• This document applies to the Direct Imagers, CCDS, and TIFKAM.
• It does not apply to Modspec, MarkIII, or the 8k imager.

If you are using one of the applicable instruments, it is possible to center on targets quickly and accurately using offset guide stars. This can be a substantial time-saver even if you don't have a lot of targets. The procedure has been made straightforward and convenient, as described here.

The Ingredients:

• The USNO A2.0 catalog is a list of 500 million stars derived from scans of the POSS-I (1950's era sky survey). We have an abridged version of this catalog on the disks of Agung and Hill. The USNO A2.0 has high astrometric accuracy (typically 0.3 arcsec); some stars have moved since the 1950s, but most have scarcely budged.
• The MIS guider unit views an off-axis guide star using a movable mirror. It is easy to move the mirror to a precise location. The geometry of the focal plane is well-known.
• It is therefore possible to "dial in" a guide star ahead of time, instead of searching for one.
• What's more, it turns out that in practice this is so accurate that if you position the guide probe ahead of time, and you know where on the TV screen to put the guide star, you can center on your object to within a few arcseconds -- without ever seeing it!

How to Use the Simple GUI That Does This:

The description here is verbose for newbies, but it goes very quickly once you've done it a few times.

First, to set up:

(0)You can do everything from Hiltner's (or McGraw's) terminal, but you need to run the program on Agung (or Hill).

(1) Pop up a terminal window using the icon at the lower left of the KDE screen on Hiltner (or McGraw). In this terminal window, type

   ssh agung         (at the 1.3m,   ssh hill)

(2) If you have an object list that you have put on Agung somewhere, change directory to where the list resides:

   cd mydirectory

(3) Launch the Python GUI skycalc with

   skycalcdisp.py

Skycalcdisp.py will come up with some windows you may not want (e.g. the site chooser). To close any window in skycalcdisp.py, use the "Hide" button provided, not the "X" button on the title bar; if you use "X" you can't get the window to come back up.

(4) Near the bottom of the gui skycalc window you'll see a button labeled Guide Star Config. Press this to get a tiny window that sets the telescope and the rotator angle.

(5) If you have a pointing list, load it into skycalcdisp.py with the Get object list button. This is not strictly necessary, but it makes the operation vastly more convenient. The format of the list is identical to the MDM pointing list, viz.

   name_no_blanks  hh mm ss  dd mm ss  equinox

   grb090909  7 50 18.23  -0 12 13  2000

OK, now to use it:

(6) Enter an object's coords into the program. You can always do this by typing the RA, dec, and equinox into the boxes in the upper left, followed by the "Enter" key to refresh. If you do have an object list loaded, you can load your object by double-clicking it on the list, or by typing its (exact) name into the "objname" box in the main window and hitting "Enter".

(7) Press the Get Guide Stars button (bottom row). A temporary bright blue xterm pops up, and a big black xwindow appears like the one at the top of this document. The red lines show the guide probe's field of regard, with the catalog stars superposed on it. Click on your favorite star, and its guider XY coordinates will appear in the blue window as shown below. (In all likelihood the blue window will be blocked by the black diagram, so you'll need to move the black diagram to see the numbers; just grab its title bar with the mouse and drag it out of the way.)

(8) Selecting a good guide star:

• Stay away from the top of the diagram -- the guide probe loses track at Y greater than about 10500. The limit in the diagram appears to be a bit too optimistic. If the guide probe loses track, use the pink "Preset" button on XMIS to reset to "Origin", which re-zeroes the guide probe coords.
• I like to stay over to the left (low X) if possible, to avoid long throws and keep things consistent.
• Avoid stars at Y less than 3000 (bottom of the screen); there's a zone there that is vignetted, you won't see them.
• Avoid guide stars brighter than about 12 (but you can use them in a pinch).
• Ideally, a guide star will be part of a little pattern that will be clear when rotated upside down and backward (I think), so you can positively ID it, but this isn't necessary. If possible, avoid stars that will be ambiguous on the TV (e.g., members of close equal-brightness pairs).

(9) Go over to the xmis window (shown below). Select the beige "x:" field, and use the delete key (not backspace) to erase the number there; type in the guider X value of your star, and hit the "Enter" key. This moves the guide probe. Do the same for Y. Note that the blue field above the entry boxes gives the actual location of the probe (though it doesn't update until the probe stops moving).

(10) Set on your target by putting the guide star near the middle of the screen. Center up your target accurately - e.g., with a test exposure, or by viewing the target on the CCDS slit jaws. Once the telescope is pointed exactly where you want, use the numeric keypad on the guider computer keyboard to move the guide box onto the star. This will be the fiducial location for all your guide stars.

(11) Take your science exposure; check carefully that you've actually centered it correctly. When it's time to move to the next target, repeat steps 6-11; to summarize, just

• Look up a guide star,
• Move the guide probe to the expected location by typing its location into xmis. (If you're brave, you can type ahead during your last exposure, but Don't Hit Enter or the probe will make its move while you're still trying to guide!)
• Move the telescope to the field, and
• Use the telescope handpaddle to put the guide star into the FIXED box on the TV.

(12) To make the big black Xwindow go away, just type a q in the black Xwindow. This kills the process that skycalcdisp.py has spawned, and takes down the blue Xterm also. These windows take up a lot of room so you'll only want one set up at a time. The skycalc window will stay up and you can repeat the process indefinitely.

Cautions:

• If the guide probe is driven past its limits, it silently loses track of where it is. If you suspect this, simply go to the pink "Preset" menu in xmis to select "Origin"; this sends the guide probe up against its limit switches and resets the coordinates. It only takes a minute or so and can save you big trouble.
• As noted earlier, the rotator angle has to be accurately known for this scheme to work.
• Don't assume this is working without checking your images, or spectra, or whatever. If you're using a CCDCOM-based direct imager, there is a "centermdm.py" program that automatically matches the stars in the image to the USNO A2.0 and tells you exactly where you are. The MDM mountaintop web server hosts a manual for this program.
• Not all objects in USNO A2.0 are real, and some are galaxies that are misclassified as stars. If you cannot see your guide star, it doesn't necessarily mean your pointing is wrong -- just dial in another. If you can't see any guide stars, then there is something wrong (maybe you're not looking out the dome, or it's cloudy, or your pointing is off). Incidentally, the limitations of the USNO are a good reason for not selecting your guide stars in advance -- you'll occasionally get a bad one, and you want to be ready to select another "on the fly".
• A word to the wise -- at the start of the night, the 2.4m telescope encoders will usually have drifted by about a minute of arc in dec during the day. Scan north/south a little and your guide star should appear.
• The USNO A2.0 has some blind spots because it was compiled from photographic plates. Anywhere a plate would burn out -- right near bright stars, in the Galactic center, and so on -- it will be useless, and you're on your own. You'll have to scan for guide stars using the dx and dy fields in the Xmis window, and you obviously can't use them to control your pointing.
• Occasionally a USNO A2.0 star will have moved significantly in the 50+ years since the POSS-I was taken. This seldom happens, but you should keep the possibility in mind and check the first science exposure of each field (you should do this anyway to correct blunders of all kinds).

Skycalcdisp.py is useful in its own right -- you can learn it by playing with it, and it has an online help text and reference manual built in. There's also an html manual (with screenshots and the like) on the MDM mountaintop web server. The java version JSkyCalc is similar to skycalcdisp.py, but it does not have any MDM-specific goodies. Java code is extremely portable; JSkyCalc it will run on Macs and Windows machines as well as Linux boxes, provided only that you have Java Runtime Environment 1.5 or higher.