Skip to content. | Skip to navigation

Personal tools
You are here: Home Users Konstantina Boutsia LDSS-3 long-slit spectroscopy cookbook

LDSS-3 long-slit spectroscopy cookbook

Daytime Preparation

Confirm with the Instrument Specialist that LDSS-3C has been setup with masks, filters and any other "Special Requirements" noted in your Instrument Setup Request Form, submitted three weeks before your observing run.

The GUI should already be running on the observer workstation but if it is not (or if it stops at any point) simply open an xterm, type “ldss3" and follow the instructions in the "LDSS3 Control Software" section of the LDSS3 Manual. Take time to familiarize yourself with the GUI controls and exercise the components (filter changes, aperture mask changes, grism changes). You should confirm that the filters you need are installed and the instrument is properly focused. While the Instrument Specialist should have verified the focus as part of the daily setup, it is always a good idea to at least verify that the focus values are near the nominal value (standard filters have a focus range of ~900-1000 at about 10C) and that as you change the filters the focus values are properly updated. Note that focus changes tend to be executed more reliably when you INCREASE focus numbers than when you decrease them. The focus values vary as a function of temperature. When the option “Auto” is selected on the “Focus” section of the GUI, focus values will be automatically updated if the temperature changes by more than 4 degrees, following an empirical linear fit calculated by the Instrument Specialists (for more details see v2.4 (2017-04-11) of the GUI) .

In the afternoon, before starting daytime calibrations, make sure that the Instrument Specialist has prepared the telescope. This includes opening the mirror covers and having the tertiary mirror facing the West Nasmyth instrument port. For long-slit observations, the calibrations to be obtained in the afternoon include biases and images of all slit masks, to be used during the night for calculating slit positions.

Turn Vsub-on, set on the GUI the combination of binning/gain/speed necessary for your science observations and start the calibration sequence (for gain we recommend low and for speed we recommend fast in all cases)

Bias: To take a sequence of bias images, in the LDSS3 GUI, set Loops to the number of exposures you wish to take, ExpType to "Bias", and the Binning X and Binning Y values to whichever values you plan to use at night, and then click on the "Start" button to start the loop sequence.

Slit mask images:  Deploy the Flat Field Screen (FF-Screen) using the DCU/Clay GUI (see manual). During daytime, the ambient light should be sufficient to obtain usable slit mask images with a 10sec exposure time. If you have to do this task during night, 3sec with the Ql lamp should be enough. If you use subraster mode, make sure that save mode is set to FULL, as this is required by the IRAF field acquisition software (Note: as of May 2017 the subraster mode presents read-out problems and should not be used). Take images for each slit mask you will use during the night/run. When you are finished, turn off any lamps you may have used, take the FF-Screen out of the telescope beam (Retract) and ask the Instrument Specialist to close the mirror covers. Turn Vsub-off.

The afternoon is also a good time to start IRAF on the observer workstation and load the LDSS3 packages by typing "ldss3" at the 'ecl>' prompt. Take some time to familiarize yourself with the LDSS3 packages (i.e. lstitch, ltoslit, lcobject, lmvslit).

“lstitch” is useful for all observing modes as a means to stitch together the two FITS files that result from a standard readout of the dual-amp CCD. The resultant stitched image will display the full CCD in the orientation matching that of the LDSS3 quick look tool.  


Fig.1: epar of the IRAF task lstitch 


Run "lstitch" on each of your slit mask images, display the resultant image and measure the pixel coordinate of the slit position that you want to use for your target(s) during the night. Record these values as they are required input for the 'ltoslit' routine, used for initial target acquisition. The slit pixel coordinates MUST be measured on the stitched image by displaying it in e.g. ds9. Measuring the slit coordinates in the LDSS-3 Quick Look window will produce the wrong results (Note: If you experience a large difference between the two amplifier when displaying the stitched image, add the bias section [1025:1152,1:4096])

If you have not already done so in preparation for your observing run, you should be sure to prepare your Observing Catalog File


Night-time Operations


Have the TO slew to the object that you wish to observe from your Catalog File. As soon as the telescope has slewed to the position of the object and the rotator has reached the desired rotator angle, let the TO know that you will need to run “ltoslit” since they need to turn off guiding and the Shack-Hartman (SH). Run "ltoslit" with the X,Y position of the slit recorded in the afternoon. Binning should be set to the values used to take the stitched image on which the slit centre coordinates have been measured.          

The script will ask you:  “Send RA/DEC offset to TCS?”, if the offsets seem reasonable type “yes” (default), if you do not want to send the offsets, type “no”. The offsets vary depending the used slit and can be large. For example for entering a source in the central slit usually a ~30" offset is needed in RA/Dec. For the red and blue slits the offsets are of the order of 100". 

Once at the offset position, the TO can start guiding and SH by setting up a guide star and an off-axis SH star.


Fig.2: epar of the IRAF task ltoslit  


The “ltoslit” script is meant just to bring your target close to the slit position. It is not meant for accurate positioning at the pixel level, because the distance is measured based on the centre of rotation of the detector. Since the pointing error of the telescope is of the order of 5” - you should expect your target to appear in a radius of <5” from your requested value. For accurate positioning of the source on a pixel level, the “lcobject” script should be used.

You should now be ready to take an image of the field and align your first target. In the LDSS3 GUI, set the Filter to the desired position, make sure the aperture and grism positions are "OPEN", set ExpTime to 10-30 seconds depending on target magnitude and Loops to 1, and then click on the "Start" button to begin the exposure. Next take an image of the field with long-slit installed (this is required to allow a precise measurement of the slit position when installed at the rotation angle of your observations - this way you may compensate for flexures of the instrument because of the rotator position on daytime and nighttime images). When the exposure is finished you will need to run the 'lcobject' routine to center the target on your desired slit position. Type "epar lcobject" in IRAF.  


 Fig.3: epar of the IRAF task lcobject  


Set the "image" parameter to the name of the image of the field and the slit image parameter to your slit image. In the "xslit" and "yslit" parameters enter the CCD coordinates of the slit (the ones you recorded in the afternoon). Enter an appropriate box size in the 'length' parameter - this is the size, in pixels, of the extraction area around the X,Y slit position that should be close to the centre of the box. A value between 500 and 800 should be sufficient. This means that when the images are displayed in the next step your target should appear close to the centre of the extraction box.

When you run "lcobject", you will be queried if you want to skip measuring the slit position on the slit image - accept the default "no" and proceed to mark the slit position in the extracted and displayed image.

Next you will be asked if you have just acquired the target. If you have run “ltoslit” before “lcobject” then the answer should be “no” (set as default answer) so that the script will extract the same area of the field image as the area extracted for the slit image. If you answer “yes”, then the script will extract the subsection around the centre of the field where the telescope should be pointing after the initial slew. We recommend that you always use “ltoslit” and you take the default answer “no” to this question. At this point, the image section of the field will be displayed and you will be asked to mark the object which you wish to center in the slit. If you hit the "space bar", the coordinates of the object center will be calculated automatically by the code. This is appropriate for stars or galaxies with compact cores. For diffuse objects, the "m" key can be used to manually mark the position you would like to have your slit centred.

After this step the program will output the required offsets to put the object on the desired slit position. 

You will now be asked if the calculated offsets should be sent to the TCS. Answer "yes" and you will be asked if you want to do a "coordinated offset". Let the night assistant know you will do a coordinated offset, and then proceed to answer YES to this question.

Our experience is that for offsets of a few arc seconds, the object will be at the appropriate place without further iterations. If the offsets are big (>5"), then you should obtain a new image and perform another "lcobject" command (and you should also check the values that you input to the previous 'ltoslit' command), but this time you don't need to measure the center of the slit. The values obtained during the first “lcobject” iteration, where the slit centre was manually selected, are conserved and displayed in the “xslit”, “yslit" parameter lines.



Fig.4: screenshot of the lcobject task output, after centering the target. Here measurement of the slit centre has been skipped but in principle it should be done when changing targets.  

Once the target has been centered, if it is bright enough a through slit image could be taken. To do this, in addition to the current setup, select the long-slit aperture (while keeping the grism position “OPEN”) and start integrating. Then define the appropriate scale cuts in your image to see if your target is visible through the slit. When you are happy with the position of your target in the slit, in addition to the current setup you can select the appropriate grism - make sure you have the correct speed and gain setting (for speed we recommend fast in all cases) and start your science integration.

For calibrations after you science target, you need to notify the TO so that guiding and SH corrections are stopped. You can then move the screen in in the DCU/Clay GUI (Deploy), turn on the He-Ne-Ar lamps and take the arc exposure. You may also take flat exposures using the Ql, Qh quartz lamps (at least 3 are recommended).



For offsetting your target in the slit after using "lcobject" to centre it, there is the iraf task “lmvslit” .   

Screen Shot 2017-07-08 at 8.16.37 PM.png

Fig.5: epar of the IRAF task lmvslit  

As “xcoord” you should put the desired move in the X axis in pixels and in “ycoord” how much the source should move in the Y axis in pixels. For example if you need to move your source 10” along the slit to the right you should put xcoord=53 and ycoord=0.

It has been verified that lmvslit is very accurate for offsets up to several arcmins. Offsets of as large as 5 arcmins have been tested and found to be working, moving in both axis. For example:

From position A: x=658,y=2256 to position B:x=1751,y=2871

lmvslit : xcoord=1093,ycoord=615

And from position B:x=1751, y=2871 to position C: x=863,y=1033

lmvslit: xcoord= - 915, ycoord= - 1838   

All coordinates have been measured in the stitched image.

It happens that during such large offsets the guide star is lost and the TO should find a new guide star appropriate for the new position. In such cases we recommend to verify the position of your source in the slit taking a through slit image before starting the science acquisition.

Cheat Sheet

Once you have familiarised yourself with the process you may follow the cheat sheet below:

  • Slew to field

  • Run "ltoslit"

  • Take direct image of the field

  • Insert slit and take direct image of the slit

  • Run "lcobject"

  • Take direct image through the slit and check if the object is seen

  • Insert grism and take spectra

  • Before changing target make sure to take arcs (always) and flats if necessary

Document Actions