GHRS Instrument Handbook
Instrument Summary -- Why Use the GHRS?
GHRS has several operational modes for target acquisition and obtaining science data. This is meant only as a brief introduction. "Acquisitions" on page 42 should be consulted for more details on acquisitions, and other parts of Chapter 4 for ACCUMs, etc.
Most targets observed with the GHRS can be automatically acquired with an onboard acquisition (ONBOARD ACQ). An onboard target acquisition observation consists of distinct phases. Phases 1 and 2 perform initialization and internal calibration functions, and need not concern the observer.
The third phase is called Target Search. A series of small angle maneuvers, called a "spiral search," scans an area of the sky centered on the initial position. The flux coming through the Large Science Aperture (LSA) is measured at each dwell point in the search. If the BRIGHT=RETURN option has been chosen, either explicitly or by default (and it is recommended), the telescope returns to that dwell point which had the greatest number of counts. If BRIGHT and FAINT limits have instead been specified, the flux is compared to these upper and lower limits at each step in the spiral, and if the measured value falls between these limits the target is assumed to be within the aperture and the search immediately stops. You may request that a field map be generated at the final dwell point by means of the MAP optional parameter. You should be aware, though, that approximately two minutes is required for each map, and that many pointings may be made during the search. (If you intend to analyze the maps in real time, the search phase must be done as an interactive acquisition.) If you wish to confirm the spacecraft's pointing, we recommend obtaining an IMAGE after the acquisition instead of using the MAP option - see Chapter 4.
The fourth phase is target locate. Although executed as a separate phase of the ONBOARD ACQ, operationally it is identical to an ACQ/PEAKUP executed in the LSA. This process measures the precise location of the target within the aperture, and requests a small angle maneuver to move it to the center. The field map of the LSA may be made before the centering maneuver is performed by specifying MAP=END-POINT. If done after the centering (in IMAGE mode), the map can be helpful for confirming that the object was placed precisely in the center of the aperture*1. The final phase of an acquisition is a flux measurement in which the flux entering the GHRS through the final target aperture is measured and inserted into the data. After centering, a second maneuver will automatically translate the object to the SSA if that is the aperture specified for the observation. An ACQ/PEAKUP with "0.25" as the specified aperture is necessary to center the object in the SSA. For some kinds of difficult targets an onboard acquisition may not work. Possible causes might be:
- The error in the coordinates is greater than a few arcsec in either declination or right ascension, so that the target lies outside the largest area that the GHRS can search in its onboard procedure.
- The object is a moving target whose coordinates can not be predicted with xb1 5 arcsec accuracy when the proposal is written. Features in the atmosphere of a planet, and comets are possible examples.
- The object has a poorly known or unpredictably variable ultraviolet flux.
- The target has nearby neighbors of similar brightness - the onboard search process could then center on the wrong object.
- The object has a spatial extent greater than two arcsec. The automatic centering algorithms may not produce acceptable results on objects comparable in size or larger than the Large Science Aperture.
- The object is too faint to get adequate counts with the maximum permissible integration time of 12.75 seconds.
In many cases these problems can be worked around by using an ONBOARD ACQ on a nearby star and then offsetting to the object of interest, or, perhaps, by using the FOS to acquire before slewing the target to the GHRS.
You may choose to obtain an early acquisition image with WFPC2, FOC, or GHRS itself. In some cases an acquisition image would be helpful, but the field of view of the WFPC2 is not needed. Stationary point sources in crowded but recognizable fields would be examples. In earlier cycles this was done as an EARLY ACQ, but now the initial image is obtained as a first Visit, with a subsequent Visit or Visits with an ON HOLD Special Requirement; see the Phase II Proposal Instructions.
The GHRS also has its own "field map" capability which will produce an image of the sky as seen through the LSA. Each map is a square array of 16 x 16 pixels, covering arcsec with 0.11 arcsec spatial resolution. A single field map requires a minimum of two minutes to take the data and send it to the ground, and much longer if each point in a spiral search is mapped or if a STEP-TIME longer than the default (0.2 sec) is used. One WFPC2 image requires from three to five minutes, but covers a much larger area of the sky. As a practical matter, if more than one field map would be needed, a WFPC2 image may be a more efficient choice. The FOC could also be an appropriate choice in some situations.
If an interactive acquisition (INT ACQ) is required, the observer must be present at STScI, prepared to inspect the image and identify the target in a timely fashion. Real-time observations are subject to many constraints and are difficult to schedule (they are occasionally impossible). Early acquisition should therefore be chosen in preference to interactive acquisition whenever possible. Your HST Phase I proposal should include a justification of your request for real-time observation.
Early and interactive acquisitions are rarely used now. The shortness of Cycle 6 for the GHRS (see "A Short Cycle 6" on page 11) is yet another reason to eschew them.
There are several modes of science data acquisition, including Accumulation Mode, Rapid Readout Mode, and Image Mode. Each of these modes may be used in conjunction with any of the optical configurations described earlier.
The first is the ability to make long duration observations with effective and automatic control of the process. The time varying Doppler shift caused by the orbital velocity of the spacecraft is compensated for automatically. The software constantly monitors a set of data quality criteria and can flag, reject, or reobserve individual integrations that fail the tests. Finally, the software can suspend the observation during scheduled or unexpected interruptions, such as occultation of the target by the Earth or passage through the South Atlantic Anomaly, and then resume when the interruption ends. The very low background count rate and absence of readout noise in the Digicons make exposures of hours duration feasible, though it is strongly suggested that these be broken into shorter segments to aid in scheduling and protect against catastrophic data loss in the event of an unexpected problem.
The second category of benefits results from the ability of the software to perform patterns of integrations at closely spaced positions on the photocathode, a process which is referred to as substepping. There are four purposes for this. At the beginning of an observation, the software executes a procedure called Spectrum Y Balance (SPYBAL) to find the optimum centering (up and down, perpendicular to the direction of dispersion) of the image on the diode array. This compensates for minor changes in the image location due to thermal or electrical drifts. The second use is to make multiple (2 or 4) samples per resolution element (1 diode width) to ensure that the digital data satisfy the Nyquist sampling criterion. This is very important when the ultimate spectral resolution of narrow features is required. Third, the background adjacent to the spectrum or in the echelle interorder region can be measured. Finally, COMB addition allows the effect of small diode-to-diode sensitivity variations to be minimized and eliminates the holes in the data due to a few inoperative channels. When substepping is used to define the detailed sampling of the spectrum and background, the data obtained at each step are accumulated into one of up to seven distinct "bins" in the memory of the onboard computer.
This overview of the flight software features is not exhaustive, but summarizes those capabilities which are immediately relevant to the acquisition of spectra in ACCUM mode. Several items, namely substepping and exposure control, require the observer to specify certain parameters. These will be described in more detail later in this Handbook.
Rapid Readout Mode (RAPID), sometimes called "direct downlink", is intended to provide excellent time resolution without the overhead times associated with ACCUM mode. The sample time can be between 50 ms and 12.75 seconds, in increments of 50 milliseconds (i.e., 1 to 255 times 50 ms). At the end of each integration the data are read out, either directly through the TDRSS satellite or to the spacecraft science data tape recorder. The flight software cannot execute all of its functions and still allow readouts every 50 ms. When RAPID mode is entered, substepping, data quality checks and exposure control features are deactivated in order to reduce instrument overhead activities to the bare minimum.
The primary factor governing the choice between ACCUM and RAPID is time resolution. In ACCUM mode, the time between exposures can be no shorter than about three seconds for the simplest STEP-PATT (see "Standard Patterns for Substepping and Background Measurement" on page 112) and is about one minute for more typical cases. If higher time resolution is required, if the source is bright enough to give useful counts in a shorter integration, and if one is willing to sacrifice the flight software control, then RAPID mode is a useful alternative. In RAPID mode, a SAMPLE-TIME of less than 0.33 sec requires the use of the 1 Mb data channel (see "Rapid Readout Mode" on page 54). Such a high data rate stresses HST's data-handling capabilities and means that only about 20 minutes of observations can be stored. A SAMPLE-TIME of 0.33 sec or more allows for essentially continuous operations.
Images may be obtained in this mode by deflecting the image of the photocathode over the 0.11 x 0.11 arcsec focus diodes. The result is a map similar to that obtained during target acquisition, but without an acquisition being performed. Note that a MAP as part of an acquisition can cover more of the sky than the LSA subtends at one time by small movements of the telescope, whereas an IMAGE is limited to the 1.74 x 1.74 arcsec area of the LSA; see "Image Mode" on page 49 and "MAPs" on page 101.
These are really modifications of ACCUM mode designed for higher efficiency in multiple observations, and they may be requested during Phase II of the proposal process. WSCAN obtains a series of spectra within a given order, incrementing by a specified wavelength increment between each. The result is a spectrum spanning a broader wavelength range than is possible with a single exposure. OSCAN works with the echelle, and uses the magnetic deflection of the Digicon to obtain spectra over a range of echelle orders. The grating carrousel is not rotated, and spectra are obtained at equal values of ml, where m is the echelle order. OSCAN is not ordinarily used for science observations. Both WSCAN and OSCAN are just shorthand versions of ACCUM that make it more convenient to write a Phase II proposal but which do not change the way in which the instrument operates.