Learning the OT tool
In Phase II, is when you have to define concisely your observations; Phase II submissions are required to contain spectroscopic acquisition sequences and all calibrations. These observations are automatically generated by the Observing Tool (OT), although they will need customising for the PI's particular use case (e.g. adding telluric standard star coordinates).
The GNIRS team strongly recommends that users start from the automatically-generated OT templates and the GNIRS OT library when preparing GNIRS observations. The library contains detailed instructions for customising the template observations: changing targets, standard stars, slit widths, wavelengths, etc
In this section a generous information is provided with the aim that you will be able to use efficiently this software in order to define the observations of your program. The best way to approach to OT is to watch the suite of dedicated tutorial videos. Below, you can find a more detailed description of how to use OT for your GNIRS observations. However, if you still are still hungry for more information about Phase II in general, please don't hesitate to visit this webpage.
GNIRS in the Observing Tool
This section is intended to explain the various OT components in more depth:
As well as the science target spectroscopy, GNIRS programs must include:
- Acquisition observations for all targets
- Calibration observations: telluric standards, flatfields and arcs; any additional calibrations
Adding a GNIRS component to an existing observation, or adding a "GNIRS Observation" gives access to the GNIRS static component. The GNIRS component is used to define the basic GNIRS configuration and looks like this:
The available items are:
- The combination of pixel scale and central wavelength (below) determines which of GNIRS' four cameras is selected. 0.15"/pixel selects one of the "short" cameras, and 0.05"/pixel selects a "long" camera. The science field of view (shown in green) will update to reflect the configuration chosen, also displayed by the position editor.
- The disperser menu is used to select one of the three gratings. This, in combination with the pixel scale and slit width, determines the spectral resolution.
- The minimum exposure time is 0.2 sec, maximum exposure is generally dictated by object brightness, sky background or radiation events (see the Observing Strategies and Known Issues pages). Coadds can be used to combine multiple integrations, useful for high background observations or other situations in which exposures are very short.
- The focal plane unit menu is used to select the slit width or the IFU mode. The slit length is determined by a separate mechanism (the decker) which is implicitly determined based on the pixel scale and whether one of the cross-dispersing prisms is used.
The central wavelength field sets the grating central wavelength. Menu choices are available and recommended for moderate dispersion (0.15"/pix + 32 l/mm grating) configurations. For higher-resolution observations, any wavelength can be entered. The proper order-sorting filter is automatically selected; this choice can be overridden using the GNIRS iterator, although this is normally only necessary for acquisition observations.
- The cross-dispersed menu selects whether a cross-dispering prism is used for the observations and if so, which one. For short-camera cross-dispersed observations, only one XD prism is available. With the long camera, either the SXD or LXD prism can be chosen. This choice has implications for slit length, wavelength coverage and spectral resolution.
- Well Depth: the bias voltage may be adjusted to increase the well depth for thermal IR (L and M band) observations or observations of very bright targets. The default is "shallow", which is appropriate for most observations below 2.5 microns.
- Position angle: this may be set to "Fixed" or to "Average parallactic." In the latter case the value for the PA will be selected at the time of observation. Selecting the average parallactic angle minimizes slit losses, but is only important in cross dispersed mode with the narrowest slits. See the refraction table for more information. If a fixed angle is selected a value of 90 degrees is recommended; it usually minimizes drift perpendicular to slit due to flexure between GNIRS and the peripheral wavefront sensors. However, any position angle - specified in degrees E of N - may be chosen. The view of the science field in the position editor will reflect the selected angle.
The read mode is automatically set when defining the exposure time for the integration, although the user can override this choice. Each read mode has a different read noise and readout overhead: see the detector and observing strategy pages for more details. Green text is informational.
The ISS port tab allows the user to select the position on the Instrument Support Structure in which GNIRS will be used. Changing the choice of port selection alters the way the field of view is displayed in the position editor, as well as various configuration options set at the telescope. GNIRS is going to be mounted on the side port from 2011A for the foreseeable future, so "side-looking" (the default) should be chosen.
To make any changes visible in the observing database used by the observatory, the program must be stored.
Selecting cross-dispersed (XD) mode
When a cross-dispersing prism is selected, the cross-dispersed tab becomes active (shown below). This tab displays the full wavelength coverage provided by orders 3 through 8, the primary orders passed by the cross-dispersing prism. This is an informational tab only, no input is required. With pixel scale = 0.15 arcsec/pix and the 32 l/mm disperser (max. R~1700), the central wavelength should be set to "cross-dispersed" (=1.65um). The observed spectrum will then provide essentially complete coverage from 0.9 to 2.5 um.
The GNIRS Iterator is a member of a class of instrument iterators. Each works in basically the same way, except that different options are presented depending upon the instrument.
GNIRS iterators are required when the observation is stepping through different GNIRS configurations or - as is usual - flats and arcs are embedded in the observations (so that the appropriate read mode, which is not necessarily that used for the science exposures, can be selected for the calibrations). If you are iterating over GNIRS configurations and also dithering, the offset iterator would normally be nested inside the GNIRS iterator, and the observe command nested inside the offset iterator.
A good way to check your observing sequence is by clicking on the "Sequence" folder. There one can view the sequence as a contiguous list of actions or as a timeline.
The offset iterator (visible in the above figure) is common to all instruments and is used to define the sequence of dithers or sky offsets to be used. The observing strategies page gives information on some of the considerations to be taken into account when setting up offset sequences for GNIRS. The offset sequence can be visualised using the position editor.
The position editor can be used to visualise the field of view and offset pattern, select and view guide stars, change the position angle, and much more. Full details of its capabilities are given on the generic position editor page. Here we point out that selecting the OIWFS causes the position editor to overlay a red, shaded "keyhole" corresponding to the science field of view in GNIRS' acquisition imaging mode (see figure below). This can be useful in planning acquisitions of complex fields, for example. The OIWFS itself has not been commissioned.
Acquisition observations are needed for all on-sky observations (science targets and standard stars) and are defined in a separate observation from the spectroscopy observations. The OT library provides templates for acquiring several types of object:
- Very bright objects requiring use of a narrowband filter
- Objects faint enough to use a broadband filter, but bright enough that no sky subtraction is necessary
- Faint objects requiring sky subtraction
- Very faint objects requiring a blind offset from a reference star
The acquisition procedure is explained here.
The library notes give detailed instructions on how to customise the acquisition templates. In the static component, all fields except the exposure time, coadds and read mode should remain the same as in the spectroscopy observation. Although the grating is bypassed by the acquisition mirror, keeping the central wavelength and grating selection fixed avoids the grating being moved between acquisition imaging and spectroscopy (and thus between the science target and telluric standard). Near-IR acquisitions for thermal-IR observations should normally keep the deep well setting used for the spectroscopy, as the detector takes a few minutes to stabilise after changing the bias voltage.
In the GNIRS iterator contains the following steps, in which the acquisition mirror, coadds, decker (which determines the slit length), exposure time, filter and focal plane unit (slit) need to be set.
- Image of the slit. Slit image exposure times are typically just a few seconds (except in the narrowband H2 filter, which should not be used for slit imaging). The focal plane unit should be the same as that used for the spectroscopy, and the decker should be set explicitly and match the spectroscopy configuration. In the library templates the slit image is offset by 10" to avoid bright targets landing in the middle of the slit, making it difficult to measure the slit centre.
- Image(s) of the target field, with optional sky subtraction. The slit and decker should be set to "acquisition" for these steps, to allow the whole imaging field of view to be seen. The exposure time and coadds should be appropriate for the source being acquired (see this table).
- Image of the target through the slit. This step should use the decker and FPU of the first step, but the filter, exposure time and coadds of the second step.
The acquisition mirror should be "in" for all acquisition observations. The GNIRS acquisition field scale and orientation are shown here, and can also be seen using the OT position editor as described above.
The red cameras, which are used for L/M band spectroscopy are sensitive enough at short wavelengths to be useful for H/K band acquisitions (see this page). This may be preferable to acquiring in the PAH band filter for some objects. The OT will give an error if the H band filter is selected for use with a red camera. This can be ignored and will be removed once software effort has been released from other tasks.
Acquisition time is charged according to the type of target, i.e., to the program for science targets, to the partner for baseline tellurics. This is accomplished by setting the Observe Class to "Acquisition" or "Acquisition Calibration", respectively.
PIs are asked to specify all calibrations, including baseline calibrations, in their Phase II OT file. The OT library example groups include telluric standards, flats, and arcs, and users are encouraged to take these groups as a starting point for defining their own observations.
Baseline calibrations generally consist of "before" and "after" telluric standards and GCAL flats, arcs and - for XD spectroscopy - pinhole flats. Calibrations in addition to the baseline set will be charged to the program, unless they can be executed between morning and evening twilight (e.g., darks, twilight flats). Flats and arcs that are taken with the data (i.e., without changing the instrument configuration) should be embedded inside the science or telluric observation. Using the "Night Baseline GCAL" option in the OT, as shown in the OT library examples, will automatically add the appropriate nighttime calibrations with the correct exposure times etc.
In principle it is possible to define single science observations containing several central wavelengths, taking arcs and flats at each wavelength before moving the grating to the next wavelength setting. However, such a sequence would involve moving the tertiary ("science fold") mirror between its sky --> instrument and GCAL --> instrument positions. The science fold mirror positioning is not reproducible enough to guarantee that the source will stay in the slit after a move, so we do not recommend this kind of observing setup (and this is one reason why the library examples all have the calibrations at the end of the sequence). PIs wishing to embed flats/arcs at multiple wavelengths should contact their NGO/contact scientist before proceeding.
Telluric standards should be defined in much the same way as a normal science observation, including an acquisition observation and the desired offset sequence. The only difference is the "Observe Class" which should be set to "Nighttime Partner Calibration" for baseline standards, or "Nighttime Program Calibration" for special/additional calibrations.