Massachusetts Institute of Technology
> George R. Wallace, Jr. Astrophysical Observatory

Prototype Medium-Resolution Spectrograph ``Demon''
Information for Users


The following is intended to be a user reference primarily covering the WAO Prototype Medium-Resolution Spectrograph (``Demon''), suitable for a user who's already familiar with the basics of observing with a small telescope (astronomical coordinates, manually acquiring target fields, guiding, etc.). It intentionally does not contain extensive information particular to the WAO 16-in telescope on which it's presently used, so be sure to also review the 16-in User Information.


slit width: 120 microns
collimator: spherical mirror, focal length 250mm, f/10
gratings: (#1) reflection grating, 50 x 50 mm in size, 600 lines/mm. Suspected blazed for 750 nm in the first order. dispersion ?? nm/mm. Spectral resolution ??
(#2) reflection grating, 50 x 50 mm in size, 1200 lines/mm. Blaze unknown. dispersion ?? nm/mm. Spectral resolution ??
imaging elements: (#1) Nikon SLR lens, focal length 28mm, f/2.8
(#2) Nikon SLR lens, focal length 50mm, f/1.8
comparison spectra: argon
detector: SpectraSource Lynxx2000 camera using Texas Instruments TC255 CCD
useful image region 241 x 322 10-micron pixels
useful spectral range: (using grating #1) estimated 5000 A to 8300 A
guiding: manually by viewing slit mirror
acquisition eyepiece: 90%/10% beamsplitter to 25 mm focal length eyepiece and projected illuminated crosshair

This spectrograph seems to have been originally built sometime in 1983, and was modeled after that described in the article ``A Grating Spectrograph for a College Observatory'' by D. Schroeder almost a decade earlier (Sky & Telescope Vol. 47, No. 2, 1974 Feb, pp. 96-99). It's a simple grating-based instrument designed for use on small observatory telescopes and primarily intended for teaching purposes. Originally fitted with a 35mm SLR camera and film as the detector, it was designed to be relatively small, simple, inexpensive, and in principle easy to use. In the article Dr. Schroeder expresses that ``Taking such spectra on his own gives a student a greater appreciation of the problems faced by professional astronomers and of the results obtained by them.'' In the case of the original Wallace version instrument the problems almost always prevailed over the results, and the instrument earned the name ``Demon Spectrograph from Hell''.

Records indicate that at least two attempts to make the instrument more functional using a CCD detector, in 1994 with a LynxxMC camera and then again in 1996 with a STAR-1 camera, had not been entirely successful. During the summer 1998 ``exorcism of the Demon Spectrograph from Hell'' the CCD camera was changed to a Lynxx2000, a wider slit was swapped in, on-chip calibration spectra were added using fiber optics, and an illuminated crosshair was added to the slit-viewer eypiece.


It's a Good Idea to familiarize yourself with the location and operation of everything listed in this section with the lights on (or during daytime) before you begin your first observing session.

The detailed instructions for multi-step procedures appear in the next section.

[external views]
Figure 1: External views
[internal view]
Figure 2: Internal view
acquisition eyepiece
An acquisition eyepiece and accompanying ``homebrew'' illuminated crosshair assembly are inset into one side of the spectrograph's telescope coupler. 10% of the light from the telescope is diverted to the acquisition eypiece by an internal beam-splitting assembly. The position of the wire crosshairs can be adjusted in 2 directions using the small metal knobs on two sides of the crosshair assembly. The crosshairs are lit by an incandescent bulb, powered through a wire connected to the spectrograph's power compartment, where you'll also find the on/off toggle switch and a brightness control knob.


filter holder
A filter slide bar in the cylindrical telescope coupler, just ahead of the spectrograph housing, is held in place with a wing nut at its far end. This filter holder accepts a 1-inch diameter round filter.

slit-viewer eyepiece
Presently a 12.5mm eyepiece with an illuminated crosshair is used to view the slit. It's oriented so that one of the crosshairs will be at or near the slit and approximately parallel to it.

The crosshairs are illuminated faintly by a red LED, whose 1/8-in power plug needs to be connected to an appropriate battery box.

calibration lamp
A high-voltage gas discharge lamp is used to generate emission lines for wavelength calibration of spectra.
lamp type: argon, Oriel #6030
operating current: AC, 10 mA
dissipation: approx. 2.7 W
rated life: 500 hr
lamp housing: quartz
replacement cost: about $260
filter: glass UV safety filter
The lamp is normally used with its glass UV safety filter because even low-intensity UV can burn eyes and skin.

The lamp's cord and plug connects to a mating cord from the transformer power supply, which in turn has a regular plug for an AC outlet. The power supply is controlled by two switches on the end opposite the two cords: the ``off'' button is black and stays in after being pressed, while the ``on'' button is red and returns to its out (un-pressed) position when released. In other words:

Don't leave the calibration lamp on longer than necessary at any given time because it uses up the limited life of the lamp. It also gets very hot, which needs to be kept in mind should it be necessary to handle the lamp outside of its mount. (which hopefully shouldn't have to happen very often)

The calibration lamp mount is the black metal square ``post'' which pokes up through the main cover of the spectrograph. Two nylon set screws hold the lamp in its mount. The screws should be only as tight as needed to prevent the lamp from shifting about; too tight and they'll crush the lamp! Conversely, if the set screws are too lose, the lamp can fall out and then break when it hits the pier (which has already happened).

With the filter in place the orientation of the lamp in the mount matters because the filter window has to face the pair of fiber optics that bring the light to the slit. The calibration lines will be brightest when the window faces the fibers directly, pointing along the long direction of the spectrograph, in the direction away from the lens/grating compartment. It's possible to crudely dim the calibration lines by repositioning the lamp so that the window is angled slightly away from the fibers.

Possible complication: If the fibers come unattached internally and poke too far into the mount, it won't be possible to insert the lamp fully.

calibration spectrum position adjustment
The two long threaded rods that poke through the top and bottom of the spectrograph's main housing are screw adjustments that move the brackets that hold the slit-end of each of the two fiber optics along the slit.

The brackets share the same track and have a limited range of motion beyond which they mustn't be forced.

collimator focus micrometer
This micrometer is a super-fine-focus of the internal spectrograph optics, which moves the collimator mirror laterally on its mount. This'll be set by observatory staff as part of installing and aligning the spectrograph on the telescope, and so shouldn't be changed as part of routine observing by users.

The grating is accessible to the user once the grating/lens compartment cover is removed. In this spectrograph it's mounted ``camera-normal'', that is, facing the camera lens rather than the collimator mirror. The grating is held to its rotating platform with two Allen-head socket screws behind the grating.

We believe our 600 lines/mm grating is blazed for 7500 A, in which case the most useful range of this grating in combination with the Lynxx2000 CCD is probably about 5000 A to 8300 A.

grating angle micrometer
This micrometer adjustment is at the camera end of the spectrograph housing but near the base of the main box, and changes the center wavelength imaged by the camera by rotating the grating platform to change the angle of the grating. The micrometer reads in tenths of an inch (i.e. ``7.50'' is 0.75 in), and has a full range of adjustment if 1.0 inch.

Possible complication: If the 1.0-in range of adjustment of the micrometer doesn't reach far enough to image a wavelength or order you need for your observing program, you'll need to readjust the ``zero point'' of the range of motion by loosening the round clamp at the bottom of the shaft of the grating platform, rotating it to reset its position, and then re-clamping.

When configured for broadest wavelength coverage using the 600 lines/mm grating and the 28mm camera lens, at the first-order spectrum the CCD chip spans about 1670 A, and the micrometer changes the center wavelength about 1195 A per tenth-inch of adjustment.

camera lens (imaging element)
A Nikon bayonet-mount SLR camera lens which images the spectrum onto the CCD detector is mounted on the CCD housing. The lens aperture ring is toward the back near the CCD housing and should always be set to the largest aperture (lowest f-number, fastest setting) to minimize loss of light. The front end of the lens is inside the spectrograph housing, so focusing the lens involves removing the housing cover before being able to rotate the lens barrel to change the focus. The correct focus tends to be just a bit back from the ``infinity'' setting.

At present Demon can be configured with either of two available lenses:

  1. 28mm focal length, f/2.8, effective aperture 10mm
  2. 50mm focal length, f/1.8, effective aperture 28mm



Keep in mind at all times when using the spectrograph that it's an intrinsically delicate instrument; even slight internal misalignment or damage caused by any rough handling will cause you and all other spectrograph-users much observing time lost later!

[ 3.1. starting up | 3.2. configuring and focusing | 3.3. initial object acquisition | 3.4. guiding | 3.5. recording the calibration spectrum | 3.6. recording CCD calibration frames | 3.7. shutdown checklist ]

3.1. starting up

  1. Power on the control computer. Check the UT clock time on the computer and make sure it's within 1 second of UTC. Adjust it if necessary using Control Panels\Date and Time.

  2. Activate the Lynxx 2k 32-bit application on the desktop.

    This will, by default, put all your saved image files on the desktop, where you can see 'em and most easily keep track of what you've really saved. You'll also be able to explicitly specify a different location for your image files if you like.

  3. Once you've started the control program, choose Camera\Use Cooler and then allow at least 15 minutes of cool-down time for the camera to reach thermal equilibrium before taking any ``keeper'' images.

    The CCD cooler is designed to reduce the operating temperature to 50 deg C below the ambient air temperature. Each 10 deg C drop in operating temperature represents an increase in limiting magnitude of about 1 mag., so the colder the night the better the thermal noise characteristics of your images!

3.2. configuring and focusing

  1. Remove the lens/grating compartment cover, currently fastened with three knurled screws. Don't lose the screws!

  2. Inspect the camera lens and grating to determine whether they're the ones you've chosen for your observations.

    If either or both isn't/aren't you'll need to reconfigure the spectrograph. New users are expected to enlist an experienced user to join them at the telescope to help them learn how to do this procedure safely!

    1. Slew the telescope to an orientation so that the lens/grating compartment is below the rest of the main optics housing. This precaution removes the danger of parts or tools dropping down into the main optics of the spectrograph.

    2. Carefully remove the lens. The lens mount is a Nikon bayonet mount: to release, rotate lens counterclockwise about 1/6-turn and then pull it forward from the CCD.

    3. Carefully remove the grating (if you're changing it). This involves removing the two Allen-head socket screws behind the grating that hold it to its rotating platform. Don't lose the screws!

    4. Use the same two screws to carefully install the desired grating.

    5. Carefully install the desired lens, reversing the lens removal instructions above.

    6. Slew the telescope back to ``home'' position to proceed with focusing.

  3. Set the aperture ring on the lens to its lowest setting (for largest aperture).

  4. Set the lens focus to between 5 ft and 2 m on its distance scale, which will bring it near correct focus.

  5. Move the grating to its ``home'' position:

  6. Turn on the calibration lamp, and dim the room lights. (Also can drape black cloth over open compartment if stray light seems to be a problem)

  7. Start the CCD's focus mode (continuous readout and display) with an 0.1 s exposure time.

  8. You should be able to see either or both of the calibration lamp spectra, each of which will look like a row of short vertical bright lines.

    If you can't see the lines, and you've confirmed the calibration lamp is on and working, and that the lamp filter window is indeed facing the proper direction toward the fiber optics, then the calibration spectrum position adjustments may be holding the fiber optics too high or too low to appear on the chip (especially if the lens and/or grating configuration has just been changed).

    To ``home'' the calibration spectra when setting up the spectrograph, first move the upper spectrum as far up as it can go, then follow it with the lower spectrum as far up as it can go as well. This should guarantee that both spectra will be positioned vertically so they'll appear in the CCD field. (Note though that the home position is only for setting things up before observing because it leaves no space between the brackets for the actual object spectrum to shine through.)

  9. Use the image of the calibration lines to more carefully focus the camera lens until the lines are as sharp as you can get them (pick well-exposed but not overexposed lines to gauge the focus).

    Complication: Different wavelengths across the chip will have slightly different foci so you may not be able to achieve sharpest focus simultaneously for the entire width of the CCD.

  10. At this point you can adjust the vertical positions of the calibration spectra to move them out toward the top and bottom edges of the CCD field so that there's space between the brackets for light from the target object to actually get through.

  11. Now you're set to adjust the grating angle to place the desired center wavelength on the CCD.

    To figure out where the grating is currently set, stop the focus mode and use the accompanying plots (Figure 1) to identify precisely which calibration lines you've imaged in the CCD field. With the grating in the ``home'' position you should see lines from the bright 7000-8000A group.

    A particularly useful function for this is View\Horizontal Scan, which plots signal as a function of column number for a horizontal cut you define by clicking within the image.

    NOTE when comparing the screen display with the plots in Figure 1 that in the plots wavelength increases to the right, while the displayed CCD image is left-right reversed and wavelength increases to the left.

    [argon lines - 5000-6800 A]
    [argon lines - 6800-9800 A]
    Figure 1: Argon I lines, approximately as seen by Demon.
    Note that the lines in the second plot (around 7000-8000 A) are several orders of magnitude stronger than those in the first plot (around 6000 A).

  12. Once you've established where the grating is ``pointed'' reactivate focus mode and leave it running while you adjust the grating angle to the range of wavelengths needed for your observing program.

    Note that if you move shorter than 7000 A the calibration lines there are much fainter so you'll need to increase the greyscale contrast and/or switch to a longer exposure to be able to see them.

    Data Quality Hint: Be certain that you have calibration lines on your image that span whatever range of wavelengths you're interested in!

  13. Stop the focus mode, and turn off the calibration lamp.

  14. Replace the lens/grating cover.

3.3. initial object acquisition

  1. A high-power eyepiece with illuminated crosshairs is used as the slit viewer eyepiece so that the crosshairs provide an in-field reference to help in locating the slit in the field of view. The crosshairs of the eyepiece are illuminated by a red LED mounted in the side of the housing. The cord connected to the LED needs to be plugged into an appropriate battery box with a brightness control.

    To establish the orientation and location of the slit with respect to the crosshairs, point the telescope at a well-illuminated surface (such as the inside of the dome with the room lights on) and have a look through the slit viewer eyepiece. Note where the slit is with respect to the crosshairs, and make a sketch in your notebook for future reference. Consider that when you're observing a dark sky all you'll see is the illuminated crosshair and the target star, and even that only when the star's on the slit mirror.

  2. Next, acquire a bright star (nice high altitude, to use as a setup star) successively in the telescope's finder, the spectrograph's acquisition eyepiece, and finally on the slit mirror. Put the star near slit for now, but not on the slit just yet.

    Possible complications:

  3. Focus the image of the star in the slit viewer.

  4. Use the fine-motion controls of the telescope to place the setup star on the slit.

    The ``magic spot'' along the slit, (i.e. the location that actually images onto the CCD field of view) seems to be in the direction toward the CCD camera head, about halfway from the center to the edge of the slit viewer eyepiece field.

  5. Now that the telescope's focused you can go back and touch up the focus of the acquisition eyepiece: view through the acquisition eyepiece and adjust its focus (the eyepiece's, not the telescope's!) until the image is... uh... how about ``least blurry''?

    Also take this opportunity to adjust the position of the acquisition crosshairs to mark the position of the setup star. As long as you don't move these crosshairs you'll be able to use them to position future objects on the slit mirror more rapidly.

  6. Move the setup star to the magic spot on the slit and shoot a test exposure of its spectrum just to make sure everything's working properly.

3.4. guiding

Guiding corrections needed are determined manually by directly viewing the target object on the slit mirror through the slit-viewer eyepiece. Guide using the telescope's slowest-motion controls to keep the target on the slit continuously for the entire duration of the exposure.

The trackball for the control computer has an extra-long cable so that you can bring the trackball to the telescope, and start the exposure after you've already started guiding.

3.5. recording the calibration spectrum

Recording the calibration spectrum seems a deceptively simple task, assuming the vertical placement of the lines was set properly while configuring and focusing. Since the exposure needed for the bright lamp is usually shorter than that needed for the faint target object, it's just a matter of turning the calibration lamp on and then off again sometime during your image exposure, right? Basically yes, but in practice it's a little more tricky though: This also means there are two exposure times to record in your notebook for each image: the total CCD integration time that the shutter is open, and the amount of time the calibration lamp was turned on.

``Ok, but what if the exposure time for a bright object is shorter than the exposure time for faint calibration lines?'' Hopefully this situation won't actually happen, but if it does, at present the only means for blocking the light coming from the telescope while still getting the calibration lines is to put an opaque ``filter'' in the filter holder and use it as a manual shutter.

3.6. recording CCD calibration frames

bias frames
To set up a bias frame on the Lynxx 2000:
  1. Choose Camera\Expose
  2. Click on Bias Frame. The exposure time will change to N/A.
When you ``expose'' the chip will just be read out---the shutter won't open and there'll be no delay from integration time.

NOTE: Taking a bias frame seems to reset the exposure time for subsequent images back to the default of 0.1 second. We are annoyed.

dark frames
To set up a dark frame on the Lynxx 2000:
  1. Choose Camera\Expose
  2. Click on Dark Frame
  3. Set the desired exposure time
When you ``expose'' the timed integration begins but the shutter won't open.

flat frames
To shoot flat frames with the Lynxx 2000:
  1. Set up the flat field of choice (screen or twilight)
    (NOTE: This step is telescope-dependent, e.g. for the WAO 16-in see 16-in User Information - Dome Flats.)
  2. Treat just like an image frame, adjusting illumination and/or exposure time until you get average counts in the range 32000--52000 (i.e. between 50% and 80% of saturation).
  3. Be sure to go back and shoot dark frames for your flats!

You may have noticed we're specifically avoiding the Flat Frame option in the Expose Image window. It seems to do some kind of normalization to the image after it's read out, but (a) in absence of any clear documentation of what it actually does do to our images, (b) in absence of any actual need for the program to do anything weird to our raw flat frames, and (c) the fact that it opens up a mysterious Console window that doesn't have an off button, the party line on the Flat Frame option is: ``punt''.

3.7. shutdown checklist

  1. Deactivate the CCD control program by choosing File\Quit
  2. Have you copied your images from the local disk?
  3. Is the lens/grating compartment cover closed?
  4. Are unused camera lenses and gratings bagged and/or covered, and put away?
  5. If you had a filter in the filter holder, has it been removed?
  6. Is the slit-viewer illuminated reticle turned off? (so you don't exhaust the batteries)
  7. Is the calibration lamp off and unplugged?
  8. Is the power for the acquisition crosshair turned off?
  9. Power off the control computer by choosing Special\Shutdown


Idea is to print a hardcopy, draw in indicated field circle(s) with a compass, then photocopy onto a transparency.

Reproduction scales of sky atlases:

        Sky Atlas 2000 =  8.41'/mm = 505"/mm
      Uranometria 2000 =  3.23'/mm = 194"/mm
 Millennium Star Atlas =  1.67'/mm = 100"/mm
        Pickles charts = 0.285'/mm = 17.1"/mm
    (default settings)


6. Information for staff

The instructions written for ``Lining Up the Optics'' when the spectrograph was originally built, and which were subsequently included in hardcopy versions of ``The Wallace Manual'' as Appendix E, recommends aligning the internal optics indoors using a laser before mounting the instrument on the telescope. (See the Demon looseleaf binder)

Substantial experience with aligning and mounting Demon during 1998 summer demonstrated repeatedly that the approach originally documented doesn't presently work in practice. It's not clear whether we were simply unable to position the laser properly relative to the spectrograph, or whether the optics on the WAO 16-in telescope may be out of collimation, or both.

Instead I was able to improvise a modified procedure I used to successfully align the optics after the spectrograph had already been mounted on the WAO 16-in telescope. It's main disadvantage is it requires a mostly-illuminated Moon as light source - that is, a clear night with a gibbous or full Moon. So if someone un-aligns Demon during dark time everyone's going to lose the rest of that run!

Note especially that the most serious design flaws of this spectrograph have to do with the inability to robustly and precisely make these crucial adjustments. Substantial amounts of the instrument need to be taken apart for access, and repeated disassembly and reassembly takes its toll. Even when you get something in the right place, more often than not merely the act of tightening the very screws that are supposed to hold the component of interest in place, will ruin its positioning. Getting the alignment process right involves an exceptional amount of finesse along with a good deal of luck, and owing to the design there's very little practical leeway in not getting it right - even small misalignments cause tremendous losses of light in the final spectra.

In short, aligning the optics of Demon absolutely isn't a task to be taken on casually, or in a rush, or without the assistance of someone with substantial experience!

That said, even the experienced will want some instructions. Full instructions have not yet been written, but here's a ``cue-card'' sparse version I wrote for myself:

  1. On a suitable night, bring a small blank white card (business card, to find beam) and find the black cloth (to cover head and instrument with cover open)

  2. Set the RA tracking rate to the lunar rate and acquire the Moon with the telescope.

  3. Remove the fiber optics, baffle, slit assembly. Aim beam at collimator center by adjusting prism orientation

  4. Aim beam reflected from collimator at center of grating by adjusting collimator tilt screws (annoying, counterintuitive adjustments!)

  5. Adjust CCD position in bracket so that beam reflected from grating goes into lens centered and on-axis

  6. Reinstall slit assembly, position so that on beam from prism to collimator (center left-right) and angle to send reflected beam to flat mirror for slit-veiwer

  7. adjust slit-viewer mirror (an especially finnicky adjustment, involves lots of swearing)

  8. reinstall baffle and fiber optics

WAO home Last modified 1999 Feb 6 Valid HTML 3.2!