--> OT Details | Gemini Observatory

Change page style: 

OT Details

You are here

This page explains how to configure T-ReCS in the Observing Tool:


Refer to the rest of the T-ReCS pages for further details of the instrument capabilities and guidelines on how best to use the instrument.

There is a T-ReCS OT library file containing examples of typical T-ReCS observations that is available for fetching from the Observing Tool database. For more examples of T-ReCS observations in the OT refer to the T-ReCS OT Example page, which shows normal sequences for imaging and spectroscopic observations.

See also the MIR proposal and phase II checklist.



T-ReCS Component

The detailed component editor for T-ReCS is accessed in the usual manner, by selecting the T-ReCS component in your science program, and is shown below:

T-ReCS component

Selecting a filter

The filter is chosen by clicking on the pull down list (i.e. the down-pointing arrow) and selecting the desired filter. The filter names match those listed on the T-ReCS filter page.

Setting the observation time

The main fields to be set here (or in the sequence, as appropriate) are: the total on-source time, the chop angle, the chop direction, the array orientation on the sky and, for spectroscopy, the combination of the disperser and focal plane mask to be used. These are discussed in order in the next few paragraphs.

The total on-source time is what it says -- the time spent usefully integrating on the target field while carrying out any chopping or nodding of the telescope. This corresponds to the usual notion of the exposure time, although for T-ReCS the on-source integration will be made up of some large number of very short integrations to avoid saturation of the detector (see this page about chopping and nodding). The on-source time is the only "exposure time" value the PI needs to be concerned about.

The total on-source time has to be set in the T-ReCS component of the OT, and it may be defined for the individual steps of a sequence in the T-ReCS Iterator component, where it is denoted as the "timeOnSource". All times are in seconds.

The individual frame exposure time (i.e. the duration of each readout of the detector) is not set explicitly because the value for optimum sensitivity of mid-IR detectors depends on the incident flux through the filters. Hence the T-ReCS software adopts an individual exposure time based on the filter, disperser and other parameters.

In normal operation we expect that the "Observe" component will be set to "1X" for most T-ReCS observations. Even if a long observation is split into several sections, it is better if one has duplicate observation steps in a sequence rather than setting the "Observe" component to a value larger than "1X", unless there are no acquisition steps in the observation. Otherwise all the acquisition steps will also be multiple, which is not desired.

For long on-source integration times we recommend that the observation be split into several shorter observations of no more than 20 minutes on-source time. A unit value of 10 to 15 minutes on-source is better, since a 15 minute on-source observation takes of order 50 minutes real time in imaging mode and of order 60 minutes in spectroscopic mode. There is only a small overhead for writing out the files after an observation and going on to another step. By breaking down long observations this way there is less risk of the loss of a large amount of data if there is a telescope problem or other failure, or if the weather conditions change. Thus rather than requesting a single 1800 second on-source imaging observation we would strongly suggest that the observation be entered as 6 times 300 seconds using the T-ReCS Iterator to define the 6-step science exposure sequence after any acquisition steps that are needed.

It is also important to remember that a long observation is not likely to be carried out as one monolithic observation, especially for programs in bands 2 and 3, due to practical requirements of queue scheduling. A band 1 program might request that a long observation be done on a single night, but even then it cannot be required unless this is justified in the phase I proposal (such as for time-critical observations) and approved by the ITAC. Thus if one is requesting a 4-hour spectroscopic observation, for example, one has to be realistic about the chances of getting this done in one block even if the weather cooperates. Depending on the program priority and the object visibility during T-ReCS queue nights the observation might have to be done over 3 or 4 nights, with the associated overheads for more than one acquisition.

The observation sequence timeline displays the estimated total time required for each Observe command. Note that previously the timeline was inaccurate because it took the time required for the first step in a sequence and then multiplied it by the number of steps in the sequence. From 2006A onwards this should be corrected in the OT and the timeline should give a more accurate account of the sequence.

Our experience in suggests that the on-source efficiency is about 30% in imaging mode and about 24% in spectroscopy mode. 

Nod time

The PI also can define the duration of a nod, called the "NOD dwell" time. The default time at each nod position is 45 seconds. The nod dwell value can be changed by the PI in the T-ReCS Iterator component. Note that shorter nod dwell times will decrease the on-source efficiency because in doing a nod the telescope has to be re-pointed and time has to be allowed for the system to settle after the nod has been completed. The time that the system waits after each nod is the "NOD Settle" time. The current best value for this parameter is 8 seconds. This also can be set in the T-ReCS Iterator component.

The number of nod pairs is calculated by the software when the observation is taken. The actual time on source for an observation will vary somewhat from the requested value (it is always larger!) depending upon internal detector settings of the mode that is being used as well as the nod dwell time. T-ReCS currently always uses an ABABAB... nod pattern.

Saveset time

Another parameter that PIs may wish to set is the "saveset time". T-ReCS is able to save the images that are being accumulated at intervals smaller than the time spent at a given nod position. In this it differs from Michelle, where there is one image for each chop position saved at each nod. For example one might spend 30 seconds at a nod position but save the images every 10 seconds, so that for each nod position there are 3 accumulated on-source and off-source images in the raw data file. This is useful if there are fast changes in the weather conditions, since the PI can then discard some images that have been compromised by bad conditions and retain "good" images. The saveset time can be set in the T-ReCS Iterator. Note that one should not have too short a saveset time with a long observation as this will cause data handling problems due to accumulating an excessive number of frames before the image is written out. Nor are short savesets useful for observing very faint targets.

Selecting a focal plane mask

Choice of one of the built-in focal plane masks is made by clicking on the pull down list. For imaging observations the "Imaging" mask -- meaning no mask at all -- is used. For spectroscopy observations the desired slit width from 0.21 to 1.3 arc-seconds can be selected. The "Imaging w/o Flexure Mask" option is not currently used. Note that any slit can be used with any grating.

If you display a view of the field with the position editor and have selected the "science area" button, then this will reflect the choice of focal plane mask (displayed as a blue box). The chop beams are also displayed (see more details about the T-ReCS position editor display, below).

Setting the position angle

The facility Cassegrain Rotator can rotate the instrument to any desired angle. The angle (in conventional astronomical notation of degrees east of north) is set by typing in the "position angle" window. The view of the science field in the position editor will reflect the selected angle. Alternatively the angle may be set or adjusted in the position editor itself by interactively rotating the science field. The orientation cannot be changed during a sequence, so this value is set in the main T-ReCS component and that value applies to the entire observation. The PA is that of the short axis of the array. The default PA of zero degrees corresponds to the long axis of the detector being oriented east-west. Chopping at 0 degrees then corresponds to chopping along the short ("y") axis of the array.

Selecting a disperser and the central wavelength

Choice of grating (for spectroscopy) or mirror (for imaging or spectroscopic acquisition) is made by clicking on the pull down lists (i.e. the down-pointing arrows) and selecting the desired item. One can choose spectroscopy in the N or Q atmospheric windows by the Low Res 10um Grating or High Res 10um Grating or Low Res 20 um Grating menu items. Note that if these are chosen the Filter should be set to be N (broad 10um) or Q (broad 20.8um) respectively. While one is able to take a spectrum with any of the other filters besides N or Q, this would be an unusual request and should be discussed with your contact scientist.

In either the low resolution N-band and low resolution Q-band spectroscopy observing modes the Grating Central Wavelength field is ignored. For the high resolution N-band spectroscopy observing mode the central wavelength must be defined. If more than one wavelength setting is needed, this can be arranged in the T-ReCS Iterator. One does not have to re-center or re-acquire when changing from low-resolution N-band to high-resolution N-band spectroscopy modes, or when changing the central wavelength in high-resolution N-band spectroscopy mode. It is a bit more efficient to do a multi-wavelength high resolution N-band spectral observation starting from long wavelengths and going to shorter wavelengths, due to details about exactly how the grating turret moves inside T-ReCS.

Setting chop angle, chop throw, and nod direction

It is important to note that guiding while chopping is only possible on one side of the chop. Thus, when observing a point source in conditions of less than ideal seeing and windshake, the guided image will be noticeably sharper than the other image.

The chop angle may be set independently of the instrument position angle (thus, for example, you can set the spectrograph slit at a specific angle and chop either along the slit or off the slit). The chop angle is defined in conventional astronomical notation of degrees east of north. The chop beams are displayed in the position editor (see more details about the T-ReCS position editor display, below). For bright targets a chop angle of 45 degrees with respect to the array is recommended, as it is for Michelle, since bright targets produce some low-level artifacts along the rows and columns.

Currently the maximum allowable chop throw is 15 arcsec. The default nod direction is beam-switching (nod parallel to and same amplitude as chop).

Saving changes

The save button accepts the latest changes and stores the program to the local database, the undo/redo button (and, transiently, the edit pencil) toggles pending and saved changes and the close button closes the science program editor (saving any changes to the local database).



T-ReCS Iterator

The T-ReCS Iterator is a member of a class of instrument iterators. Each works exactly the same way, except that different options are presented depending upon the instrument. The iteration sequence is set up by building an Iteration Table. The table columns are items over which to iterate.

The items that are available for inclusion in the iterator table are shown in the box in the upper right-hand corner of the T-ReCS Sequence Component. Selecting one of these items moves it into the table in its own column. Each cell of the table is selectable. The selected cell is highlighted green. When a cell is selected, the available options for its value are displayed in the box in the upper left-hand corner. For example, when a cell in the filter column is selected, the available filters are entered into the text box. When a cell in the exposure time column is picked, the upper left-hand corner displays a text box so that the number of seconds can be entered. 

Rows or columns may be added and removed at will. Rows (iteration steps) may be rearranged using the arrow buttons.



Visualisation in the Position Editor

The Position Editor displays not only the primary science beam (Nod A Chop A, equivalent to Nod B Chop B) but also the two chop beams (Nod A Chop B and Nod B Chop A). Display of the chop beams is toggled using the Position Editor View...Display Chop Beams menu. The main beam is shown with solid light blue line, the chop beams in dashed light blue. (Click on the image for a larger view).


As an alternative to setting the chop angle and throw using the T-ReCS component described above, these parameters can be set interactively using the Position Editor in drag mode, by clicking on either of the central blue square "handles". (Note the caution given above about not changing the chop throw). As usual, you can also set the instrument position angle interactively by dragging the blue triangular handle located next to the main beam. (The chop beams rotate too, of course, although their alignment relative to the main beams is unaffected).

Using the Position Editor View...PWFS FOV Display...Display at menu you can also display the vignetting pattern of the peripheral wavefront sensors as they appear at the other chop positions. (The WFSs move too slowly to follow the chop and so remains fixed at the same x,y position in the focal plane, corresponding to a different apparent position on the sky when the secondary mirror is tilted).

List of Iterator Items

The following items are available for use in creating a T-ReCS observing sequence:


  • chopAngle: the angle of the chopping, degrees E of N
  • chopThrow: the size of the chop in arc-seconds, up to 15 arc-seconds
  • dataMode: setting for whether the data should be saved to a file ("Save", the usual situation) or not saved; for initial acquisition steps this should be set to "Discard All"
  • disperser: the grating to use, as in the T-ReCS component
  • disperserLambda: spectroscopic central wavelength (highres10 mode)
  • filter: which filter to use for the step
  • mask: which slit to use, if any
  • nodDwell: the time spent at each nod position A or B in seconds
  • nodSettle: the time to wait after a nod for the telescope to recover from being moved; set to 8 seconds
  • obsMode: mode of taking data--stare, chop, nod, or chop-nod; except possibly for acquisition steps it should be left as "chop-nod"
  • timeOnSource: the total on-source exposure time
  • timePerSaveset: the time between saving images during the observation, in seconds; while this can be set by the PI, it is not useful to have more than a small number of savesets per nod position and for most purposes the default value is sufficient.
  • windowWheel: which of the T-ReCS windows is to be used; this will normally be set to "auto" and not changed by the PI

Gemini Observatory Participants