MM-MOS at the TNG


Mask preparation using the Interactive Mask Design Interface
NEW !  IMDI Version 2.0 released. The new release of IMDI includes support for the new E2V CCD at LRS/Dolores and the new set of grisms.  In addition, some known issues in previous versions have been fixed.

Getting the IMDI software:

1) Request the most recent version of the IMDI package to the TNG resident astronomer; be sure that the IDL Astronomy User's Library is installed in your system and included in the IDL_PATH.
2) Unpack the IMDI tarball in your working directory.
3) Create a file "mos.dat" appropriate for your version of IDL. For example,
cp mos_60.dat  mos.dat  
to run IMDI under IDL versions 6.x. Use mos_54.dat for IDL versions 5.x.
4) Run the setup file:   csh chooseCCD  E/L
where E is the flag to be used if the preimaging was taken with the new E2V CCD, and L is used for images taken with the old LORAL CCD.
5) To run IMDI, start IDL and type the following commands (the first one is needed to correctly display the image on a 24-bit screen):
device, decomposed=0
@ mos

Note for Linux Users. There are two known issues when using IMDI with Linux X-window systems. The first is that the IMDI display is not restored (i.e. it remains black) when windows overlap. Second problem, changes in the IDL color table are not supported with the 16/24-bit graphics that is the default in Linux. As the results, the color controls enabled by the COLOR button do not take immediate effect on the displayed image. Both problems will be fixed in the next version IMDI V.2.1. In the meantime, we suggest the following workarounds. 1) Edit the X-windows parameter file (e.g., /etc/X11/xorg.conf) by adding the Option "BackingStore" (and the "Virtual" option if  you run IMDI on a small-screen laptop) as in this example.
2) Load your image, experiment a bit with [COLOR] to find the best table (although the default MOS table will do in most cases), then  re-[LOAD]  the  image. The new color table will remain in effect until you quit IDL. This feature works on Linux only with IDL 6.x. It is convenient to do that before starting any mask design.

Using IMDI.

Using the IMDI is intended to be self-explaining. Run it by typing "@ mos" from IDL (Vers. 5.4 or later). Images taken with the DOLoRes are displayed on a   3-window screen,  from which targets are interactively picked up and centered. Slits are adjusted under control of a real-time anti-overlap algorithm. 
Opening a pre-imaging FITS file using menu_bulo.gif (1826 bytes)   is the first  step.  The program can read both ".fits" and ".fts" image files. The upper window (``main window'') of the Interface is used for image selection, while the lower rigth window (``zoom window'') is used for fine centering of the targets. The lower-left section gives an overview of the field (currently unused). The color look-up table of the image can be modified using menu_bucol.gif (1699 bytes)

Note, however, that MOS is the recommended lookup table (also VOLCANO was found  to be useful in case of very faint targets).


IMDI interface


When IMDI is started, the basic setup information (CCD, grism table, etc ...) are read from a TNG/DOLoRes setup file provided with the release of the mask design software. Just choose your grism from the menu. The spectral range can be edited by changing the upper and lower wavelength in the small windows (you must hit the RETURN key in each  field, due to a known limitation of IDL).
Note. You may notice that the default wavelength ranges for the grisms are different from those listed in the DOLoRes documents. This is not a mistake. While the latter are effective ranges determined by the size of the CCD in the case of a central (long)slit, the default grism ranges for the MOS are related to the wavelength range covered by the grisms. 
Some data on the target field are read from the FITS header and displayed in the middle part of the main panel. This information will be inherited by the mask files and transferred to the associated documentation. 
Objects are selected by clicking once (left button of the mouse) on the main window.  Then click on the zoom window with the right button to center the target. The size of the box can be changed by using the middle button. Before that, choose the appropriate centering method menu_cent.gif (2079 bytes) in the main panel. Rectangles are used to outline the location of the spectra.  These are computed using information in the IMDI instrument database and the image header. The slits are interactively adjusted by the user under control of an anti-overlap algorithm.
Using "centroid", the slit appears mis-centered in the zoom window. This problem only affects the graphics, not the computed centroids, and should be ignored. 
The [import ascii] function allows importing of a list of targets from an ASCII catalog produced by automated image analysis and classification programs (e.g., Sextractor) or by hand. The first three columns must be IDENTIFIER, x_pixel, y_pixel. Any further columns are ignored.

Clicking on a target activates the Slit Control Panel. By this mechanism, you can either choose a new target or pick up and modify a previously defined target. The length, setting, and angle of the slitlet are defined by sliders in a self-explaining way. The minimum length of slits is 8 pixels (2"), even if you specify 0. After interactive editing, save the slit, delete it, or just quit without changes. The buttons menu_busa.gif (2182 bytes) and menu_bure.gif (2175 bytes) in the main panel allow you to suspend mask editing at any time and resume it later.  The "saved configuration" is a temporary IDL structure not to be confused with the final output of the mask design.

The IMDI anti-collision algorithm warns you in real time about spectrum overlap by highlighting the overlapping spectra. Our advice is not to pack the spectra too close to each other, letting some extra space between spectra. This will make reduction much easier ! 

Still, the final decision about accepting these warnings is left to user's responsibility, although there are software checks against slitlets violating severe mechanical constraints. In particular, slits too close to the edges are rejected by safety checks. An automatic slit optimizer is not foreseen with the first release of IMDI. 

Full resolution image
A useful feature allowed by the mask design software (and by Mask-Mode MOS observations) is stacking of two rows of spectra. Using a low-dispersion grism in combination with an appropriate filter to reduce the spectrum length, more than one spectrum can be aligned along the dispersion direction. (Note, however, that this mode is not offered as a standard facility of DOLoRes.) A multiplex factor of 2  will allow up to 80-100 spectra per mask (10" slits). 
Reference stars. At least two reference stars, well spaced across the dispersion direction, must be included in each mask. The presence of the reference holes (in practice, slits with minimal length) is required to speed up MOS pointing at the telescope. Of course, the two stars must be on the same coordinate system as the targets.


Each mask slit setup, once satisfactory, is saved to a mask-cutting file   using menu_buwr.gif (1777 bytes)     After finishing your design, for each mask you will end up with:

a mask-cutting drawing in the standard Drawing Interchange File Format (DXF), for instance held_AOT9_123_mask01.dxf

a documentation file, logging the DOLoRes and telescope setup associated with the mask, all useful information from the image header, and a listing of the slits in plain ASCII format (e.g., held_AOT9_123_mask01 .log);

a finding chart in postscript format (not yet available in the Window environment) (e.g.,

After packing all files and a detailed README file in one gzipped-tar file (e.g., held_AOT9_123 .tgz), the P.I. shall notify the DOLoRes Staff Astronomer who will provide file transfer instructions. Masks will be manufactured by TNG mechanics staff in advance of the run. Do not forget to specify in your e-mail the requested slit width (1.1 or 1.6 arcsec) for each mask.

return to MOS home page back to MM-MOS page