How Can We Help?
Echelle Spectrograph Users Manual
Bill Kunkel, Steve Shectman, Mark Phillips, Ian Thompson, Nidia Morrell
December 2001, June 2010, June 2017, March 2019
This document describes the operating procedures for the 100″ du Pont telescope Echelle spectrograph. With the principal optical components held by a short optical bench attached at right angles to the telescope optical axis mounted directly on the lower face of the IMB for freedom from flexure, the instrument was originally designed for use with a two-dimensional “2D-Frutti” photon-counting detector. In early 2000 this was replaced by the TEK#5 CCD camera with a 2k by 2k format and 24 micron pixels, providing an order separation of not less than 11 pixels in the red, and more in the blue. In December 2009, the TEK#5 CCD was replaced by a SITe2K CCD, having the same format and pixel size. Simultaneous wavelength coverage extends from ~3700 to 7000 Å at a fwhm resolution of about 45,000 or 7 km/s with a one arc-second slit. The longest non-overlapping slit projects to six arc-seconds on the sky.
The optics employ a cemented triplet lens of 500 mm focal length operating as both collimator and camera in a Littrow arrangement. Immediately behind the movable entrance aperture plate there are two filter slot positions, one a wheel, now fixed open, the second permitting the insertion of an external filter, normally a diffuser to permit obtaining “milky flat” frames from daytime sky exposures. Behind the shutter blade the beam is deflected 90 degrees by a small right-angle prism. A negative lens on the exit face of the prism converts the telescope’s f/7 beam to f/5. Light then travels to the objective lens, through the cross disperser to the echelle grating, returning through the cross disperser and the lens a second time, passing immediately below the prism to a plano-concave field-flattener fixed to the exit surface of the echelle case body where the CCD camera is mounted. Comparison light from a Th-Ar hollow cathode lamp (mounted on the back face of the spectrometer body) is introduced via a movable prism above the aperture plate. Visual inspection of the dewar on the instrument shows a variation of spacing between the dewar face and the spectrometer body. This tilt in the dewar compensates for a chromatic variation of the focus with the orders, and is the normal appearance.
The CCD data acquisition system is started by clicking on the ‘Echelle’ icon on the bottom tool bar of Clarity or Caballo.
‘Echelle’ first launches the CCD configuation windows (startup and hosts), which are similar to that of other du Pont instruments.
The CCD startup windows
In the startup window: update the observer’s name and under “CCD” select SITe2K-1.
The number of overscan lines and columns can be modified, although the default 128 is the recommended value.
Clicking ‘Startup’ will start the CCD camera gui:
The CCD Camera GUI
From the “Camera” pull-down menu, go to Data_Path and write the name of the directory where the data will be stored (usually, although not necessarily, a subdirecotry of DATA having the observing date as its name). If this directory doesn’t exist, it will be created. You can just click on ‘DEFAULT’ and a the standard directory name will be offered.
“ExpTime” is the exposure time in seconds for the next observation. (For this parameter the entry window will be highlighted green after the number is typed. The value is not recorded by the program until the user hits Enter). The exposure time can be changed any time during the exposure by entering a new number and hitting Enter.
Loops: Setting up a loop (= number of
integrations wanted) can be a convenient way to take flat field series
or make synoptic observations.
ExpType: The exposure type (saved in the fits header) can be set to “Object”, “Bias”, “Dark”, or “Flat”. Selecting “Bias” will automaticaly set the exposure time to 0 seconds. Selecting “Dark” will disable the shutter for the exposure.
Binning: The CCD can be binned by factors of 1, 2, 3, or 4 in rows and columns.
Subraster: An echelle image does not fill in the full detector area, thus some readout time can be saved by setting a subraster; instead of using the “Full” frame readout which is the default (as indicated by the “Full” label in the right side of the CCD camera GUI). Clicking on the button next to the “Full” label will start the subraster window.
The subraster section can be entered as XcYcWH (X_center, Y_center, Width and Height) or X1X2Y1Y2 (starting and ending column and row numbers). These coordinates are in binned pixels if the CCD is read in a binned mode. The full requested number of overscan columns will be added. After typing the subraster values, click the “APPLY” button, and then “DONE” to activate the subraster mode. It is also possible to use the option “Save” to save the selected subraster to a file.
A useful subraster for the echelle is as follows (the following lines will be saved to the selected file name):
1 1025 751 2048 1500
This will save columns 1 to 2048 and rows 1 to 1500. Even if you used the X1X2Y1Y2 format to enter the subraster coordinates, the file saved by the system will be in XcYcWH format.
The subraster region will be marked by a yellow rectangle on the Quick Look tool window.
You can also chose between saving in ‘full’ or ‘minimal’ mode, clicking in the SaveMode button. Saving in full mode, will fill in with zeroes the pixels outside the subraster region, leaving the original file size.
“Start” begins a science exposure. The
exposure can be paused (e.g., for clouds) and resumed by toggling the
“Pause” button, but observers should remember that cosmic rays are
accumulating while the exposure is suspended. The progress of the
exposure can be seen in the bar-graph immediately to the right of the
ExpTime entry. This graph fills to the right during the exposure and as
the CCD reads out the bar empties to the right.
The “Snap” (for “snapshot”) button automatically sets the 4 x 4 mode, takes an exposure, and — unlike the “start” command — does not update the frame number at the end of the exposure. It is meant for a quick lookout before taking an exposure, or for cleaning the chip after an ‘Abort’, but not for science.
For a “File” number of nnn the frame will be saved as file ccd0nnn.fits (e.g. ccd0012.fits). The file number will be updated by one for all exposures except a “Snap”. Note that if a frame with the same number as the current exposure already exists on the disk, it will be overwritten. The file number can be edited by clicking on the ‘Camera’ menu on the top tool bar on Clarity and Caballo and then selecting ‘File Number’. The Camera menu also allows to edit the data path for the directory were data will be stored:
Stop/Abort: Exposures can be stopped or aborted. If the “Stop” button is clicked during an exposure the shutter will close and the frame will be read out. If the “Abort” button is clicked during read out, the data will be dumped and the file number not updated. A wipe of the chips follows, but since the user will not generally know when it is finished, it is recommended that a short “Snap” exposure be taken to insure a clean chip before the next science frame is taken.
The readout speed can be set to “Slow”, “Fast” or “Turbo”, having the following characteristics (for 1 x1 binning):
Speed Gain Noise (e-) Readout time (s) Saturation
Slow 1.05 6.0 79 > 64,000 DN
Fast 1.55 6.3 67 >64,000 DN
Turbo 2.75 7.9 54 >55,000 DN
Linearity tests have yet to be carried out for the SITe2K CCD.
From the “Camera” pull-down menu you can start the Quick Look tool where the exposures will be automatically displayed:
The Quick Look Tool windows
From “Camera” you can also start the Dewar_Status window which is useful to monitor the detector temperature.
For details about the Quick Look tool refer to the IMACS users manual:
IRAF is used to display the data and to make quick extractions of the echelle orders.
To start IRAF and DS9, click on the IRAF and Sunburst buttons on the Clarity tools panel on bottom of the screen. Then, in the IRAF window, ‘cd’ to the directory where the data will be stored (as defined in Data_Path) and resize the DS9 window to the current image size by typing:
Control of the instrument has evolved with the years. Spectrometer focus remains an entirely manual adjustment. The projection optics for the Th Ar comparison light remain as in the original design and positioning of the aperture plate is controlled with a bidirectional contact switch without encoder readout. These latter two functions are activated by the left two of the three toggle switches on the Echelle Control Box. The filter wheel multi-position switch is currently disabled, and the voltmeter switch serves only for engineering diagnostic purposes.
The operations are described below in the order in which they would normally be performed during an instrument setup operation. This will normally be performed by the day technical crew. Their help should be also requested if any further modifications to the original setup are needed during an observing run.
There is no encoder reading the aperture plate position and so the
positioning has to be done by eye. To begin, you need to illuminate the
aperture plate so that it is clearly visible on the TV screen. The dome
lights are too bright and will produce saturation. Using a high
integration value is not recommended because the aperture has to be
moved during the process, and having the integration set to more than 1
sec will cause the screen to take a long time to refresh. If the light
leaking into the dome is not enough, the flat lamps can be used with a
voltage setting near 350.
Once the aperture is properly illuminated, you can use the aperture switch on the control box to move the aperture, always watching the changes on the TV screen. Pushing the switch up will cause the plate move to the right on the TV screen. Move it this way until the large number 1 aperture (0.06 x 0.2 in) is seen near the screen center. Keep moving the aperture in the same direction and the left edge of the isolating window which lies below the slit plate will become evident (this may need a new adjustment of the illumination level).
Mark the middle point of that window edge with one of the moving cursors (e.g. xy3). The cursor will be kept in that place during all the spectrograph operation.
Now gently move the aperture plate further in the same direction until you clearly see the right edge of the isolating window, and place another cursor (for example, xy4) at its middle position.
Both edges of the window below the aperture plate are now marked. Although those will change slightly between setups, they should be not far from xy3: 571, 587; and xy4: 578, 535 (values obtained on February 9, 2010).
Next, you should place another cursor (e.g. xy1) at the middle point in between xy3 and xy4, thus marking the center of the isolating window. This is where the slit to be used will be placed.
Whith aperture 1 perfectly centered, both edges of the window are barely seen behind, but we will give you another hint to make sure that the centering is right.
Move left the aperture plate (by pressing the switch down) so one of the small 0.75″ x 0.75″ holes (number 4 or 5) reaches the central popsition. Be aware that none of those will fall exactly on the xy1 cursor, because those apertures are below (ap. 4) and above (ap. 5) the midline, respectively.
This small aperture will be used in order to focus the spectrograph. But before starting the focus sequence, take a 90-100s exposure of the ThAr lamp (move the comparison mirror to the ‘IN’ position; turn on the ThAr lamp, set the current to 15 ma, and press ‘START’ on the CCD gui). To keep the ThAr lamp alive as long as possible, turn it off as soon as the exposure is over and the readout starts.
Now use IRAF to display the lamp image you just obtained and check that it is similar to this one:
If you get something like this instead:
it means that the window center in not properly marked, and you are
getting on your image light that comes through an adjacent aperture
(either left or right) as well as the desired one.
Note that the bright saturated lines in the upper part of the image are duplicated (all lines are duplicated indeed, but the effect is easier to see in the brighter ones). If this happens, go back to the big number 1 aperture and repeat the procedure to mark the edges of the isolating window and its central position.
Take a new test ThAr exposure, and if it looks good, you can start focusing the spectrograph.
The comparison mirror is controlled with the leftmost toggle switch on the Control Box. The motor takes about two seconds to move the comparison mirror into place, and a red LED comes on. The comparison mirror sometimes sticks, and when that happens the LED will come on for only a short time — about half a second. The recovery is to try again. The shutter is controlled by the CCD acquisition program, and the shutter control toggle switch on the Echelle Control Box is disabled.
Shutter timing tests have not yet been performed.
The thirty apertures machined into the plate are listed below in order
from left to right on the slit viewing TV. Note that the tiny holes
(numbers 2 through 5) are hard to locate by eye on the TV monitor
because of their size. Their spacing along a straight line connecting
all aperture centers is uniform however, and positioning them with the
help of a third cursor xy3 on the TV guider may prove useful. When
slits longer than four arc-second are used the closest orders toward the
red end of the spectrum will begin to overlap.
|0.06 x 0.2 in
|0.75″ x 0.75″ (arc-sec)
|two 0.75″ x 0.75″
|spaced 4.0″ apart
|0.75″ x 0.75″
|hole, 5″ below midline
|0.75″ x 0.75″
|hole, 5″ above midline
|0.5″, 0.75″, 1″, 1.5″, 2″, 4″
|0.5″, 0.75″, 1″, 1.5″, 2″
|0.5″, 0.75″, 1″, 1.5″, 2″, 3″, 4″, 8″
|0.5″, 0.75″, 1″, 1.5″, 2″, 4″
A Th-Ar hollow-cathode lamp serves as the comparison source for the
Echelle spectrograph. The power supply should be turned on only when
the lamp is being used. The current should be set between 10 and 15
ma. PLEASE DO NOT EXCEED THE 15 ma LIMIT. We may not have a spare
Th-Ar lamp. Typical exposure times run from 20 to 90 seconds, depending
on slit width.
Initial focusing of the spectrograph is normally done by LCO staff. Nevertheless, changes of temperature during an observing run of greater than 3 degrees Celsius are cause for redetermining the focus setting, and so the observer should become familiar with the procedure (but ask the day technician to show you how to adjust the focus before you try). The basic steps of this procedure are covered in this section.
Focussing consists of taking a series of ThAr exposures, moving the lens for each exposure, and analysing the frames with the IRAF script fechelle. The lens position is controlled with a micrometer screw, and motion steps of 0.20-mm or 0.25-mm give the best results. Micrometer adjustment should always be from larger to smaller values to ensure smooth motion of the lens.
Focus observations should be made with one of the 0.75″ x 0.75″ holes (#3 or #4 from the apertures list), this images onto the chip as a point source. A satisfactory criterion for focus quality is a fwhm of about 2.0 – 2.2 pixels. The sharpest focus in the X-direction differs from the Y-direction by about 0.3mm in the micrometer setting. An “optimal” best setting half-way in between can be achieved with a degradation in fwhm in both axes no larger than 0.05 pixels. The observer has to decide on the compromise that will meet the desired observing goals. Caution must be exercised in deciding on a compromise with focus along dispersion (the X-axis) and along the slit (the Y-axis): the two differ just enough to require some thought.
The following plot showing the best empirically determined focus settings for different Cell temperatures can be used as a starting reference:
The lens is held in dove-tailed ways clamped via two dark-anodized hand
wheels some two inches in diameter. These wheels should be turned half a
turn counter-clockwise, and then be tapped sharply with the heel of the
palm to snap them free of “stickage.” Once freed, a clockwise motion
of the micrometer screw to lower values may be applied. If the value to
be set is higher (this will be the case if the temperature increased),
always back off the micrometer by a full turn (0.5mm per turn) beyond
the destination setting, and approach the final position from greater
values. When the desired setting is attained the hand wheels must always
be clamped. Leaving the clamps loose leaves the lens in an
To run the echelle focus script “fechelle” you must first load the “c100echelle” package by typing c100echelle at the IRAF prompt. Now type fechelle. The script will ask for a frame name. Since the comparison lines will retain their relative positions on each focus frame, the table of line positions needs to be determined only for the first frame. The script asks for the line list name. Enter a new value or accept the default. The script will then ask whether to use the existing table, or measure new values, and offers a default (the last table used). If a new table is to be generated, the observer positions the cursor at each feature, and adds the line to the table by hitting the space-bar. As many as a hundred features may be selected. The selected lines should be as widely distributed across the frame as practical, since the true focal surface does not coincide completely with the chip surface. When the last feature has been added, the loop may be exited by striking the “q” key. If during the loop one wants to exit because a mistake was made, the control-C sequence WILL HANG IRAF, requiring IRAF and the xgterm window to be reloaded! Instead, use the “I” key to produce a controlled abort. Alternatively, one can restart the task of marking the features by using the “r” key. On proper exit of the loop, the script will measure the profile of each line, and then plot a map of the measured fwhm values, with the symbol size proportional to the fwhm in each axis for each measured line, as well as the mean fwhm values in the x and y directions. A paper copy of the map may be obtained by typing “=gcur” at the IRAF prompt, which puts a cursor in the plot. With the cursor in the plot strike the “=” key, and when the print – done message appears, strike the “q” key to exit. The plot will print to the Laser-Jet printer in the observing room.
Fringing in the CCD is just visible at the longer wavelengths (longward of H-alpha), so that any flat field illumination must reflect the wavelength illuminating the local neighborhood of each pixel. A sufficiently bright and smooth illumination is obtained from the day-time sky, which is very much “hotter” than any lamp (equivalent to about 10,000K), and so produces useful flat field illumination even near the H and K lines of CaII. The solar absorption spectrum may then be smoothed over optically by placing a diffusing opal glass (called ‘diffuser’) in a filter slot on the side of the spectrograph. The opal glass is located far enough behind the entrance aperture to spread each wavelength over a bell-shaped distribution some fifty pixels in diameter. Since the order separation between adjacent orders is never less than 11 pixels, each pixel will be illuminated by light not more than 400A different from its working wavelength. The opal spreading is sufficiently smooth to remove all traces of solar daylight absorption features. The opal glass diffuser is mounted in a bracket that slides into a slot in the side of the spectrograph located to the to the far left of the focus hand-wheels on the spectrograph body, just above the face mounting the CCD dewar. The slide is stored in the filter loading room off of the dome floor. The slot must be kept covered with black tape when the slide is not in place.
A member of the day crew should be asked to open the du Pont dome about 3 hours before sunset when the sun is far enough from the meridian to permit positioning the dome so that no direct sunlight reaches the interior of the dome or falls on the upper end of the telescope. Doing this much later puts telluric molecular absorption bands across the spectrum, and flats should be done before the last hour of sunlight. Good exposures are obtained, depending on the sky, the season of the year, and cloud cover (yes, this observing can be done in overcast conditions!) in times between 180 seconds or longer. A milky flat should look similar to the following picture:
To process the “Milky Flat” make a median-averaged copy of the bias-subtracted and bad-column-corrected frame, running a “box-car” (available in IRAF images.imfilter package) some fifteen pixels on a side on the copy, keeping the frame dimension the same. The box-car may perhaps be made shorter in the y-direction, and longer in the x-direction. The “Milky Flat” is then divided by the box-car smoothed frame, yielding a new “flat” whose value is nearly unity over the useable portion of the frame. The edges of the final normalized flat are noisy, but, of course, these parts of the frame are not used anyway.