OSMOS Mask Making

These descriptions for multi-object mask making is specific to the C4 effort and measuring the caustics. However, the are probably of general use to anyone making multi-object masks for OSMOS on MDM 2.4m.

The OSMOS Mask Software (OMS) is available here: http://www.astronomy.ohio-state.edu/~martini/osmos/oms.html It is very important that you use the latest version.

Step 1: Gather your ingredients: the images, the star catalogs, and the two galaxy catalogs (one with SDSS redshifts and one without). For the C4 project, we have created MOSAIC SDSS r-band fits images around the C4 clusters. The images go out to a  radius of 1 virial radii (2.0*RHO200_RV1500/0.75). The star catalogs go out to a 1 virial radius or 60 arcminutes, whichever is smaller. The galaxy catalogs go out to 1 virial radius. Example files are attached below. Note that the format of the data files must be matched exactly and are described in the OSMOS documentation (see External Links on the side-bar).

Step 2: Assuming you have OSMOS installed, setenv OMSROOT /Users/christoq/share/oms_v170  (or whevever you installed it) and run oms.sh in the bin directory of the OSMOS distribution. this should bring up the following window:

Step 3: Load up an image (File -> Open). You will probably need to adjust the levels some. I found it best to zoom in on a bright galaxy (e.g., the BCG) and click "Auto Set Cut Levels", the use the left mouse button (on Linux, Firefox) to drag across the bottom color bar to get it looking ok. You can also go to View -> Cut Levels and click 98% or whatever and then adjust accordingly. Similarly, you can change colors and scale levels by clicking View->Colors. For instance, the color "ramp" is standard B/W. An example loaded image looks like this:

Step 4: At the top of the winow, click OSMOS -> Init Mask. A geen shaded circle might appear in your viewing window (depending on your zoom level). Zoom out to see the OSMOS footproint:

Step 5: Make sure the instrument is set-up properly. For now, we are using straight slits that are 1" wide and 10" long, using the "Camera", the VPH grating, and the g-band filter (since we want to make sure we get the H-K break at these low (z<0.15 redshifts). Go to OSOMS -> Config/Layout to adjust:

Step 6: Load in the reference star catalog by going to Data Servers -> Local Catalogs -> Load From File and choose a star catalog appropriate for this cluster. Once loaded, you should see a data window:

To be able to see these data on the image, you will need to set a plot symbol option. Go to Options -> Set Plot Symbol and it will bring a new window. I choose purple circles that are 40 pixels wide for the reference stars. I also had to add the RA and DEC columns to the Used table (by highlighting and using the arrows).:

After clicking "Add Symbol", then clock "Apply" and "Ok" (you must click ALL 3)  the stars will be identified on the image with purple circles and a single star will be brightened when you click on a specific star in the catalog window:

The next step is to load the galaxy catalog (denoted as the 'target' catalog). Make sure the IDs in the galaxy catalog are different than the star ID's, or you will be highlighting both when you click on something in your catalog pop-up window. This is done the same way as for the reference stars outlined above (load catalog, assign symbols, etc). The "X"s note the galaxies we want to observe.

You have a choice of picking which stars to use as reference stars/guiding stars first or assign slits to galaxies. Dan chose to do the latter. Because we are interested in the high density regions of our clusters (for fixing completeness), it may be wiser to skip the "auto slit" function and place the slits by hand over the galaxies of interest. Having the galaxy catalog pop up window to highlight galaxies makes it easy to see which ones you want.

Knowing that the spectra project along a 1D space on the CCD, by eye it is fairly easy to optimize the location and rotation angle of the mask and which galaxies you observe to get the most spectra per mask. Moving the mask is done by clicking on the filled green circle in the center of the mask and dragging it to a new location. Rotating the mask is done by clicking the filled green circle in the center of the mask with your center mouse button. The amount the mask turns each click is specified under OSMOS -> Config/Layout -> Rotation Step.

WARNING: When placing slits, the target galaxy can't be too close to the sides of the mask - if it is, part of the spectrum we need will extend off the detector. The closest a target galaxy should be to the side of the mask is roughly halfway between the white line and the edge (note this applies to the sides, not the top and bottom). You'll have to take this into account when positioning the mask over the target galaxies.

WARNING: The instrument cannot be rotated about the full 360 degree motion. To be safe, the rotation angle should be +/-80degrees.

Now you are ready to place slits.  Here is what a galaxy target file will look like:

ID      RA      DEC     R_MAG   BIN     BIN_NEEDS       PRIORITY

--      --      ---     -----   ---     ---------       --------

100     08:48:51.286    +09:01:21.64    14.6596660614   1       1       1

101     08:49:11.847    +09:03:29.93    16.7495632172   1       1       1

102     08:48:55.737    +09:09:47.03    15.4274330139   2       1       1

103     08:48:34.009    +09:06:26.10    16.3035373688   2       1       1

104     08:49:21.465    +09:04:33.99    16.8543510437   2       1       1

105     08:49:25.716    +09:15:46.70    16.0981082916   3       2       1

106     08:48:24.923    +08:51:33.58    16.5887012482   3       2       1

107     08:48:29.075    +08:35:21.97    16.5973129272   5       0       1

The galaxy catalogs have a column which lists the total number of galaxies needed in that bin and another column listing the bin number.  Our goal is to get the required number of galaxies in each bin and a few that already have spectra (which are in a separate file). In the above example, bin 1 needs just 1 galaxy and either of the first two galaxies will suffice. Bin 2 needs 1 galaxy and you have three to choose from. Bin 3 needs 2 galaxies and we really want both. Bin 5 doesn't need any galaxies, but there is one to observe.

Start with the bright galaxies in each bin and place slits until you reach the BIN_NEEDS. After those are targeted, fill out the mask as best you can. To place a slit manually, click the OSMOS -> Config/Layout -> Center Slit option and a pop up window will show up called Pick Object.

Your cursor will be different, and by scrolling over a galaxy in the image, you can click on it to place a slit. (Note: you may have to click twice). 

A yellow slit and two parallel red lines will be placed over the mask. The red lines represent the region the spectra will be projected over and no other slits can be placed. This process can be repeated (by clicking Pick Object at the bottom of the Pick Object pop up window) until you have as many galaxies with slits over them as possible. If you want to delete a slit, just click around it with your center mouse button and a pop up window will ask you if you want to remove it. Warning, its a little finicky, so you will have to click around a bit.

Now you need to pick your reference stars. Go to OSMOS -> Config Layout  and change the slit width AND slit length to 5" each (a box). With your reference stars loaded into the window with their own plotting symbol, pick 5 or so reference stars by clicking on them just as you clicked to add the galaxy targets. Note in the following image how a 5x5 box is drawn around the star (and the red lines are also drawn on as well.

DO NOT USE THE OSMOS -> Reference Target  option. It DOES NOT CREATE THE RIGHT MASK.

You can then repeat this process for as many stars as you want, but 4 or 5 should be sufficient.

Then you need to pick a guide star. Go to OSMOS -> Guide Stars -> Add and click ok at the pop up window. This process will work the exact same as above picking reference stars, except you are looking for a star in the light red box which is slightly offset from the mask and bigger. 

The manual says to try and pick a star in this box, but outside the mask field of view. One is sufficient.

After you have your main targets, reference stars and guide star set, target a few galaxies which have been observed by SDSS. You can do this by loading in the name_redshifts.cat file which contains all of the galaxies in that region with SDSS redshifts. We want a few that have been observed to check our reduction algorithms.

Finally, there is no need to waste sky, so fill in the mask with any object that looks like a galaxy and does not have an SDSS redshift.

To name your mask, go to OSMOS -> Config/Layout -> Project Name. Name is after the cluster. We are naming them after their C4 row IDs (4 digits, like 0024). Then go to OSMOS -> save OSMOS and click through the Ok's to finish saving the mask. The mask files will be output to a directory specified in the .sh file, but on my machine the path looks like so: 

/Users/giffordw/.oms/SET/osmos.1.1291test.oms

You will want to collate these mask files (.oms, .gbr, and .epsf into a single C4_NNNN/mask directory so that we do not lose track of them for later).

To view a mask you've already created, go to OSMOS -> Load Setup and load one of the .oms files that you created. To go back to mask making, click OSMOS -> Reset Mode ot OSMOS-> Quit Mode and than OSMOS -> Init Mask (note that this will DELETE the mask in memory (but not if you saved it to file).

Here are some screen shots of a completed mask:

Zoomed in and a new color-scheme to highlight the slit (red) and the dispersed spectral region (green):

The following zoom-in shows some reference stars near the field center. Notice the reference targets are the boxes while the wider slits are the galaxies:

JNote the yellow circle away from the field-of-view: it is the guide star.

At the bottom of this page you will find all the files you need to try it out. Note that the image

is compressed using FPACK and so you will need to FUNPACK it (CFITSIO).

One final note, OSMOS uses a GERBER-style file for machining the masks (this is one of the files that is output when you save an OSMOS mask).  It is worth examining the .gbr file using gerbv to see that all of you slits are actually in the circular mask region. For instance, in the case below, there is a slit on the edge in the upper left corner (with a box around it):

[NOTE: I was unable to build GERBV from source on RHEL 5.*. But I was able to download from the RHEL repositories (yum install).]

This "bad" slit should be removed the mask file using the OSMOS software. The best way to do this is to use GERBV and select the bad slit (draw a box around it). It should change to a new color (white on my screen). Then click Analyze -> Display Selected Object and it will appear in the left hand "Messages" panel. There should be an identifyier, like D24. The .gbr files appear to start their labeling at 11 (instead of 1). So this is actually slit #14 in the OSMOS mask. Go back to the mask and find this slit # (along the top of the mask), right click on the slit and click "Delete".