[Top] [Prev] [Next] [Bottom]

35.5 Assessing GHRS Acquisitions

There are no jitter files available for the example we just discussed, program 5297, and so in this section we show another case drawn from GHRS ISR 079. We will return to program 5297 when we discuss the science assessments.

35.5.1 Understanding GHRS Acquisitions

Most GHRS targets were point sources (stars) and were acquired with the normal on-board target acquisition mode. Other types of targets were acquired either by blind offsetting from a nearby star, with an FOS acquisition, or offsetting from coordinates determined from a WFPC2 image. FOS acquisitions and WFPC2 offsets were discussed in the GHRS Instrument Handbook. The LSA is larger than the GHRS PSF, and using a blind offset to position a target into the LSA could introduce a wavelength shift in the data. Each OnBoard Acquisition is composed of up to eight steps, some of which are optional. The first and last steps always occur:

  1. Initialization.
  2. Deflection calibration (DEFCAL).
  3. Target Search: 3 x 3 (or 5 x 5) spiral search. Search stopped when within BRIGHT/FAINT limits or pointing returned to dwell scan position with most counts.

  4. Locate: center the target in the LSA.
  5. Peakup in SSA.
  6. Blind offset.
  7. Post-offset peakup.
  8. Final Flux measurement (ZFLUXM). A DEFCAL is an internal calibration that used the calibration lamps to locate the spectrum on the photocathode. The location of the spectrum on the photocathode depended on thermal and geomagnetic factors. This spectrum was not saved, but the location of the spectrum on the photocathode was compared to coordinates stored in the onboard data base. An offset was determined and used by the onboard computer during the subsequent Search and Locate phases of the LSA ACQ. A DEFCAL observation only generated log files (.shp, .udl), not science data.

    The target search phase slewed the telescope in a spiral search pattern. At each dwell point, the flux in the aperture was measured. All targets were first acquired using the LSA because the SSA is so small. (Note, however, that the SSA ACQ/PEAKUP algorithm is the same as that used for the LSA Return-to-Brightest (RTB) target search.) The target is assumed to be somewhere within the LSA 3 x 3 (or 5 x 5) spiral search pattern. The separation between dwell points during the LSA search was 1.52 arcsec on the sky (for an SSA ACQ/PEAKUP 5 x 5 spiral search, the separation was 0.052 arcsec). Figure 35.6 displays normal LSA 3 x 3 and SSA 5 x 5 spiral search patterns (from OMS jitter files). The spiral searches are plotted in V2,V3 space.

    Most LSA acquisitions from mid-1993 on specified the BRIGHT=RETURN (RTB) option. This feature automatically returned the spacecraft pointing to the dwell position with the highest total counts. Since a 32-bit onboard register was used to store and compare the measured counts, overflow was not a problem. Alternatively, BRIGHT and FAINT limits may have been specified for the spiral search. For this case, a 16-bit register was used to store the measured counts and wrapping of the counter could result if a diode exceeded about 65,000 counts. The spacecraft pointing stopped at the first dwell position that had a flux between the selected limits. If no dwell position was found to meet the requirements, the spacecraft pointing remained at the last position.

    The locate phase consisted of y balance, coarse-x balance, and fine-x balance. This process determined the location of the target in the aperture, and commanded the telescope to slew to position the target in the center of the aperture. The y balance routine compared the counts in the upper and lower half of the aperture and used single step offsets to find the optimum centering of the target. The x balance algorithm computed a coarse offset to place the peak of the target on one of the two center acquisition diodes. The peak was then single stepped toward the other center diode until the peak count was sensed on the second diode. Offsets in y and x were computed and used to slew the telescope. Sometimes this phase helped pull targets into the aperture that were not well-centered by the search.

    A small angle maneuver was performed to move the target to the SSA. The LSA shutter is closed when the SSA is used. An SSA ACQ/PEAKUP was used to perform the centering in the SSA. In practice, the SSA ACQ/PEAKUP is identical to the LSA RTB except the SSA was used and the Locate and balance phases are not performed. The pointing was returned to the dwell position with the most counts. For the SSA example in Figure 35.6, the target was found at the first dwell position. The offset in the return path to the first dwell position was due to a velocity aberration correction.

    The flux was summed over the eight target acquisition diodes after the completion of the locate phase (and any additional peakups); this value was written into the keyword ZFLUXM, available in the Unique Data Log (.udl or .ulh) and the target acquisition science header, if one was generated. Executing the following IRAF command returns the flux value:

cl> hsel z2il0102t.ulh $I,zfluxm yes
z2il0102t.ulh 16092

35.5.2 Return-to-Brightest

Not every GHRS target acquisition method produces data to examine. After the acquisition, if specified in the proposal, IMAGEs or MAPs will create science (.d0h) files. The Return-to-Brightest (RTB) acquisitions, implemented in 1993, record the flux at each dwell position. At first (April 27, 1993, SOGS Build 31.1) this information was dumped only to the trailer files (where the first dwell position is numbered 00) but eventually .d1h files were created to contain this data. See GHRS ISR 079 for more details.

Figure 35.7 shows examples of contour plots for two LSA ACQ .d1h/.d1d pseudo-images. The target was found at the first dwell position for the first example, while for the second example, the target was found at dwell position 6. For both examples, the spacecraft pointing was returned to the dwell position with the maximum flux.

During the spiral search, the telescope slews were parallel to the GHRS x and y detector coordinates. The first slew in the sequence is in the positive x direction. The next slew is in the negative y direction, followed by a slew in the negative x direction. All subsequent slews are performed to complete the spiral pattern. Figure 35.6 displays the spiral search geometry and the corresponding detected counts at each dwell position for the two LSA ACQs presented in Figure 35.7. Figure 35.8 displays two examples of the SSA 5 x 5 spiral search. The upper search pattern found the target (a Cepheid binary star) at dwell position 1. The lower search pattern found the target (a B0-B2III-I star) on the edge of the spiral pattern at position 19. The spectra for this program were fine, indicating the target was found and subsequently moved to the center of the aperture.

35.5.3 OMS Products and Guiding

The guide stars used for pointing control during an observation can be found in the Observatory Management System (OMS) observation log header and in the SHP (.shd/.shh) header file. The OMS data header file (.cmh or .jih) contains information about the observation. The header is divided into groups of keywords that deal with a particular topic (e.g., spacecraft data, background light, pointing control data, line of sight jitter summary). A more complete description of OMS products can be found in Appendix C of Volume I.

There were three commanded guiding modes: FINE LOCK, FINE LOCK/GYRO, GYRO. If the acquisition of the second guide star failed, the spacecraft guidance could drop from FINE LOCK to FINE LOCK/GYRO or even to GYRO which could result in a target rolling out of an aperture. It must be noted that single guide star guiding is acceptable. A GHRS target could remain in the SSA even over multi-orbit exposures. The guiding at the time of the observation can be found in the OMS header keyword GUIDEACT.

cl> hsel z2il0102j.cmh $I guideact yes
z2il0102j.cmh "FINE LOCK"
You can check the OMS header keywords to verify that the guiding mode was the one requested. Archive users should check the guiding mode to verify that no changes occurred during the observation. If you suspect that a target rolled out of the aperture during an exposure, you can check the counts in each group of the raw science data.

A "type 51" slew was used to track moving targets (planets, satellites, asteroids, comets). Observations were scheduled with FINE LOCK acquisition, i.e., with two or one guide star. Usually, a guide star pair stayed within the pickle during the entire set of observations (obset), but if two guide stars were not available, a single guide star may have been used. This assumes the drift was small or perhaps that the observer stated the roll was not important for the program. An option during scheduling was to drop from FGS control to GYRO control when the guide stars moved out of the pickle. Guide star handoffs (which are not a simple dropping of the guide stars to GYRO control) could affect the guiding and may be noticeable when the jitter ball (V3 vs. V2) is plotted.

On January 19, 1996, there was a failure in tape recorder #2. This recorder had been used for recording engineering data for subsequent playback to the ground. The failure was permanent, causing engineering data to be obtained only during real-time TDRSS availability (typically 80% or so during each orbit). Consequently, there are gaps in the engineering data and gaps in the OMS products.

Displaying Spacecraft Motion

IRAF/STSDAS tools are available to read the OMS tables and graphically display the spacecraft motion during an observation. The package of tools to work with tables is called ttools. For most observers, the .cmi or .jit STSDAS table will contain all the information needed to determine if there was motion or jitter during an observation.

For example, plotting "v2_dom" versus "v3_dom" will generate a plot that looks like a ball of string, which is why it is sometimes called the jitter ball. The same plot for a target acquisition will display the spiral search and locate phase that occurred. See fifth page of paper product example.

st> sgraph "z2o40402j.cmi v2_dom v3_dom"

After entering the plotting command, the command will plot the data in the graphics window and auto scale using the minimum and maximum data values. The plot may be misleading due to the auto scaling of the axis units, so be careful.

35.5.4 Acquisition Anomalies

Status Codes

If a target was not in the aperture, the acquisition may have failed and generated an error message also known as a Status Buffer Message, or the acquisition software may have centered on noise or a background object. All of the flight software status messages have their corresponding numbers placed into the keyword FINCODE. The FINCODE value for each data group are stored in the UDL (.uld/.ulh) file. The STSDAS task obsum can be used to display the individual group FINCODE values. The FINCODE value of the last data group is extracted and written to the science header (.d0d/.d0h). FINCODE=102 is a normal end of an observation. Any other value of the keyword indicates a possible problem, see Table 37.7.

The GHRS Carrousel

All acquisition mirrors and gratings were mounted on a rotating carrousel. There were times when the carrousel failed to lock into position if a target acquisition took longer than expected to complete, due to carrousel movement. The subsequent observations may have been affected and exposure times shortened (FINCODE=106).

Figure 35.9 shows an example of a SSA ACQ/PEAKUP that started late and completed later than the expected time as specified by the SMS (Science Mission Specification). A guide star re-acquisition took longer to complete than normal and there were two loss of locks before the start of the SSA ACQ/PEAKUP 5 x 5 spiral search. The acquisition did not complete normally. The time window used by OMS to extract the telemetry was for the originally scheduled time. The OMS .jit table contains part of an unrelated slew as well as the start of the spiral search pattern. For this unique case, a request was sent to the STScI help desk (help@stsci.edu) for a new OMS observation log. The software engineers determined that the OMS software could not handle this special case, moving the start time of the observation to extract the spiral search pattern information. OMS software was to be modified to handle these special cases. Currently, the anomalous observation log is in the archive.

Figure 35.9: SSA ACQ/PEAKUP Started Late Due to Long Guide Star -Re-acquisition

False Fine Lock

A false lock, an anomaly that is not specific to GHRS, was simply a failure of an FGS (Fine Guidance Sensor) to lock onto the null point of the S-curve for the acquired guide star. Instead, it locked onto noise, onto a fake fine error signal at the edge of the FGS interference field of view, or onto a very weak S-curve. These sources of noise may be influenced by:

Loss of Lock

During a loss of lock, spacecraft motion can be large (>0.2 arcsec). FGS control of the spacecraft is given up, and the FGSs execute a sequence of maneuvers to re-acquire the guide stars (sort of an on-the-spot guide star re-acquisition). During a loss of lock recovery, the spacecraft Take Data Flag (TDF) was set to "no." The function of the TDF is to provide a simple reference for the science instruments to query about the stability of the spacecraft pointing. During a loss of lock, the GHRS stops acquiring data, but the LSA shutter does not close. (If an observation is taken in the SSA, the LSA shutter is already closed.) Once fine lock pointing control has been re-established, the TDF was set to "yes."

A loss of lock did not necessarily result in the loss of science if the guide stars could be re-acquired and the observation completed before the next scheduled activity. However, this was unlikely for most GHRS observations and a delay usually forced an observation to time-out.

Suspect Spoiler Guide Stars

The Guide Star Catalog (GSC) was created with entries down to 15th magnitude from digitized photographic Schmidt plates. Suitable guide stars were selected for each observation by the Guide Star Selection System (GSSS) to be used for fine pointing and guiding of HST. Most (>98%) guide star acquisitions and guiding were successful. The main source of problems was very close binary stars that were unresolved on the Schmidt plate, causing FGS fine lock failures. There were infrequent failures (1 to 3 a year) due to misclassification of an object in the GSC (blend of two stars or a faint circular galaxy).

A spoiler star could result in acquiring the wrong star with the FGS and the target would not be in the aperture due to this mispointing of HST. This could also result in a shift in the pointing following a guide star re-acquisition when the scheduled star was acquired. The target could roll in and out of the aperture resulting in changes in the counts between observations obtained during different orbits.

The differences between the predicted (PREDGSEP) and the actual (ACTGSSEP) measured separation between Guide Stars can be used to test for a spoiler star. For the following example, the science observation z2zx0207 started with the target in the aperture, was halted for earth occultation, and the observation completed on the following orbit. However, the guide star re-acquisition (REACQ) locked up on a spoiler star. The target was not centered in the aperture during the remaining exposure time. Observation z2zx0208 was noise. Observation z2zx0209 was halted for earth occultation and upon the REACQ, the target was moved into the aperture. For this observation, the first part of the exposure resulted in noise, while for the second orbit, the target was in the -aperture.

cl> hedit z2zx020*.jih PREDGSEP,ACTGSSEP .
z2zx0204j.jih,PREDGSEP = 1328.046 <-- SSA ACQ/PEAKUP
z2zx0204j.jih,ACTGSSEP = 1326.466
z2zx0205j.jih,PREDGSEP = / <-- SPYBAL
z2zx0205j.jih,ACTGSSEP = /
z2zx0206j.jih,PREDGSEP = / <-- WAVECAL
z2zx0206j.jih,ACTGSSEP = /
z2zx0207j.jih,PREDGSEP = 1328.046 <-- first science obs.
z2zx0207j.jih,ACTGSSEP = 1328.019 guide star REACQ
z2zx0208j.jih,PREDGSEP = 1328.046 wrong pointing
z2zx0208j.jih,ACTGSSEP = 1328.019
<-- guide star REACQ
z2zx0209j.jih,PREDGSEP = 1328.046 <-- correct pointing
z2zx0209j.jih,ACTGSSEP = 1326.475
z2zx020aj.jih,PREDGSEP = 1328.046 <-- no further problems

Figure 35.11 presents the overplotting of the pointing for each observation (v2_dom versus seconds). This shows the change in HST pointing for this program. An HST Observation Problem Report (HOPR) was filed by the PI and the observations were repeated using different guide stars.

Figure 35.11: Jump in HST Pointing due to a Spoiler Guide Star

Single FGS Mode

On February 12, 1993, FGS 2 was temporarily turned off. Some observations shortly thereafter were executed with guiding from a single FGS. Later, a guide star acquisition could fail, resulting in use of a single guide star. For GHRS this could become a problem if the observation was being done in the SSA because the star could roll out of the aperture, as seen by a decrease in counts in the data. At some point, it was decided that single FGS mode was perfectly all right for LSA observations because the amount of expected roll wouldn't take the star out of the LSA, but it was not to be done for SSA observations. If you see a large drop-off in counts with SSA data, one explanation could be use of a single FGS.

Recentering Events

During a recentering event, pointing control of the telescope by the Fine Guidance Sensors was paused until the spacecraft motion excursions became small enough that FGS guiding could be resumed. Recenterings typically lasted a few seconds, the total spacecraft motion was less than 0.1 to 0.2 arcsec, and after the recentering was finished the pointing position should be the same as before the recentering (to within ~7 milliarcsec).

During a recentering event, the GHRS continued to acquire data. The PI (or archival researcher) is the only one who can determine whether or not a recentering event affected the target acquisition or data. Careful comparison of the time of recentering and the time during which the individual exposures were made can be obtained from the OMS Obs Log. Plotting the OMS "recenter" column versus "seconds" will indicate whether or not a recentering event occurred. The recentering event may have occurred during a time when data were being read out, in which case there is no problem. If the recentering was short compared to the exposure time of a given group of data, it probably had no effect on the data.

35.5.5 Images and Maps

Location of Targets Within the SSA

The location of a single star within the Small Science Aperture cannot be determined with an SSA image.1


To determine the orientation of the GHRS apertures on the sky, you need the value of the PA_APER keyword in the SHP header (.shh). This number is the position angle of the +y axis of GHRS measured from north through east. The +y axis is the direction from the LSA to the SSA. The +x axis is the direction of increasing wavelength.

The trick is to figure out how to display your image so that it has the same directional sense as Figure 34.1. The problem is that tasks typically display +y up and +x to the right but what is needed is +y down and +x to the right. In IRAF this would be something like:

cl> display test.hhh[*,-*].

Once you have this orientation then you apply the offset angle (PA_APER) to get northeast lined up.

You should also know that RA_APER1 and DECAPER1 are the predicted RA and Dec of the center of the aperture used, for the beginning of the observation. GHRS PEAKUPs effectively re-zero the coordinate system so one would need to use the OBS logs (jitter files) to get the actual pointing.

[Top] [Prev] [Next] [Bottom]

1 Malumuth, E., in Calibrating HST: Post Servicing Mission, 1995, pp. 227.- 241.

Copyright © 1997, Association of Universities for Research in Astronomy. All rights reserved. Last updated: 01/14/98 15:45:39