A Guide to Making Slit Masks for the Double Imaging Spectrograph
      by Doug Ingram - University of Washington Astronomy Department

			(Last Update:  4/24/95)

This step-by-step guide is an attempt to summarize the technique I used to
develop a slit mask based on candidate objects detected in a DIS exposure.
Please contact me (ingram@astro.washington.edu [This address does not
exist anymore -Eric {deutsch@}]) if you have any input or questions
regarding this!

1)  Decide what your limiting magnitude will be based on the amount of
    time you have on the telescope.  In a 10-minute exposure at low 
    resolution, I was able to easily detect a V=20 quasar (fairly flat
    spectrum).  You can use this number as a benchmark to within a factor
    of two.  Note:  low dispersion on the DIS is 6-7 A/pixel.

1a) Until there is an reliable guider on the telescope, you should assume 
    that <50% of your time in the field will be spent integrating in spectral
    mode.  This is because, in my experience, the longest observation you
    can make with reasonable assurance that the star hasn't drifted out of
    your slit is no more than 10 minutes (this was constant throughout the
    night for me, though others have reported more problems close to the
    Zenith).  Rotation of the telescope is not a problem.  The other 50%
    of your time will be spent offsetting in imaging mode every 10 minutes
    to ensure your object is in the slit.  Provided the DIS is correctly
    switching between modes, which it has problems with.

2)  Select an object for the center of your field and assume that the slit
    covering this object will be centered at (0,0) in pixel coordinates.
    Find (x,y) offsets to the other objects in your field and choose an
    optimal slit width and length.  For our project detecting faint galaxies
    (r = 20-23), we chose 3" wide slits (because of the drift) and 18" long
    to get plenty of sky.  As it turns out, our slits weren't long enough
    to get enough sky pixels in the blue for the small slits (we ended up
    having to get sky from long slits at the edges designed for this
    contingency)...thus, I recommend 30" long slits.

    NOTE:  For the parity I describe here, you should REVERSE
           the sign of your x-offsets (see step 7).  

3)  Use the following numbers (preferably in the red for greater accuracy)
    to translate your pixel offsets into millimeter offsets:

	Scale in the slit plane = 170.6 microns/arcsec
	
	Red chip pixel size = 0.610 arcsec/pixel
	Blue chip pixel size = 1.086 arcsec/pixel

4)  With your map of millimeter coordinates and slit sizes, you should now
    be able to make a slit mask with any precision drafting program you wish,
    including IDL and Vellum.  The following step describes the process for
    using Vellum at UW:  (there is currently no IDL procedure to do this;
    however, the potential for automated mask-generation is there with a 
    little effort by the first brave soul).

4b) On the APO Observers' Disk on Apache (the observing Mac), you should find
    a template file for the slit mask in the Vellum folder (or you can find
    something in my folder, under Ingram).  Open this up and have someone
    show you the basics of Vellum (or at least show you a manual).  It would
    be too difficult here to describe exactly how to create your file, but
    Vellum is not that hard to use (maybe 30 mins tops to learn what you 
    need to know).  It is similar to other Mac drawing programs, in case
    you are familiar with MacPaint, MacDraw, Canvas or Photoshop.

4c) When your slit mask is ready in Vellum, print out a hard copy to check
    visually that everything looks reasonable.  If it does, then print out
    another copy at 100% size to a postscript file and then put that
    postscript file onto a public machine, such as hardy or stein.

5)  Once the file is ready for printing, I recommend a 2540 dpi printer
    if you can find it.  Looking through a microscope at a 600 dpi printout
    of my slitmask, the edges were very rough and could possibly have
    caused unpredictable effects in the dispersion.  

5a) At the UW, you can have your slit mask printed out on transparent film
    negative (2540 dpi) at the "Print Plant" or UW Publications Services.
    The print plant is located along the Burke-Gilman trail about 1/2 mile
    to the West of the PAB (ask me for specific directions).  In order
    to submit a job there, you have to do the following:

	- Ask them to tell you the location of the script (a Unix pathname)
	  that you can run to create a special "lino" directory on your
	  account that the Print Plant can access to print the postscript
	  file on their Linotronix.  You can create this directory yourself
	  by typing "mkdir lino" in your home (on, say, stein) and then
	  typing "chmod g+w lino" to set protections the way they want them.
	  Place your postscript file there.

	- Get a copy of the UWPS print job request form from the Print
	  Plant and fill it out (best to go over there the first time
	  and ask to talk to the coordinator for the Astronomy dept.
	  in person...I believe her name is Robin and her number is 543-0773)
	  He or she will lead you through it step-by-step, if you like.
	  If you are familiar with the process already, you can get a copy
	  of the fax order form from Eric.

	- In 24 hours, they should have your printouts ready for pickup
	  (they will also mail them if you trust campus mail services
	  not to mangle your printouts).  Take a ruler over there and
	  check it out to be sure the spacing matches your coordinate map.
	  Give them at least 24 hours lead time.  Printouts cost $11/page
	  for 2 or more 8.5 x 11 films.  Have a budget number ready.  You
	  can rush the order but it costs "a lot extra".

6)  What kind of results can you expect?  Well, for my slit mask, the
    dark region was very dark (at least a factor of 1000 light blocking)
    and the edges were very sharp under a microscope.  I do not know yet
    the transmission characteristics of the film, but it looks to be
    very good in the blue (>90%...I have sensfunc's posted for my standard
    stars taken through this film negative...they are near Eric's office)
    In any case, I don't think it would be all that difficult to use
    the film template directly and just cut out the slit holes with a
    very sharp knife down in the shop (but I have no experience with this).

6a) My first measurements of the acetate printed out on the Linotronix
    indicated that it was 2% too large.  Also, the person I worked with
    at UWPS told me that the last time they calibrated the print size,
    it was off by 1 part in 64.  However, when the slit mask was put in
    down at APO, all of my slit locations were within 1 pixel of my
    desired coordinates (i.e. as close as I could measure), and the offsets
    were large enough so that a size discrepancy as large as 2% should
    have been detectable.  So I see no reason to play around with different
    magnifications.

7)  Once you have the slit mask printed out onto transparency, it is
    simple to install.  The edges of the aperture are marked in the file
    and will be placed down on a small ledge in the slit wheel (these
    lines will not be visible on the chip).  The proper parity, I believe,
    is that the slit mask should be installed right side up into the slit
    wheel in the bottom window of the slit wheel if you are standing in
    front of it.  You should take an image of the slit mask assembly with
    low illumination to confirm this.  In any case, flipping the slit mask
    over to another parity is a trivial task once you know which way you
    need to go.

8)  Total time from imaging your field to having a slit mask in the slit
    wheel down at APO can be as short as a week (at first), but I recommend
    more than that if you want sleep.  Federal Express delivers to APO
    directly if you need to use such a mailing service.  Once you have
    practice at this, the turnaround time could conceivably be much shorter.

9)  Pray for clear nights.