Observing procedures

This section gives practical guide-lines for conducting observations with SIS. Observing recipes summarize the recommended procedures for imaging, long-slit and multi-slit observations. The different steps are then separately presented and discussed.

Overview

OSIS is easy to use and reconfigure. However, this flexibility necessitates vigilance on the part of the observer to assure correctly set observing parameters and to properly conduct an observing sequence.

Recommended observing steps are given below. These have been developed from extensive experience, and have shown to be very efficient. We encourage you to follow these recipies and avoid short cuts which can waste more time than they save.

The basic recipes are, in fact, not very difficult. The most complex sequence of observations with OSIS (i.e. in its long slit or multi slit mode) can be divided into 10 successive steps for a given field:

• Focussing
• Field acquisition
• Guiding
• Direct imaging exposure
• Aperture mask design
• Aperture mask drilling and installation
• Centering in the slit(s)
• Spectroscopic exposure
• Calibrations
• Data evaluation

Recipe 1: Summary of steps for direct imaging

Here is a condensed summary of the sequence of actions recommended to take direct images with OSIS:

• Rotate cassegrain bonnette to proper position angle (the position angle can be determined in advance to assist in finding a suitable guide star).
• Select Filter in the OSIS form (and no grism, no aperture mask).
• Focus telescope with CAF, on a star close to your field or on the field itself.
• Acquire field. This could be done with binning, but detector reading time is now short enough to avoid going to binned mode. Centering is very accurate if the coordinates are accurate and if a reference star was pointed to first.
• Start guiding, either with Cassegrain bonnette only or with active guiding.
• Offset with OSIS-OFFSET (with telescope or OSIS) to precisely center the desired field. This can also be used for shifting the field on the detector in a sequence of exposures.
• Exposure: select the proper image raster with RASTER, then enter the desired exposure time and object name in the EXPOSE form and click on "accept" to start the exposure.
• Obtain calibration exposures: you need biases (and darks), flat field images as well as images of photometric standard fields taken the same night to properly calibrate your data.

Recipe 2: Summary of steps for long slit or multi-slit spectroscopy

• Take a direct image, full frame and binning 1 x 1 following the previous "recipe", from which the aperture mask could be defined.
• Design aperture mask by working on the image in a LAMA session. Do not forget to add one or two reference apertures (round or square) for centering stars.
• Drill and install aperture mask: drill the aperture mask with LAMA, then install the mask in the slide and the slide in OSIS.
• Reacquire field, with the same bonnette orientation and the same guide star as for the direct image at step 1.
• Take a field exposure without mask. Measure the position(s) of your centering star(s).
• Take an image of the mask using the sky as a background. Measure the center of the corresponding reference aperture(s) and compute and perform the required offset. Avoid using the halogen lamp in that step. Exposure time is shorter, but you loose time for moving in and out the calcor mirror, and you loose the guiding.
• Center targets in the slit(s) by taking a new image through the mask after OSIS-OFFSET. If the star is not perfectly centered in the aperture, offset again. note: If you are confident in the reproducibility of pointing at step 4, you can skip steps 5 and 6 and immediately try to obtain an image through the mask.
• Select grism and filter from the OSIS form. A filter may or may not be necessary.
• Spectroscopic exposure: enter the desired exposure time and object name in the EXPOSE form and click on "accept" to start the exposure.
• Obtain calibration exposures: you need biases (and darks), spectroscopic flat fields, wavelength calibration spectra, direct images of mask (eventually obtained at step 6) as well as spectra of spectrophotometric standard stars to properly calibrate your data.

Cassegrain bonnette rotation

The Cassegrain Environment, containing the Cassegrain bonnette (with guide probe, etc.), the entire MOS/SIS assembly, as well as auxiliary and support equipment, can be rotated to allow any position angle on the sky (the "bonnette angle"). Rotation of the Cassegrain environment is controlled by a hand paddle in the control room. Ask the support astronomer or the O.A. for assistance.

Bonnette rotation with SIS should be carefully considered in spectroscopy mode. A single step of the bonnette rotation encoder is 0.05 degree, hence the repositioning accuracy could be, in principle, of the order of two pixels over 2048 pixels on the CCD. In fact, because of mechanical inertia, it is quite difficult to stop at a given position angle with this degree of precision. Moreover, tests conducted in February 1995 show that, for two images taken with identical readings of the encoder, the residual rotation can amount to 0.2 degree. This is likely a more realistic value for the rotation accuracy and corresponds to about 7 pixels over a 2048 pixel field, or 0.6 arcsec.

This is not a problem if the alignment of the aperture mask with the object field can be done within these tolerances; however, if a mask is to be used over several nights, we strongly urge that a single position angle be maintained for all fields to be studied during this time. On the other hand, it may be appropriate to use position angles chosen for each individual field, to allow selection of guide stars so as to minimize occultation by the guide probe. In this case, we recommend that the entire procedure, from direct imaging, to mask creation, to the spectrographic exposure, be completed for a given field before rotating the cassegrain environment to a new bonnette angle. The situation is less critical for programs involving imaging only.

Field acquisition and centering

• Give your object coordinates to the Observing Assistant, and have him select the SAO star closest to your field.
• The O.A. sends the telescope to the selected SAO star.
• The O.A. centers the SAO star on the guide TV (the XY location on the screen corresponding to the center of the CCD for your run should be well known; make sure it is so with the T.O. or support staff).
• The O.A. will then perform the ``Local Pointing'' operation: this measures the pointing-error offsets for the SAO star, and automatically applies them to your object coordinates.
• If your object is bright enough (limiting magnitude in V is typically 19 to 20, depending on the seeing) it will appear within a few arc seconds of the CCD center position on the guide TV monitor. If the object is not bright enough for the guider TV, or if there are uncertainties in the coordinates, we recommend that you take a short CCD exposure in a binned mode (4x4)to confirm that you have indeed acquired your target.
• The O.A. will find a suitable guide star for the cassegrain bonnette guiding. Once your object is seen on the CCD, any further offset is better done with the OFFSET procedure and, for precise centering of objects into the slits, with OSIS-Ofst (see chapter 5)).

Focussing

Running CAF efficiently

• Check to see if there is an appropriate star in your field: take a 1 s exposure in direct imaging mode (no filter) with full raster and 4x4 binning. Any star with a maximum intensity of at least 20000 counts in such pictures can be used for focussing. If there is no such star, ask the T.O. to point to a nearby HST guide star (SAO stars are generally too bright).
• Integration time should be at least 10 s in order to "average" the high frequency components of the atmospheric turbulency. The maximum counts on star images with the bi-prism should be a few thousand above sky level in order to permit an accurate determination of their centroids. Adjust your exposure time to meet these requirements.
• Use fraster to select a focus subraster of at least 250 x 250 pixels; the separation between the images given by the bi-prism are ~120 pixels with SIS. If you measured the position of your focus star on a frame binned 4 x 4, do not forget to multiply the coordinates you have read by four.
• If one of the two images given by the bi-prism falls on a bad column of the CCD, it is best to move a little in the X direction (use an offset or direct a telescope motion from the O.A.); a motion of a few arcsec is enough.
• Normally, you have to run CAF only once. If another exposure seems necessary (cf. above), click on "Quit" and wait for the disappearance of the CAF exposure window. Starting the next CAF exposure too early can result in a bad configuration (bi-prism not in place).
• The sequence of commands in CAF is as follows:
  • click on "cursor" in the SAOIMAGE window
  • click on "region"
  • click on the left image given by the bi-prism (or on every left image of the pairs if several stars are measurable in the field)
  • click on "write"
  • hit "return" on the keyboard
  • hit "q"
  • Change the telescope focus (the position of the Cassegrain secondary mirror) by the suggested value with the handpaddle.
  note: CAF uses the iqe function for computing the centroids of the images. The next time you open the iqe window, you will have to change the file name back to "current.fits" as well as the parameters (activate fwhm and other options).

"Manual" focussing: IQE

If for some reason CAF is not available when you are observing, you can focus the telescope manually.

To do so use a focus star of magnitude 14 -- 16 and take successive exposures (with at least 10 s exposure times), changing the telescope focus with the handpaddle by steps of ~5 to ~10 telescope focus units between each exposure. Using IQE, compute various image parameters for each exposure. The image statistics of importance are the FWHM's of the image along the X and Y axes (or along minor and major axis of the best fitting ellipse) and the maximum (peak) value. To be in perfect focus, you need (i) a perfectly symetrical image on the display monitor, (ii) X and Y FWHM's as small as possible and as identical as possible, (iii) a peak value as high as possible.

Offsetting: OSIS-OFFSET

• Take an image of your field with proper raster and binning settings.
• Measure the coordinates of your object on the CCD. This can be done simply by putting the cursor of the SAOIMAGE window over the star or, more accurately, by using the centroid option in IQE.
• Convert the measured coordinates by taking into account raster and binning. The coordinates to be entered in the OFFSET form are always without binning and full frame. For instance, if you use a full frame and 4 x 4 binning, the measured coordinates should be multiplied by 4; if you use a subraster 200 x 200 centered in (1024,1024) and no binning, you should add 924 (= 1024 - 200/2) in X and Y to the measured coordinates. We recommend that you not use complex combinations of subraster and binning options if you want to avoid mistakes at 4200 m!
• In the OSIS-OFFSET form, enter the initial coordinates measured and the final coordinates desired. The displacement will correspond to the difference between final and initial coordinates. So, if you want to move by X = 20 pixels, you could just enter Xi = 0 and Xf = 20.
• Select the desired mode of offseting (telescope or OSIS), and "accept".

Guiding

With Cassegrain bonnette

Although you will probably prefer to use the active guiding system provided with OSIS, at least in the visible, it is also possible to guide with the Cassegrain bonnette, i.e. without image improvement. This could be the only choice for (very rare) "empty" fields down to magnitude ~18 or in case of heavy atmospheric absorption that reduces the limiting magnitude (hopefully, also a rare circumstance). In such special cases, the Cassegrain bonnette can search for a guide star in a larger field. This is also the normal procedure for taking centering exposures and for keeping good telescope tracking while you are searching for your guide star with the OSIS probe.

The O.A. is normally in charge of moving the Cassegrain probe until a suitable star is found. He usually records the XY position of the Cass bonnette, as well as the XY position of your star on the guiding TV. This will save a lot of time for centering if you plan to come back to the same field on a subsequent night, as it will ensure that the telescope is on the same location on the sky.

The reproducibility of the recentering is a few pixels.

With active mirror

• Make sure that the oscilloscope is on; if not, turn it on and leave it on for the entire run.
• Find an appropriate guide star with the cassegrain bonnette and start normal guiding (O.A.). From now on, the cassegrain guiding will continue running, until switched by a TCS command from OSIS, just before the exposure is begun.
• In the "Guider" window,, move the OSIS guide probe to Garage position.
• Take a full frame image of the field (eventually with binning) and choose a suitable star for fast guiding,
• Execute OFFSET to send the guide star outside of the field of the CCD, at chosen coordinates Xf , Yf.
• Enter the Xf and Yf coordinates of the star in the X and Y field of the guider window. Select Guide_star to send the probe to the guide star and initiate fast guiding.
• Check to see if the guide star is properly centered by looking at the oscilloscope and at the percentage of flux in each quadrant (activate the "update" box for new values). The total flux normalized for 1 s integration time, percentages in quadrants and position of the star's centroid are displayed.
• If needed, recenter the guide star with the control arrow, after having selected the steps in X and Y with the slide bars (two or three units on each axis if you are not too far from center). Note: good centering means equal flux in each of the four quadrants, but not necessarily a maximum of the total flux, because of the gaps in the lenslet array assembly. This is especially true with excellent seeing. The best indicators are the oscilloscope and the percentages per quadrant.
• Choose the integration time for fast guiding according to the total flux of the star.
  Typical values:
  • 100 ms integration can be used if the net flux (above sky) is at least 1000 counts/s.
  • 30 ms can be used above 3000 counts/s.
  • 10 ms (the fastest reasonable rate) above 10000 count/s.
  • This assumes a dark sky (level around 3000 counts/s); during grey time, use longer integrations to obtain the same S/N.
• If you want to observe in spectroscopy mode, you can now take a direct image of the field, an image of the mask, and compute the offset needed for centering the objects in the slits. Use OSIS-Ofst since you are under fast guiding. In imaging mode, you can go to next step.
• Ask the O.A.to switch to OSIS guiding (i.e.: he activates "Agoff", then "Moson" on his console). From now on, the low frequency guiding is controlled with the error signals sent by the OSIS electronics.
• Start science exposure.

Imaging exposures

A direct imaging sequence with OSIS proceeds as follows.

• After having focussed and centered your field, make sure you are guiding with the Cass Bonnette or the active mirror (previous sections).
• Select the proper filter in the OSIS form. Be sure to also select an empty grism position and an empty mask position.
• Select the proper raster and binning (RASTER form).
• Enter the exposure time (in seconds), the object's identification, and any desired comments in the EXPOSE form.
• Check again in the feedback and status windows that everything is ready according to your requests.
• Click on "accept" to start the exposure.
• After a few seconds, an expose window will appear, with the exposure time being counted down. When the exposure is over, it will say that the CCD is reading. Then, normally, the image will be displayed as a grey scale in the SAOIMAGE window.

Spectroscopy exposures

The sequence is similar to direct imaging, except that you should have centered your objects in the slits with OSIS-Ofst and chosen a grism.

When taking several long exposures of the same field, we recommend that you check the position of the reference star(s) with respect to the mask between each exposure, since instrument flexures, although small, are cumulative. Such a check does not take a long time if you use a sub-raster around your reference star.

Aperture mask preparation

Selecting parameters in the LAMA session

Once an image of your field has been acquired, you can process it to prepare the mask. This is done on another HP terminal with the LAMA account (login: lama; same password as the SIS data acquisition account).

The Lama session manager menubar with its accompanying icons is displayed.

FIGURE 21. The LAMA menubar

First select the setup form. It asks for the instrument in use (MOS or OSIS) and for the CCD name (the important parameter here is the pixel size; it is automatically recognized from the name of the device). This will set the scale (i.e. the correspondance between pixels and arcsec.) for the mask design.

FIGURE 22. The Lama setup form

You then need to go into the Grism ENG form and give the grism identification and parameters of your spectra: i.e. the central wavelength and wavelength range you want to cover. This will set the limits of spectra that will be overlaid on the field image when you select your objects. This is useful for defining the area of full wavelength coverage, or when you want to cut two or three series of slits per column with low dispersion grisms and/or wavelength range limited by a pass-band filter. However, be aware in that case that zero order images of the slits corresponding to a given series could fall on the spectra of another series.

To design a mask, select lama mask in the menu bar. The form will require the following inputs from you:

• Select the name of the fits file corresponding to your field image (without ".fits" extension, but with the file type letter, usually "o" for object image). It could be, for instance: "299742o". Select the extension for the name of your mask design file. The name of the final file will then be something like "299742o.l0"; this allows you to have different mask files for the same field with extensions .l0, .l1, etc.
• Select the width and length of the slitlets, the size of the point apertures, and the width of the curved slits (this option is not yet working and can be omitted). All sizes are given in arcsec, as required by your program.
• Select "create YAG file from mask definition". After selecting "accept", the field image will be displayed in SAOIMAGE.

FIGURE 24. The Lama mask form

Preparing a mask file

• In the SAOIMAGE menu bar, you will now see a "MOS" button. Clicking on it will give you access to the mask design menu. You still have access to other functions such as changing grey levels with "scale" and "color" or zooming with "pan".
• Select, for example, "slit" in the MOS menu. This will display a GREEN slit with width and length you specified upon start-up in the image window. Drag the slit by pressing and holding down the left mouse button and positioning the cursor on the first object position; fine tune the slit position by moving the keyboard arrows, and then clicking with the left mouse button. To select this slit for later cutting type "s" on the keyboard -- this will make the slit turn YELLOW. To display another GREEN slit for another object, click with the left mouse button and start again. You will not be able to move any YELLOW slit again unless you type "d" (for delete or de-select) on the keyboard: this makes the YELLOW slit GREEN again and removes this slit from the list of slits to be cut later.
• Pan mode allows very accurate slit positioning. Click on "pan" and then on your selected object for centering the zoom window, then select high zoom magnification. Return to "MOS" for centering of the slit and validate by typing "s" on the keyboard.
• Also choose one or two "centering" bright star(s) in the field and
• center round apertures on them (by selecting "point" in the MOS menu and process as slits). These are invaluable to quickly set the objects into the mask apertures.
• Once you have completed your mask design, select "do it" in the SAOIMAGE/MOS menu. The program will then create the appropriate mask design file (for instance 299742o.l0 as previously specified) which may be recalled on any SAOIMAGE display at later stages, as well as a specially formatted "yag" file (299742o.l0y in this case, the "y" being for yag) for mask cutting with the LAMA machine. note: Recall that at this stage the aperture coordinates should lie in the range 105<X<1898 and 185<Y<1794 in order to be actually cut on the mask; outside of this range, the laser will try to cut them on the mask frame itself! It seems wise to take a safety margin of a few pixels more within these limits to avoiding difficulties in cutting apertures at the very edge of the field.
• With the "read mask" button, you can display a previous mask file and add new apertures (perhaps with different slit geometries), thus creating a new version of the file. In the past, computer crashes during the design of a mask were not unusual. Although the problem is less frequent now, it is always wise to save your data from time to time when you are creating a mask with many slits.
• When you are happy with the current version of your yag file, note this file number for cutting with the LAMA and proceed into the LAMA room.

FIGURE 25. Selection of objects and slit positioning in SAOIMAGE

Cutting the mask with LAMA

Ask your support astronomer to be present when doing it for the first time.

When everything is ready, enter the YAG file name on the terminal (299742o.l0y in our example).

Cutting time depends on the number of slits to be cut, their size, and on the number of passes made with the laser. This parameter can be adjusted in the cutting program, but the default number is four passes and normally produces very clean cuts. If, for any particular reason, you want to change this parameter, ask the support staff in advance.

After finishing the mask cutting for the night do not forget to shutdown the LAMA following the procedure in the manual.

Aperture mask installation

To remove the mask slide from OSIS, first, in the control room, send the mask slide to position 1 (open) with the OSIS procedure. In the dome, remove the octagon cover (pull radially on both handles, then rotate), and the completely unscrew the OSIS slide screws, while supporting the mask slide with one hand. N.B.: it is important to unscrew totally, even if it seems that the slide can be removed before that. It is obviously also quite important to prevent the mask slide from falling on the ground! Remove the old mask slide and insert the new one, pushing it all the way up. Screw the OSIS slide screws completely in again, and replace the octagon cover. (The latter might be the most difficult step if you are not used to it).

Back in control room, first enter the new mask identifications in the OSIS form. Then send the mask slide to the position of one of the masks and take a direct image with the halogen image lamp. This will allow you to compute the offset needed for centering the objects in the slits.

Calibration

Imaging: recommended calibrations

Spectroscopy: recommended calibrations

Flat-fields

Photometric standards

Darks

Biases

Spectroscopic flat-fields

Wavelength calibration spectra

Spectrophotometric standards

Data evaluation

IMAGE, IQE and GRAPH

The IMAGE form permits a grey scale display of a selected image. If the image name in this form is "current.fits" the last image transferred to the disk will be displayed. Normally, it appears automatically in the SAOIMAGE window as soon as the reading of the CCD and the transfer to the disk are over. If it does not, open the "MODES" icon and select automatic image display.

The IQE form allows for the computation of various statistics on a FITS file, such as the mean, standard deviation, and minimum and maximum values. It can also give a good estimation of the image quality. Select a non-saturated but bright star and a box size at least 20 pixels wide. The spatial indications returned by IQE (FWHM, centroid position, etc.) are in pixels.

With the GRAPH procedure you can check an object's profile in direct imaging. It also offers the capability to do a quick sky subtraction from an object's spectrum. To do so, select cut number 1; enter the position of cut on the object spectrum and width to average; similarly, enter the subtraction position and width to average at a location representing the sky background. Graph will then plot the spectrum of the object after subtraction of the average sky spectrum. Cuts can be along the X or Y axis and up to 3 different cuts can be displayed simultaneously, with or without sky subtraction, and for the same or for different files. File identifiers "current.fits" and "previous.fits" are permitted.