Embedded Files
These instructions are best used after you have read the "official" mask making software instructions (see below). These instructions here are written from the perspective of a new user, and you might get some added insight.
Some resources:
General:
http://users.obs.carnegiescience.edu/clardy/imacs/maskmaking.html
Specifically for creating object tables to feed into the mask software:
http://users.obs.carnegiescience.edu/clardy/imacs/catdata.txt
Note: On our "old" RHEL here at U-M Astronomy, I have found that the version for Fedora Core 5 seems to work. It is installed in the directory called masks. Currently, I have chosen to run the software in that directory, as opposed to fighting the "path battle". So I keep all object and reference tables (see below) in there and all generated files go in there.
There are detailed directions available in the masks/ directory (mginstr.txt). Here are my boiled down items for making masks. You still need to read those detailed directions.
The goal is to create the needed files(s) to send to Magellan so that they can cut a slits and square holes in a giant piece of aluminum. This "mask" goes into the IMACS instrument and is how the galaxy light from the field then gets dispersed onto the CCD for your analysis. To make this work, you need to know where the galaxies are and what size slit you want. Slits will have a width and a length. The width controls your resolution of the dispersed light through the grism (where the light goes after passing through the slit).The length of the slit depends on the galaxies themselves. You want them long enough to cover the galaxy and get some sky as well. But the longer they are, the fewer you can get onto the mask (crowding). You can also control angles (e.g., to match the position angle of a galaxy), but we won't bother with that here. Second, you need enough bright (see below) stars with acquisition boxes such that you can orient the telescope and make sure your mask is aligned properly on the sky. Proper alignment means that all of your galaxies will be centered inside their slit, but they are so dim that you need to use the alignment with stars (which require short exposures to see their light through the mask). Finally, you need to "fill" the CCD with dispersed galaxy light. The idea with IMACS is that you can keep your wavelength coverage small enough such that each set of columns on the CCD can contain TWO galaxies (see picture below). The number of columns dedicated to each spectrum depends on the slit width you chose and where it is on the sky.The number of rows depends on your choice of wavelength coverage (i.e., where they grating is placed relative to the CCD) as well as where the galaxy is on the sky.
Pictures are worth a thousand words. So let's start by showing you an example field we want to collect spectra within:
In the above imagem you see some white dots (targets), white squares (filled with black circles) and black circles (stars). The white squares filled with stars are the alignment boxes. The square is the field-of-view of IMACS. The curvey things at the top and bottom are where the guide cameras are. There need to be some bright (see below) stars in their.
In the following picture, we see the same view after the target algorithm has been run:
In the above picture, the green slits are where the galaxies are targeted and the red squares are the acquisition boxes.
Finally, this is what the CCD plane will look like with the dispersed light:
In the above image, it is important that no light from two slits overlaps (which would make for a very challenging reduction process). Note that the SECOND image above is what the mask actually resembles after it has been cut.
(1) Create the object tables using the IDL mask making software I created (attached here). In pure astronomy fashion: the code is the documentation. That said, you need to know the XMM OBSID, the names of the cluster targets, and the location of any matching catalogs (e.g., DR3 for the XCS). For NXS masks, I use three or more files: the galaxy catalogs for the XMM fields; the one or more galaxy catalogs per cluster; and the DR3 for XCS. Make sure the NXS OBSID matched the XMM OBSID (use Nicola's redshift pages to help with that). My IDL code priortizes cluster members as -3 for those with mags 21.5, and -1 from mags 21.5 to 22,. A priority of -1 is given for other galaxies in the field that are near the color magnitude relation, but not in the clusters. XCS point sources are matched to 21.5 and given priority 0. Finally, and this one takes some time, you will need to look at the NXS image and identify sky targets. This will help with background subtraction. You need to identify 6 or so areas that have no galaxy light . I put these in a separate object file and give them priority -4.
(2) Run intgui from the masks/ directory. Don't specify anything on the command line. You will see a new window open up that looks like the image below. Enter the shortened OBSID for the "Observ. File" name (drop the first 0 and the last digit in the real OBSID). Give "Title" the mask name the full XMM OBSID. Use your name as the Observer (no spaces).
(3) The following figure clarifies the values of the parameters I chose. I center the field to make sure the clusters are visible and hopefully so that the clusters sit within a CCD and not across a gap. The slit width should almost always be 1" for the right resolution with the R200 Grism (see the user manual for how to determine the resolution as a function of the slit-width). Slit lengths depend on the sizes of the objects. I like to make them wide enough to get some sky, but not wider than ~10 or 12". NOTE: Under "Options", you can click "SkyWin" to see the image of the target locations on the sky (the figure with the blue background above). MaskGen will try to make a mask and make the other two images above. It is nice to use as a guide, but note that it does not generate the files you need to submit. In order to do that, you need to run the task maskgen from the command line (see Step 5 below). You will need to click on "Object Files" to load in the catalog you created in step 1.
(4) Pick some alignment stars on the skywin view in the intgui. Usually 4 or 5 is good enough. I create a separate file for these reference data. To do this, I identify the stars by clicking on them in the SkyWin view, noting their magnitudes and positions, and placing them in a separate file I use then load after clicking the "Object Files" (i.e., I use multiple files per cluster, one containing the targets and another containing the alignment stars). I pick alignment stars that are ~14.5 in V (read right off the SkyWin plot).
(5) The MaskGen button in the IntGUI window runs a mask generation for you.. When it does, it will output something like:
Mask Making Software release 2.09 is Current software.
Process file 02154010.
GS1 at 921.6 99.3 is good
GS2 at 914.1 290.3 is good
== Writing to "./02154010.obs"
GS1 at 921.6 99.3 is good
GS2 at 914.1 290.3 is good
-Executing: "maskgen -q -c 02154010"
GS1 at 921.6 99.3 is good
GS2 at 914.1 290.3 is good
34 objects removed as FAR off detector.
92 objects flagged outside detector.
78 objects flagged as duplicates.
Objects - Slits Alignment Other
On Mask: 212 3 3
Removed: 877 1 0
Totals: 1089 4 3
** CONFLICT between Gap "Right_Gap" <and> Object "170151".
** CONFLICT between Gap "Right_Gap" <and> Object "170121".
** CONFLICT between Object "170717" <and> Object "171209".
** CONFLICT between Object "170745" <and> Object "170756".
** CONFLICT between Object "170745" <and> Object "171246".
** CONFLICT between Object "171115" <and> Object "171133".
** CONFLICT between Object "171141" <and> Object "171148".
** CONFLICT between Object "171246" <and> Object "170756".
** CONFLICT between Object "169642" <and> Object "169650".
** CONFLICT between Object "169668" <and> Object "169680".
** CONFLICT between Object "169668" <and> Object "169682".
** CONFLICT between Object "169680" <and> Object "169682".
** CONFLICT between Object "169712" <and> Object "170167".
** CONFLICT between Object "169712" <and> Object "169722".
** CONFLICT between Object "169712" <and> Object "169723".
** CONFLICT between Object "169722" <and> Object "170167".
** CONFLICT between Object "169722" <and> Object "169723".
** CONFLICT between Object "169723" <and> Object "170167".
** CONFLICT between Object "169768" <and> Object "170158".
** CONFLICT between Object "169768" <and> Object "170162".
** CONFLICT between Object "169768" <and> Object "170165".
** CONFLICT between Object "169768" <and> Object "169783".
** CONFLICT between Object "169783" <and> Object "170158".
** CONFLICT between Object "169783" <and> Object "170162".
** CONFLICT between Object "169783" <and> Object "170165".
** CONFLICT between Object "170158" <and> Object "170162".
** CONFLICT between Object "170158" <and> Object "170165".
27 remaining conflicts.
Writing object file with use counts to 02154010.obw.
Window created; title: (dbP 02154010) MaskGen Debug (V2.12)
Note here that there are 27 conflicts that HAVE NOT been resolved. These are my high priority objects (Priority -3) that it refuses to choose for me. Note that the mask generation resolves all other conflicts, but not those with the "MustHave" Priority. The ONLY way to resolve these conflicts is to edit your target file and either remove them, or change their priority. Ideally, you should do the latter and give them a priority of -2 and run again. But you could run into an infinite loop, so at some point, it is best to just remove some of the conflicts. Choosing which of these you really want can take some time (i.e., ALWAYS get the BCG). When you finally are able to run MaskGen (from the button on the IntGUI) without any conflicts, quit and save.
7. Run maskgen from the command line:
command-line>maskgen name i
where name is the shortname you used for the "Observ. File" when running intgui. This will create you final files that you then upload to the LCO website.
Page updated
Google Sites
Report abuse