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

11.3 POSITION Mode Pipeline

The pipeline usually calibrates POSITION mode data in two distinct stages. The first stage processes each single observation in a stand-alone fashion, ignoring the other observations belonging to the same HST visit. The second stage relates all the individual observations to one another so that an astrometric "plate" can be produced. Here we describe the sequence of events recorded during a POSITION mode observation and then the two stages of POSITION mode data calibration. See page 11-8 for a sample of POSITION mode pipeline output.

11.3.1 Position Mode Observations

In a typical POSITION mode observing program the astrometry FGS sequentially observes in Fine Lock several stars distributed about the pickle. Any temporal variability in the telescope pointing will contaminate the measured relative positions of these targets. Thus, the measured positions of all targets must be mapped onto a common, fixed coordinate system before an astrometric plate can be assembled.

Experience has shown that FGS astrometry is sensitive to HST body jitter and FGS drift. The jitter can be eliminated using the guide star data, whereas the drift is removed by applying a drift model derived from check star data. A check star is a target observed multiple times during the visit. Typically the observing strategy should involve at least two check stars, and they should be observed at least three times each.

The dataset for a given observation includes the Slew to the target, the Search, the CoarseTrack, the WalkDown, and the FineLock tracking, as well as guide star data over the same interval of time. These data provide information used by the pipeline algorithms to determine backgrounds, to locate the interferometric null, and ultimately, to pinpoint the position of the star relative to the other stars observed in the visit. These data can also be used for photometric studies.

The Slew portion of the observation is used to measure the background. During the slew the astrometer's photomultiplier tubes count the photons in the 5" IFOV, registering the background and serendipitous stars. These star spikes are removed in the pipeline's background averaging process, via a trimmed mean.

Upon completion of the slew, the FGS microprocessor assumes control and begins the acquisition of the target as described in "Target Acquisition and Tracking" on page 9-16. After the completion of CoarseTrack, the DataValid and TransferHold flags are set to 1. Then the FineLock acquistion begins and the DataValid flag returns to zero until tracking in FineLock begins and the DataValid flag is set once again (see Table 10.2). The FGS will continue to track its target in FineLock until the DF 224 computer terminates the activity and slews the IFOV to the next target in the sequence. This process repeats for each exposure until the end of the visit.

When the target is faint or the data are noisy, the onboard DIFF, SUM algorithm suffers from poor photon statistics and therefore might cause the FGS to mis-identify true interferometric null. The pipeline corrects for this problem using PMT data gathered during the slew, WalkDown, and FineLock tracking as described in step 5 below.

11.3.2 Processing Individual Observations

POSITION mode pipeline processing for each individual observation in the visit executes the following steps:

Inspection of the flags/status bits to locate the data fields recording:
Computing the centroid of the IFOV, taken to be the median of the instantaneous x,y positions during the FineLock/DataValid interval, in the astrometer as well as the guide star positions in the guiding FGSs. Standard deviations about these centroids are also computed.
Updating the HST state vector, specified in the header files for the beginning of the observation, so that it is accurate for the temporal midpoint of the FineLock/DataValid interval.
Gathering photon statistics on:
Applying the DIFF and SUM corrections to both axes of only the astrometry data to locate the true interferometric null. This algorithm determines the slope of the fine error signal near interferometric null as a function of position in the pickle, using a library of reference S-curves, the target magnitude, making use of the background data computed above, and the difference in the photomultiplier averages computed during the WalkDown and the FineLock/DataValid intervals (see Figure 9.12). This correction tends to be small for bright stars (V < 13.5) but can be as large as 5 mas for faint (V > 15) stars.
Converting the raw telemetry encoder positions to instantaneous x,y detector coordinates using several parameters, such as the star selector lever arm lengths, and offset angles. The lever arm and offset angle are known to vary in time. They are monitored by an ongoing program called the Long Term Stability Monitor (LTSTAB) which executes multiple times a year. The values applied in the pipeline are determined by interpolation of the LTSTAB results.
Correcting the x,y centroids in the astrometer for Optical Field Angle Distortions (OFAD).
Correcting distortions in the astrometer arising from the pickoff mirror and aspheric mirror.
Removing differential velocity aberration from the x,y centroids using the updated HST state vector, a JPL Earth ephemeris, HST's V1 RA and DEC , the V3 roll, and the V2,V3 position of the alignment point. This correction is applied to both the astrometer FGS and the guide star FGSs. The pipeline produces output files that log these corrections, the associated standard deviations about the centroids, and the photometry averages from the four PMTs.
At this point no further processing on the individual observations are possible. The next step is to combine the measurements of the individual targets to correct for POSITION-mode jitter and FGS drift.

11.3.3 Assembling the Visit

The goal of this segment of the pipeline is to map all of the positional measurements of the individual targets onto a fixed but arbitrary coordinate system. It involves POSITION mode de-jittering and application of the DRIFT correction.

POSITION Mode De-jittering

The pipeline accounts for spacecraft jitter during the visit by establishing a fixed but arbitrary reference frame determined by the x,y centroids of the guide stars within the guiding FGSs. The HST pointing control system uses the position of the dominant guide star to fix HST's translational position and that of the roll guide star to fix HST's orientation. The output products of the pipeline processing of the individual observations include the x,y centroids of the guide star positions evaluated over the same time interval as the astrometer centroids. During the course of the visit any change in the x,y centroids of the dominant guide star within its FGS is interpreted to be HST translational jitter and is removed from the both the astrometer and the guide star maintaining HST roll. Next, any motion of the roll guide star with respect to the dominant guide star perpendicular to the line between them is interpreted as uncompensated roll of HST about the dominant guide star. The pipline then removes this roll from the astrometry data. Typically the size of the de-jittering correction is less than a millisecond of arc when averaged over the visit but can be as large as 3-5 mas for any given observation, such as when HST transits from night into day.

De-jittering is not performed at a 40 Hz rate because that would introduce noise into the dataset. Instead the time-averaged centroids of the guide stars are computed for the same time interval that the astrometer was in FineLock/DataValid. The positions of the guide stars in the first exposure, corrected for differential velocity aberration, define the reference frame for the remainder of the visit. So, for example, if the dominant guide star x,y centroids measured during the Nth astrometry observation differed from those in the first observation by (dx,dy) = (1 mas,1 mas), then the appropriate conversion to dV2,dV3 is applied to the roll star and the astrometer's local x,y centroids. This procedure creates a fixed but arbitrary coordinate system for the entire visit.

POSITION Mode Drift Correction

After the FGS data have been de-jittered, there will remain an apparent motion of those astrometry targets which have been observed more than once within the observing sequence. These check stars provide the data required for the next and potentially large correction. The drift correction model assumes that the astrometer is a rigid body which both translates and rotates in the HST focal plane during the course of the visit and corrects the measured positions of the stars in the visit for contamination by this motion.

The time-tagged positions of the check stars are used to generate a model for this drift, and the time-tagged positions of all the stars in the visit are adjusted by application of the model. Three separate models can be applied:

Linear: Translation only, no rotation.
Quadratic: Translation only, no rotation.
Quadratic and roll: Translation and rotation. The choice of model depends upon the number of check stars available and the number of times each is observed. Clearly if there is only one check star in the visit the rotation model cannot be applied. Also, if check stars are not observed frequently enough (three times or more), the quadratic models might not be reliable. The pipeline applies all three models, providing three sets of corrected centroids. It is the responsibility of the GO to decide which set is best. The output of the fitting program includes fit residuals and ².
The size of the drift correction is typically 6 to 12 mas under two-FGS guidance. The amount of drift appears to be related to the intensity of the bright Earth projected down the V1 boresight during target occultations. This intensity, and hence check star drift (generally), is highest for targets in HST's orbital plane and lowest for those at high inclination.
When only one FGS is used for guiding, the telescope is not roll-constrained. Under such circumstances the check stars can reveal very large motions, up to 60 or 70 mas over the course of the orbit. Nevertheless, this drift can be successfully removed from the astrometry data, provided the proposal contained an adequate check star scenario. For example, the overlay of the plates from two separate POSITION mode visits, each measuring some 20 stars in an astrometric field distributed throughout the pickle yielded an rms residual of about 1 mas, even though one of the visits had one-FGS guiding and check-star drifts on the order of 30 mas.
The Data Analysis chapter (Chapter 13) discusses the techniques of plate overlays and determinations of parallax and proper motion.

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

stevens@stsci.edu