osmos_inner_slit_observing

The inner slit is used because it disperses the spectra to the red side of the CCD where it is more sensitive. The tricky part about using the inner slit is that you need to place the target well of the center of the pointing (assuming the pointing is the target. So the offsets are huge.

I'm acquiring with John Thorstensens documentation, which is available here. But I found that some details are missing, and so this page helps (go to Long Slit Acquisition). Some folks have suggested that the oalign scripts work faster. I have not tried those for Long Slit Acquisition.

The order of operations I describe was suggested by Mark Wagner and is consistent with their protocols for ToOs. Note that they use roi4x1k to drop off the top and bottom of the CCD and make readout faster.

During the day, take biases and flats

In prospero window (<> are filled in by you, not including the < or > ):

ArXe arcs:

Choose your configuration and set-up:

pr> ccdbin xbin=1

pr> ccdbin ybin=1

pr> filename <you choose>

pr> call roi4x1k

Get ready for the Arc:

pr> comp ArXe

pr> exp 90

pr> filter2 6

pr> filter1 6

pr> disp 1

pr> slit 1

The filter commands are supposed to be for the "OPEN" slots. Confirm this with the little slip of paper below the monitor and on the Prospero Status window.

The disp command puts in the VPH grism (again check paper/Prospero)

The slit command puts in the slit (again check paper/Prospero)

Below are what your Prospero Status and XMIS windows will look like:

Now, go to to the XMIS window and turn on the lamps and put the finder mirror in.

pr> mgo 5

As the data are finishing, take a look in ds9. Here is an example:

Peak counts are 3.5k in the bottom left on top of a bright line. This might be a little low for your needs, and so you might need to increase the exposure time accordingly.

After you take the ArXe lamps, switch to HgNeNe in the XMIS window:

In Prospero, change your set-up slightly to reflect the different lamps and required exposure times:

pr> comp HgNeNe

pr> exp 1

You should see the Image Name/exposure time updated in the Prospero Status window:

Take the data.

pr> mgo 5

As the data are finishing, take a look using ds9 (note the really bright lines compare to ArXe):

Peak counts are ~25k in the bottom left on top of a bright line.

Now, take the flats. Turn off the arc lamps and turn on the Flat lamp in the XMIS window:

pr> flat flat

pr>exp 600

Confirm changes in Prospero Status window:

As the data are finishing, take a look using ds9:

Peak counts are around 20k in the bright region. Notice the fringing (this is the MDM4k).

Note down everything in your log. An example log file is:.

Next, we acquire with John Thorstensens documentation, which is available here. But I found that some details are missing, and so this page helps (go to Long Slit Acquisition). Some folks have suggested that the oalign scripts work faster. I have not tried those for Long Slit Acquisition.

The procedure is summed in these steps:

(1) Take a direct image of the slit (this can actually be done with a lamp in the daytime).

(2) Start the guider and take a short direct image of your field.

(3) Calibrate the WCS of the direct image of the field.

(4) Using this calibrated image, find the RA/DEC of where you'd like to place the slit.

(5) Stop the guider and move the telescope to this new location. Re-start guiding.

(6) Move the telescope in very small chunks to get the object right on the slit location.

Note that OSMOS has 4 amplifiers and the spectrum is dispersed across the CCD. You might want to have the object be above or below the center-line where the amplifiers meet. Ditto left/right if you are using the center slit.

Let's begin by going to your target location. Not a bad idea to offset to a bright star once you are there, go into "movie-mode" and re-center the pointing to that bright star. Then go back to the field and start guiding.

Put the disperser in, but leave everything else open. I've found that you need roi4k to get a good WCS in every image. You probably also one 1x1 binning.

The set-up is:

pr>slit 1

pr>filter2 6

pr>filter1 6

pr>disp 6

pr> exp 5

pr> ccdbin xbin=1

pr> ccdbin ybin=1

pr> filename <whatever>

pr> object acq

Take the image of the slit.

pr> go

It should look something like this:

The slit is the vertical bar. Notice that it is on the far left of the CCD. Note that this is a picture of the flattened image (not the one you took) to bring out the slit more.

Now, pull the slit out and take the direct image of the field:

>slit 6

>go

It should look something like this:

The object I am after is in the Green circle. This is the unflattened image, which is why you see the four amplifiers. Also, this image does not have a WCS.

Now, kill ALL of your open ds9 windows. Then, in a terminal, make a directory called "mycenters" and go there. Copy the oscrtrtask.py python script there:

mdm24ws1:obs24m% cp /usr/local/pkg/thorsoft/scripts/osctrtask.py .

Type, pyraf and then

-->pyexecute('osctrtask.py')

and then:

-->epar osctrtask

which brings up a window like this:

Fill in the numbers. The acqnum is the NUMBER of the image without the slit. slitnum is the number WITH the slit. You'd better have given them the same filename prefix (e.g., "d." such that d.0017 is the sky image and d.0016 is the slit). Fill in your target positions. And hit execute. You need to confirm the images in the terminal window and then it runs by itself.

 Here's the first image you see:

Notice how much better the image looks. I have moved my cursor to a point on the sky (in the cyan box in the zoom window). My goal is to point over here and get the new sky coordinates, such that the telescope can be offset to get my target right on the offset slit. Note that oalign and oscrtrask will calculate the shift in dx and dy (used on the guider). But since these shifts are large, I prefer to move the telescope closer first.

This pixel location that I have been using to get me close to the slit is: x=727, y = 1922. This stems from the fact that the slit is located at x = 3319, while the center is at 2065 and I want the y position to be at 2124 while the center is at 2032. Note, these numbers can be determined to higher precision to get the telescope moved closer to the offset slit position. Once I read off these coordinates at this pixel location from the calibrated image, I enter those coordinates in the XTCS window:

Then, send off the coordinates. Stopping guiding, move the telescope, and start guiding again. Hopefully you have offset the telescope to get you close to the slit.

Now, take a new acq image (w/o the slit) and run oscrtrtask as it should nominally be run to get the dx and dy offsets to apply to get the target precisely centered on the slit.

Now, adjust your set-up accordingly:

pr> disp 1

pr> slit 1

pr> call roi4x1k

pr> object <name>

pr> exp 1200

such that Propsero Status looks like:

Take some science images.

pr> mgo 3

Hopefully, they look like this: