strike the observer as akin to the excercise of balancing a six-legged table. However, well organized, the task can be separated into a series of well-defined virtually independent operations which, done sequentially, can reduce an otherwise frustrating excercise from an entire afternoon to a straightforward procedure taking at most the better part of an hour.

     

    Even so, the observer should respect the precaution that focusing adjustments are never entirely repeatable: the platform's position permits slight displacements parallel to its surface so that the six screws never come to rest repeatably at exactly the same place, and a final "touch-up" focus series with the collimator is always required. Four grey hand-screws, two at each end of the detector platform, always leave some doubt about where the platform is fixed for any perticular setup. Changing a set-up during the night can involve a significant loss of observing time, which should be kept in mind.

     

    When a repeat setup is to be used, the daytime technician will have set the detector platform at the latest corresponding orientation noted in the logbook. In such cases it usually suffices to perform a focus sequence with the collimator alone. Collimator steps of 25 or 50 dial units suffice for a satisfactory focus sequence. At the 85mm camera a satisfactory collimator setting should lie between 250 and 550. At the 200mm camera the collimator should lie between 350 and 450. Outside these ranges, and especially when working at steeper grating angles, astigmatism arising from anamorphic magnification may require resetting the detector platforms. Then the full-blown adjustment procedure described below must be applied.

     

    The preparation for a complete setup involves the same procedure, whether a previous set-up, adopted from the log-book, is to be modified, or a new, untried, setup is to be attempted (such as when an untried CCD detector is to be installed, for example). This applies equally to setups involving cross dispersing "grisms". It is useful to think of the task as leveling a three-point kinematic support, with the other three screws destined merely to anchor a final adjustment at the end of the procedure. For definition, let us call the height of the platform along the optic axis the PISTON position, and the other two degrees of freedom as TILTs. The PISTON adjustment is read from the mean of the gold and the red dial gauges. TILT along dispersion is read from the difference between the black and the gold dial gauges, and TILT along the slit (or along the cross-dispersion) is read from the difference between the gold and the red gauges. A special long-handled hexagonal key is provided for doing the adjustments.

     

    Before starting, the four grey hand-screws should be backed off a half a turn. Since we have three gauges, it is easiest to select three screws for preliminary loosening, so that during the adjustment process (from start to final tie-down) there is one fixed point, and one screw, for each dial gauge. The final tie-down at the end of the entire adjustment excercise is described further on. For now, at the start, the two corner push screws at the red dial gauge should be loosened several turns. The red dial may move considerably while loosening. Its desired setting may be re-established once the platform is held by only its three fixed points. At the other end the corner push screw at the back of platform should be loosened several turns. At this stage the observer has full control over the platform: at the red gauge the center bronze pull screw controls the red gauge setting; at the black and gold gauges turning the bronze pull screw in the middle controls the gold dial, and turning the forward push screw controls the black dial. In practice the dial gauges cannot touch the platform at the same point where the screws act, so an adjustment of one screw may slightly affect the other two dials; but to a fairly good approximation thinking of the adjustments in these terms leads to a very rapidly converging iteration of adjustments. (Moving the red dial by, say, 50 dial units may move the dials at the other end by two or three units, so they may require a "touching up" if only the red position was to be changed.) PISTON is adjusted by moving the two bronze pull screws so the gold and red dials move by the same amount, followed by adjusting the corner push screw until the black dial tracks this same displacement (that is, the difference between black and gold is kept constant). TILT along dispersion is adjusted moving only the left corner push screw, and moves the black dial alone. TILT along the slit is adjusted by moving the two bronze pull screws in an opposite sense, so that the gold and red dial sum to the same amount, followed by adjusting the left corner push screw so that the black dial displacement tracks that of the gold dial (keeping the difference between them constant).

     

    At this point focus series for definitive adjustments may be begun. To a good approximation at the 85mm camera one piston unit corresponds to roughly 25 collimator units (and somewhat less when steep grating angles are involved). At the 200mm camera one piston unit corresponds to roughly five collimator units. If the better focus direction in a collimator focus series lies above 400, then to bring focus closer to the desired 400 (lower), at the 85mm camera adjustment of dial gauges should go to lower values, while at the 200mm camera the dial readings should go in an increasing sense (the dial gauges were mounted reading in opposite sense!).

     

    Since platform adjustments do not repeat with precision, all focus series should be done adjusting the collimator. Once the position or sense of best focus has been determined, a platform adjustment should be made, and a new collimator focus series repeated.

     

    Fastest convergence to a desired adjustment is obtained by doing in sequence: FIRST: getting the PISTON right. PISTON focus should be measured only at the center of the image; even when quite apparently focus quality at the edges of the chip may be miserable. Once the piston adjustment is acceptable, then SECOND: TILT along dispersion should be pursued. A TILT sequence is done not by stepping the collimator, but by stepping the difference between the black and gold dials by 10 dial units at the 200mm camera, or by 5 dial units at the 85mm camera. If the tilt is far out, with focus degrading rapidly along dispersion, larger tilt steps may be advisable. THIRD: a TILT sequence of focus along the slit may be done (no CCD is mounted in its dewar perpendicular to the optic axis, so it should not surprise that a difference bewteen red and gold dial readings gives superior performance). LAST: After all three adjustments have been completed, and especially when several TILT adjustments have been necessary (as in echellette mode), then a final collimator focus series will be advisable. However, this final collimator series should be delayed to after the final platform tie-down has been done.

     

    The final TIE-DOWN is to restore the platform to its definitive six point support. The definitive dial gauge readings should be noted in the log or some other record before starting. Then the tie-down is begun at the black and gold dial end. The loose corner push screw should be snugged up until a dial begins to move. As the corner push-screws take hold, the gold dial will begin to move. When this happens one may anticipate how much motion of the gold dial is likely, and compensate for this by turning the middle bronze screw in the opposite sense. A balance between the two corner push screws is selected by watching that the difference between the black and gold dial is preserved. Once the black and gold dial end is fixed, at the red dial the two corner push screws are snugged up. It is quite normal for the red dial to move outrageously during this process. A compensating "anticipatory" adjustment of the middle bronze pull screw should be attempted, iterating between the corner push screws and the middle pull screw toward a final satisfactory tie down. Once all six points are fast again, the four grey hand screws must not be forgotten. Failure to tie these down may produce unacceptable flexure when the telescope moves to different positions on the sky.

     

    Now a last collimator focus series should be done, and the definitive collimator focus setting noted in the log.

     

    5. FOCUSING THE TELESCOPE

     

    The guiding optics that view the slit mechanism rely on standard 35mm camera lenses, and their chromatic correction is generally not adequate. A star seen sharply on the slit may not be in proper focus and may indeed illuminate the collimator improperly, with consequent signal loss and spurious wavelength displacements. There are two methods for achieving proper focus of the telescope at the slit. One uses the slit as a knife-edge, with an eyepiece that may be pushed into the diverging beam immediately behind the slit allowing the performance of a "Focault" test. The alternate method relies on taking a series of exposures of the same star at several different positions along the slit. The methods produce equivalent results. The observer is cautioned: he should ascertain that a blocking filter is installed in the guider optics so that the same spectral region be imaged on the slit as being observed by the spectrometer. A BG39/2 suffices for wavelengths shortward of H-alpha, and an RG610/2 for spectra to the longward.

     

    5.1 The knife-edge test

     

    To prepare the spectrometer for a knife edge test a guiding paddle must be connected at the cassegrain position. The eyepiece at the spectrograph is pushed into the beam to a detent (and not all the way in). A switch on the auxiliary electronic box permits opening the shutter. If red blocking filters are used, these should be moved out of the way.

     

    Note that there is no occular in the eyepiece. The purpose is to see not the slit, but an image of the entrance pupil: the luminous annulus of the telescope primary with the shadow of the secondary. If a portion of the slit blocks the beam, that portion of the pupil will appear dark. Normally, while out of focus, only a narrow stripe of the pupil will appear illuminated. When the slit is oriented East-West, moving the telescope North-South with the guiding paddle moves the position of the luminous strip across the pupil. The strip is always blurry, and in motion, as atmospheric "seeing" deflects incoming light. As focus is approached, by moving the focus buttons (note that two buttons must be pushed simultaneously, one serving an "enabling" function), the stripe becomes wider, more blurry, and tends to move away from the pupil center. The observer must use the North-South guide motion to keep the strip centered in the pupil. When one is close to focus the entire pupil will appear illuminated. Since it is hard to get the eye close enough into the eyepiece to see the entire pupil at once, one must move one's head sideways to sample the entire pupil.

     

    When the entire pupil appears illuminated the final focus touch-up is achieved by using the North-South motion to test if the direction from which the light seems to enter the pupil (or leave it) is still noticable. Best focus is that position where the illumination appears uniformly "all over" the pupil simultaneously, without any preferred direction of Left-to-Right or Right-to-Left. When proper focus has been satisfactorily achieved, the Night Assistant must be asked to note the value by typing fset on the telescope console computer.

     

    On finishing, the observer is advised to remember: (1) to close the shutter switch; (2) to pull out the behind-the-slit eyepiece, and (3) to return the blocking filter to the proper position.

     

    5.2 The focus series

     

    If the Focault test repeats poorely, one may opt for a focus series. Before the start, one obtains the best possible focus by visual inspection of the star on the slit. The focus value is noted, and then focus is increased to thirty units higher than this value. The star is placed at one end of the slit, the plan being to sequentially place the star at six more positions along the slit, with a telescope focus sequentially lower in steps of ten units. A larger gap along the slit must be left at either the start or the end for orientation. Usually the entire operation can be executed in a multiple sequence of, say seven 10 second exposures. Set the LOOP COUNTER to 7<cr>; select seven positions along the slit, with a larger gap between the first two. The night assistant positions the star sequentially to each position, changing the telescope focus in a downward direction between each exposure. After each repositioning has been completed an exposure is made clicking left on the multiple button. When the loop is completed the remainder of the frame is read out in the normal manner.

    The image produced should show seven closely parallel spectra. A cut perpendicular to dispersion generally shows the weaker spectra also somewhat broader -- let us assume the chip is SITe2, which displays spectra along rows: The best assessment of focus quality is obtained using the implot package. The current version of IRAF cannot implot .fits files, so convert .fits ot .imh format by typing rfits ccd###.fits 1 ccd###.imh<cd>, where ### is the appropriate frame number, and implot ccd###<cr>. Set the averaging in implot to ten by typing in the graph display :a 10<cr>; then plot some columns at the middle of the chip, by typing in the graph :c 900<cr>. Center the seven profiles by enclosing them in an imaginary rectangle for which the lower left and uppar right corners are marked, positioning the cursor and striking the <e> key at the opposing corners. The FWHM of each profile is then measured marking a point at the base of a profile to the left, striking <p>, then moving to the right side of the profile base and striking <p> again, which produces a description of the profile giving its height, row position, and FWHM.
     

    The difficulty inherent to this method is that the slit may be hard to see with stars bright enough to expose adequately in ten seconds (between third and sixth magnitude), and in bad seeing it may readily be missed entirely. Some trial and error repeats may be required initially.

     

    6. FLAT FIELD PROCEDURES

     

    Flat field exposures must be taken when spectrophotometry is to be done, or when one wants to control fringing (Moirè patterns) at the red end of the spectrum. While the spectrometer contains an internal quartz halogen lamp, use of this lamp to control Moirè patterns is not advised. The geometry of the illumination from the spectrometer halogen lamp is sufficiently dissimilar from the telescope illumination to leave significant residual fringes. For the nost effective mitigation of Moirè fringing with the TEK1 chip longward of H-alpha the illumination from the dome lamps is recommended, even though exposure times will be substantially longer.

     

    7. PRECAUTIONS WORTH CONSIDERING

     

    Because of the non-negligible aberrations of the CANON camera lenses and their flimsy (commercial) construction, some "tricks" are frequently employed to mitigate the degradations of the instrument's performance. Several are described, as derived from experience with this instrument.

     

    Beam geometry

     

    Differences between the beams illuminating the optics from the comparison or calibration sources and from the telescope represent a source of error that appear in two forms. One is in the flat-field exposures described in section 6 above. The other is in the precision with which wavelengths may be measured. The problem is different in work with a long slit compared to short slit work. In short slit work (which represents by far the major use the instrument has seen) the instrument relies on a cardboard mask placed on the collimator to shape at least the outer envelope of the beam geometry of the comparison lamps to resemble that of the telescope. This permits obtaining radial velocity measurements to 5km/s with the 831/mm grating, or 3.5km/s with the 600/1.25u grating in second order red or third order blue. When slit lengths longer than two arc-minutes are required, the mask must be removed to minimize vignetting (this requires dismounting the collimator module; the cardboard mask should be saved, despite its homely aspect). Removing the mask will result in some degradation of the velocity precision.

     

    Mechanical flexure

     

    Mechanical flexure deforming the spectrograph with the changing orientation of the gravity vector displaces the image formed at the chip. With the slit in the East-West direction, in my experience, the effect is more marked in declination (since gravity then varies in the direction of dispersion) than it is in right ascension. If the observer wishes to obtain some idea how strong the effect is, this may be ascertained during an afternoon, taking a series of expoures with the slit East-West, and stepping the telescope at fifteen degree intervals in declination. At each position two exposures are made: one using the internal comparison lamp, and another with the He-Ne lamp on the end of the telescope illuminating the dome; the latter uses the geometry of starlight illumination. The difference between them is a measure of how tolerant to flexure the instrument would be if the comparison illumination were simultaneous with the object illumination. With the slit East-West the variation with time of the component of gravity parallel to dispersion (primarily North-South) is usually tiny, so the experiment described is a fair indication of flexure tolerance; in my experience the error is usually smaller than the precision of internal measurement errors. On the other hand, spectrophotometric exposures that keep the slit in a vertical plane compromise this immunity to flexure somewhat, and precision in radial velocity may then be less than desribed in section 6 above. Quite apparently observing programs seeking velocity precision should prefer velocity standards at declinations comparable to those of the program objects, and keep the slit East-West.

     

     

    APPENDIX A

     

    Rebooting ...

     

    The SPARC

     

    Solaris is the assumed operating system. Occasions arise when a computer must be rebooted. This process is closely but not entirely similar at the two telecopes. The procedure is described here, on the assumption that SPARC shows the login prompt :

     

    Login with: obs100<cr> or: obs40<cr> and follow this with the password. After a short while a number of windows will have awoken. Only the console and an xgterm window are essential. Others may be killed by clicking RIGHT on the upper border and in the menu that pops down clicking RIGHT on the <quit> button. (If the console window is the only window that appears after a minute's wait, put the cursor in the console window and type: xgterm &apm;<cr>, with the cursor in the background, click RIGHT to produce a menu, then drag right over the IRAF button to the xgterm button and let go. Follow this, typing in a window with: ximtool &<cr>, or, with the cursor on the background as before, click RIGHT and drag RIGHT over and beyond the IRAF button and in the second menu release on the ximtool button; an empty display window should appear after a short while. Type set stdimage=imt1024.

     

    Moving or resizing a window: If the windows that have been created are inconveniently located, they may be moved by clicking LEFT on the upper bar and dragging the window to a convenient location. If the size of a window is not what is wanted, place the cursor on a border or a conter until the arrow becomes a circle; then, holding the left mouse-button down, drag the circle in the direction in which you want the window size to change.

     

    Starting the aquisition window: Now put the cursor in the window and the directory into which the data are to be written. In this window type ccd<cr>. After a few seconds a dialog window will appear posing several questions: answer by selecting the telescope, the detector, the overscan desired, and the shutter (the spectrograph uses the Uniblitz). Some seconds later the control window shoud appear again.

     

    At the SWOPE telescope there is no guider, and the observer must do the acquisition and guiding by with the image of the slit shown by the monitor.

     

    The MODspec control PC CLONE

     

    The PC clone may occasionally hang. One gets out of the hung state, rebooting the PC: by holding down <Alt> and <ctrl>, then striking <del>. The reboot takes about a minute. The program controlling the comparison lamps and comparison mirror is started by striking 6<cr>, where the 6 refers to the GOMOD batch file. The program will ask if the collimator is attached and the collimator value. a y<cr> and <cr> alone as answer will now start the program. If some other message should appear, beginning with: "The cy233 is not attached ...", then assistance should be requested.

     

    On successful startup the control window shows a top line with the basic instructions, followed by six lines that control, in order: (1) the comparison lamps, (2) the comparison mirror, (3) the blocking filters, (4) the collimator, (5) the aperture wheel, and (6) a report of the shutter status. Only the first three of these are used. The highlighted symbols confirm direct readings of the instrument status. The software makes no assumption about the instrument state.

     

    The comparison lamps

     

    Comparison lamps are identified by a two letter code, followed by a one or a zero to turn ON or to turn OFF respectively. The available lamps are Hollow Cathode (Fe-Ar): hc1 or hc0; an argon lamp labeled deuterium: du1 or du0; a helium lamp: he1 or he0; a neon lamp: ne1 or ne0; and an incandescant lamp: in1 or in0.

     

    The comparison mirror

     

    The mirror is positioned into the beam with cm1<cr> and removed with cm0<cr>. The moving arm on which the mirror is installed also serves as an internal cover to protect the spectrograph innards from outside dust. So it is useful to leave the spectrograph with the cm1<cr> command at the end of the night. The mirror may also be moved manually, durning a small brass knob to the right of the slit controls. The position of the mirror is reported to the PC from four Hall effect sensors.

     

    The filter wheel

     

    Six filter positions are available. Normally the day crew will leave the filters properly identified. It may occur that a filter has been changed without placing the corresponding label in the software. If in doubt, ask a technician to confirm the filter placement. One of the six possible filter positions may be selected by typing fl#<cr>, where # is a digit between 1 and 6. Filter names shown on the screen may be changed from the setup window by striking the <F9> key and selecting filter setups.

     

    The aperture wheel

     

    When used without a slit, the mechanism positioning aperture masks may be operated giving an ap#<cr> command, where # is a digit between 1 and 4. The aperture wheel allows replacing the standard slit with photographically pre-fabricated aperture masks. The aperture masks are built on a 2.0 x 2.0 inch format, and carriers for these are found in the spectrograph filter box. Aperture masks have not been used since 1989.

     

    Help

     

    Striking the <F1gt; key will show a "help" screen identifying all the commands understood by the software. <esc> exits from the help screen.

     

     

     

Document Actions