Observatory Support
Pointing

Chapter 2: Header Keyword Definitions

Header Keywords

Below is a listing of keywords which can be found in the image headers. The .jih file is detailed here. The old .cmh header contains a subset of these keywords. In some cases, the same keyword's name changes between the .cmh and .jih files. In these cases, both names are listed before the description. The _jif.fits file contains nearly all of the same keywords as the .jih ASCII header file, but with the exception of some keywords affected by the new style FITS associations.

Note: The calculation of some keywords below requires the position of the spacecraft or the velocity vector, parameters which are not available in the FN telemetry format (see TLMFORM). For exposures taken in FN format, these keywords will be unpopulated. If telemetry switches occur during an observation, such keywords will be blank when switching from AN or HN to FN and may or may not be blank when the switch is from FN to AN/HN. In this latter case, if the AN/HN portion is long enough to encompass an update of the telemetered HST position data, then a value will be provided; otherwise the keyword is blank.

PROPOSID (jih) / PROPOSAL (cmh)

TARGNAME

STARTIME and ENDTIME

Note: The HRS has the unique capability of pausing (idling) during an observation for an occultation or South Atlantic Anomaly (SAA) passage. These idle periods are pre-planned. When idling occurs during an HRS observation, the duration as indicated by the STARTIME and ENDTIME will encompass the entire observation including the occultation or SAA passage. However, the tables will only contain data during non-idle periods.

SOGSID

CONFIG (jih) /INSTRUME (cmh)

PARALLEL (jih) / PRIMARY (cmh)

One important characteristic to note when analyzing pointing data associated with PARALLEL-SECONDARY is the effect of the differential velocity aberration. The telescope (and so the observation log software) computes a differential velocity aberration correction for the PRIMARY aperture only. Therefore, the secondary will experience a small drift (cyclic as the telescope moves in its orbit). The effect of the error is a deviation in the positioning of the secondary target of up to 50 mas peak to peak for a full orbit (~ 96 minutes).

Note that when WFPC2 (or WFPC) is the parallel instrument, the aperture used in the calculations of jitter and pointing in the WFPC2 observation log is always the UWFALLFIX aperture.

OPERATE

• FOC operate--The FOC is most affected by the amount of time since high voltage turn-on. In order to achieve an optical and thermal stability, the FOC high voltage must remain on for 75 minutes prior to the start of an observation, however the longer the stabilization period, the better. In this mode, all electronics units and high voltages of one chain are turned on. The FOC can remain in this state for several hours prior to activation. If 12 hours of idle time occurs, the FOC must be returned to a hold (or high voltage-off state) for the sake of the operational lifetime of the instrument.

• WFPC2 operate--The WFPC2 low voltage is always on and so the relevance of using this field to study thermal stabilization is limited. Only after a decontamination process where heaters are cycled and after an instrument safing event does this parameter become an important indicator of the stability of the instrument. After the completion of a decontamination or safing event, the WFPC2 will require several hours before thermal and optical stabilization is achieved.

• FOS operate--When in operate mode, the FOS will have the logic supply, quiet supply, and high voltage supplies turned on for the side of the instrument that is active (Red or Blue). Only one side of the instrument may be active at a time. The high voltage will be brought up to its nominal value for that side at least 30 minutes prior to the beginning of an observation.

• HRS operate--In the HRS operate mode, the low voltage, high voltage and Detector Electronics Box (DEB) for one of the two detector sides are on. The low voltage is always on for the Side 2 detector. Because of this, an extended warm-up time is not needed for either side. The DEB generates a great deal of heat and therefore, the HRS is really never thermally stable. As a result, a periodic, frequent, and automatic re-calibration of the detector coordinate location of the apertures is performed (DFCAL).

• HSP operate--The HSP operate mode refers to a high voltage state but is slightly different for the 4 image dissector tubes and photomultiplier tube (PMT). For the POL and VIS detectors, this state refers to the lowest high voltage state prior to the ramp-up to the highest or operating high voltage state. A 30-minute delay is built in prior to the ramp-up. For the UV detectors and the PMT, the ramp-up has already occurred and a 30-minute delay is built in prior to the start of an observation.

• WFPC operate--Not relevant in the context of stability after a decontamination event. This parameter is for engineering purposes only.

TLMFORM

Given below are the rate possibilities of TLMFORM, the quantization of the time tags in the observation logs, and brief comments on the primary use of the format.

TLMFORM	Rate (sec)	Display Rate (sec)	Comment
AN	0.5	0.5	Common format
FN	0.05	0.125	Older format, mostly replaced with HN
PN	0.15	0.125	Pointing and gyro observations
HN	0.05	0.125	Most commonly used format

Note: The rate column indicates how often the engineering telemetry is updated for a given format. The display rate column indicates the time resolution of the Observation Log files, the time separating lines of data in the file.

CONTENT:

• FN: - Contains a majority of the FGS telemetry points such as the PMT counts, encoder positions and flags, sampled 40 times per second. Does not contain the S/C velocity or position vectors. Therefore some column entries using velocity and position will not be defined.

• PN: - Contains a subset of FGS data, sampled at 0.15 seconds, and all associated GYRO information; does contain the S/C velocity and position vectors.

• AN: - Contains all temperature telemetry points and some FGS data and is most useful when studying the thermal trends of the SIs and vehicle. Does contain the S/C velocity and position vectors.

• HN: - Combines PN and FN format information.

APERTURE

APER_V2

Note: Aperture positions may be refined and updated or sometimes re-defined; each Observation Log's calculations are done using what was the operational value for the aperture in question, at that time.

APER_V3

ALTITUDE

The eccentricity of HST's orbit oscillates between ~0.0002 and 0.0012 over a period of about 56 days. Taking into account the oscillation of the orbital eccentricity and assuming a spheroidal Earth, the altitude range is 580 to 630 kilometers, The algorithm for the Earth radius assumes a spheroid (elliptical cross-section):*1

Where:

• equatorial_rad = 6378.14 m

• flat_factor = 0.00335364

LOS_SUN

HST is generally restricted from pointing within 50 degrees of the sun. The tolerance for off-nominal roll angles on a particular date is dependent on the value of the angle between the sun and the line of sight. In general, the HST aft shroud is hottest when this angle is close to 135 degrees (maximum surface area exposured) and coolest when the angle is 180 degrees.

LOS_MOON

SHADOENT

SHADOEXT

LOS_SCV

LOS_LIMB

The inputs to the limb angle calculation are the altitude of the HST (and hence the GCI coordinates of the HST position in its orbit), a telemetered parameter which is the cosine of the angle between the V1 axis and the Earth's center, and an assumption of a spheroidal Earth. This CMI table contains the limb angle as a function of time.

ZODMOD

EARTHMOD

MOONMOD

GUIDECMD

• FINE LOCK = two FGSs in fine lock,

• COARSE TRACK = two FGSs in coarse track,

• FINE LOCK/GYRO = one FGS in fine lock (pitch/yaw) and gyro (roll),

• COARSE/GYRO = one FGS in coarse track (pitch/yaw) and
  gyro (roll),

• GYRO = GYRO control only.

GUIDEACT

Because of acquisition anomalies, the guiding mode initiated may not be used throughout the observation. The default guiding modes used in cases where fine-lock was not achieved has varied since HST launch. Until 1993, FGSs that did not achieve a commanded fine-lock could default to coarse track. Early in 1993, as a conservation measure, FGS2 was restricted from coarse tracking; a few months later, on day 104, FGSs 1 and 2 were similarly restricted. Currently, if one or both FGSs failed to achieve fine lock, then the pointing would fall back to either FINE LOCK/GYRO or pure GYRO.

GSD_ID

If an anomaly occurs during the guide star acquisition process, the guide star keywords will be populated, however GUIDECMD and GUIDEACT will be different, the guide star separation keywords (PREDGSEP, ACTGSSEP) may differ, and an anomaly keyword will appear at the end of the header record warning of a guiding failure or degradation.

GSD_RA

GSD_DEC

GSD_MAG

GSR_ID

GSR_RA

GSR_DEC

GSR_MAG

GSACQ

PREDGSEP

Note: If the guide star acquisition failed, PREDGSEP will still be populated with the predicted separation, however the actual separation (ACTGSSEP) will be blank. If GYROS are used (GUIDECMD=GYRO) or single guide star guiding (GUIDEACT=FINE LOCK/GYRO, COARSE/GYRO), PREDGSEP will be blank.

ACTGSSEP

Note: If the commanded guiding (GUIDECMD) is set to a mode with a single guide star only (FINE LOCK/GYRO, COARSE TRACK/GYRO, GYRO), this field will be blank. If a two-FGS guide star acquisition failed into GYRO mode, this field will also be blank and an anomaly keyword will appear at the end of the header record.

Note on the actual minus predicted guidestar separation: The PREDGSEP and ACTGSSEP keyword values will differ. The difference is predominantly due to the positional errors in the Guide Star Catalog which will typically result in differences between these keyword values of 0.1-0.3 arcsec. FGS to FGS calibrations affect this as well. Depending on when the last FGS calibrations were done, there can be another 0.1 to 0.2 arcsec total contribution. For reference, the nominal operational FGS search radii are between 15 and 30 arcsec. If either guidestar is not found within this radius the acquisition will fail. Even if both FGSs acquire guidestars, the predicted minus measured separation must pass the coarse mode angle check which is set at 6 arcsec. If the difference is greater than 6 arcsec the acquisition should fail. It is possible given an unfortunate confluence of older FGS calibrations and worst case GSC error that the proper guidestars could have been locked up while exhibiting a predicted-measured value of nearly 1 arcsecond. This is rare however and in recent times this value is usually less than 0.2 arcsec. For differences between about 1 arcsec and 6 arcsec, the observer should suspect a "spoiler" field star, or a binary guidestar. This will grossly affect the target positioning.

One will also sometimes encounter the problem of an FGS "false lock". This peculiarity of the FGS interferometric tracking results in accidental guiding on a local minimum in the S-curve, and can be identified by an offset pointing and abnormally high jitter. When multiple observations are made with the same guidestars, the ACTGSSEP keyword will be very constant. An observation affected by false lock will exhibit a 0.3 to 1 arcsecond offset with respect to the normal observations' ACTSGSSEP keyword, as well as an unusually high GSSEPRMS

GSSEPRMS

Note: When comparing this value from a set of observations with the same guidestars, "false locks" discussed above can be identified. Look for unusually high GSSEPRMS keywords compared to the others in the set. As discussed above, the ACTGSSEP will also be an outlier if false lock is the culprit. The value of GSSEPRMS can also vary when there are large slews during the observation or during moving target exposures due to residual optical distortion errors in the FGSs. This keyword can be blank for the following reasons: (1) For short observations, GSSEPRMS is not meaningful in this context and the keyword is blank, (2) If the commanded guiding is set to a mode with a single guide star only (FINE LOCK/GYRO, COARSE TRACK/GYRO, GYRO), this field will be blank, (3) If a two-FGS guide star acquisition failed, this field will also be blank and an anomaly keyword will appear at the end of the header record.

NLOSSES

LOCKLOSS

The HRS and FOC are interrupted during their data-taking when a loss of lock event occurs. If enough time remains in the exposure, the data accumulation will resume when lock is re-established. The WFPC and WFPC2 close their shutters during loss of lock provided that the exposure time exceeds 3 minutes (otherwise the instrument continues to expose even though the star may be drifting out of the field). The FOS, on the other hand, continues to observe. For all SIs but the FOS, the actual exposure time is the commanded time minus the LOCKLOSS. Note that the LOCKLOSS is accumulated over the duration of the observation window.

Loss of lock is the result of the limited dynamic range of the FGS interferometry mode and is normally detected on-board when the total flux received is not as expected or the star is suddenly found too far from the nominal position (an adjustable parameter set currently at 100 milliarcsec). A typical single loss of lock event lasts about 3 minutes, which is the time it takes to re-acquire the guide stars and re-initialize guiding. These events are managed by the onboard computers through the use of flags.

NRECENT

A knowledge of the onset and duration of recentering events is important because the position of a target within an aperture could change significantly (due to gyro drift and spacecraft disturbances) thus degrading the pointing quality. The trade-off, i.e., degraded pointing versus loss of FGS lock and significant reduction in exposure time (3 minutes or more during lock losses), has been judged acceptable.

RECENTR

V2_RMS

Caution: The jitter is referenced to the first data point. Therefore, slews (small angle maneuvers during target acquisitions, peakups, spatial scans) and moving target tracking will introduce an apparent drift. This motion will also contaminate the calculation of the jitter statistics presented in the header and artificially inflate the values (hundreds of milliarcsec or more). To determine whether a slew was in progress, the SLEW flag in the tables may be consulted (a 1 will appear in the SLEW flag column). The filtering algorithm is not applied if slews occurred during the observation.

Note: This field is blank when in GYRO mode. Also, when TLMFORM=FN, no compensation for differential velocity aberration is performed and therefore a small drift will be included in the calculation of the jitter. This problem will be corrected in a future enhancement.

V2_P2P

See keyword V2_RMS for a discussion of cautions in interpretation when the vehicle is slewing.

V3_RMS

See keyword V2_RMS for a discussion of cautions in interpretation when the vehicle is slewing.

V3_P2P

See keyword V2_RMS for a discussion of cautions in interpretation when the vehicle is slewing.

RA_AVG

See below for important details on the calculation, accuracy, and cautions in interpretation of the attitude.

Note: When TLMFORM=FN and the guiding mode is single FGS+GYRO or GYRO, the keyword will be blank: the absence of the velocity information in FN format would result in a +/- 20 arcsec error in the pointing calculation.

DEC_AVG

Note: When TLMFORM=FN and the guiding mode is single FGS+GYRO or GYRO, the keyword will be blank: the absence of the velocity information in FN format would result in a +/- 20 arcsec error in the pointing calculation.

ROLL_AVG

The CMJ, CMI and JIT tables contain the roll angle tabulated as a function of time.

Note: When TLMFORM=FN and the guiding mode is single FGS+GYRO or GYRO, the keyword will be blank: the absence of the velocity information in FN format would result in a +/- 20 arcsec error in the pointing calculation.

Pointing Calculations and Sources of Error

The 3 methods used in the calculation of the RA, Dec, and roll depend on the 3 current guiding scenarios: Two FGS Fine Lock Guiding, Gyro Control, and Single FGS plus gyro.

Two FGS Fine Lock Guiding: This mode will produce the most accurate pointing. Absolute accuracy in this mode can vary over time as calibrations and geometries change, but can be expected usually within the range 0.2 - 1 arcsec. The nominal operational FGS search radii have been between 15 and 30 arcsec. If either guidestar (correct expected brightness) is not found within this radius the acquisition will fail to achieve 2GSFL. If both FGSs acquire guidestars, the predicted minus measured separation between them must pass the coarse mode angle check which is set at 6 arcsec. If the difference is greater than 6 arcsec the acquisition will fail. It is possible given an unfortunate confluence of worst case FGS calibrations and worst case GSC error that the proper guidestars could be locked up while exhibiting a predicted-measured value of nearly 2 arcseconds. This is rare however and the value is usually a few tenths of an arcsec. It is possible for spoiler guidestar(s) or guidestars with high proper motions however to pass the brightness check and separation check while still producing an absolute pointing error of up to 6 arcseconds.

Gyro Control: Pitch, yaw, and roll control is performed by the Rate Gyro Assembly. The pointing is calculated using the data from the gyros. These vectors, (quaternions), are corrected for the differential velocity aberration between the V1-axis (for which the vectors are reported in the telemetry) and the aperture reference position. Because of the gyro drift, this mode produces the least accurate absolute and relative pointing. Absolute accuracy of the aperture RA and Dec will be 2-50 arcsec (15 arcsec 1 sigma) and positional drift will be of order 1-5 milliarcsec per second of observation time.

Single FGS Pitch and Yaw plus Gyro Roll Control: In this mode, a single FGS is used to control the pitch/yaw of the spacecraft and the roll control is handled by the gyros. The absolute pointing is calculated from the Rate Gyro Assembly output. The jitter information from the guide star in the FGS is then factored in to the absolute pointing. Finally, the combined pointing profile is determined at the position of the aperture (i.e., absolute and differential velocity aberration corrections are applied in all telemetry formats except FN). This guiding mode will produce less accurate pointing than the two FGS guiding mode because of gyro drift, in this case, a drift in roll about the single guide star. Because of the drift, relative pointing during an observation will be affected, but absolute pointing can also be impacted due to possible roll biases at the time of acquisition. Expect absolute accuracies 0.5 - 5 arcseconds. The error due to the roll drift and other perturbations during the observation will normally be ~1-2 milliarcsec/sec of roll about the dominant guidestar. Larger values are occasionally seen. A roll rate about the guidestar introduces a much smaller translation drift at the SI than the pure gyro mode. Note however that in 3 gyro mode prior to 28 August 2005, roll drift would continue to build during occultations as well as visibility periods. Example: During a single guidestar WFPC observation, if the roll drift about the guidestar 700 arcsec away from WFPC2 is 1.5 mas/sec, this results in a translational drift at WFPC2 of 0.005 mas/sec. In a 10 minutes exposure, that produces 3 mas. After 3 orbits the uncorrected drift totals 87 mas or 2 PC pixels. (In Cycle 17, the behavior of single guidestar roll accumulation across orbits is TBD.) Note that the V2V3 position of the dominant guidestar can be found from the jitter table's (_jit) columns V2_DOM & V3_DOM, while the science aperture location is obtained from jitter image (_jif) header keywords APER_V2 & APER_V3.

Roll angle changes (roll drift) during an observation:

• The roll will change when the RA and Dec change via spatial scans, offsets, moving targets, or other slews, since such small angle maneuvers will change the direction of north with respect to the aperture reference position. The pointing calculations properly record this.
• Gyro drift will cause a change in the roll angle for Gyro Control and Single Guidestar observations.
• Thermal expansion or contraction and jitter (single and two FGS guiding modes). The peak-to-peak amplitude of the breathing in the direction of pitch and yaw is less than 15 milliarcsec. The effects on the roll are still to be determined.

Users may note small shifts in images from multiple visits at the same pointing, with the same guidestars, and SAME ORIENT specified. HST sets up a given roll for an observation with gyros, and acquires the dominant guidestar. For visits with the same guidestars, target, and roll specified, the Pointing Control System (PCS) will place the dominant guidestar at the same location within the FGS to high precision (1 or 2 milliarcseconds). The subdominant (roll control) guidestar is then acquired and tracked in fine lock. The source of the small roll non-repeatability is in this step. If the subdominant guidestar is not found in exactly the predicted location, the PCS will not adjust the roll to place the subdominant in the same location. This is due to the fact that the PCS, once calibrated via the Fixed Head Star Trackers (FHSTs), can normally execute a commanded roll better than if it relied on the FGS subdominant guidestar location, with its ~0.3 arcsecond catalog error.

The accuracy with which HST is rolled for a given visit depends then not on the subdominant guidestar but on the amount of gyro bias in the PCS. Typically, an FHST update is made some time prior to a full guidestar acquisition. In an FHST update, the PCS reconciles the accumulated gyro drift with FHST (fixed head star tracker) star fields, providing a "reference update" and zeroing out most of the PCS error. In most cases a successful FHST update is made within a reasonable time before an acquisition, and under these circumstances the roll accuracy will be around 0.003 degrees. This angle of roll around the dominant guidestar works out to a positional shift of about 73 mas at the subdominant guidestar (assuming 1400 arcsecond separation), which, when this number is compared to the error in the guidestar catalog positions, illustrates why the guidestars are not used to "correct" the roll. At the WFPC2, this shift is 38 mas and just under a WFPC2 PC pixel. Therefore, multiple visits at the same specified roll, target, and with the same guidestars will under nominal circumstances show repeatability at this level. It is not uncommon for scheduling constraints to affect the time between FHST updates and GS acquisitions. Such acquisitions that take place with on older FHST update, or an inaccurate one, can suffer from larger roll deviations, 0.01 upward can occasionally be seen. Note: The jitter files, which determine roll based on guidestar locations in the FGSs can be used in the case of visits with the same guidestars and same roll, to determine quite accurately the amount of actual roll change incurred between visits, though the commanded or specified roll would be the same for each visit.

Sources of Error:

Here we discuss two types of errors:
1. Errors on positioning a target at the intended reference. (commanded positions vs. independent measurement from an image)
2. Errors between the Observation Logs' calculated target position and a measured value (Obs Log vs. independent measurement from image)

• Relative guide star position error (~ 0.3 arcsec) and target coordinate error such as proper motion or target coordinates specified from a catalog containing a rotation or zero point offset with respect to the Guide Star Catalog (single and two FGS guiding modes);
• Fine lock of an unintended or binary dominant guide star (single and two FGS guiding modes). Only the dominant guidestar is important in the case of target positioning, since the roll about that guidestar (appropriate for the target positioning) is made by the pointing control system without FGS input, and thus is not affected by the location of the subdominant guidestar in the FGS;
• "False lock" of the correct dominant guide star. (single and two FGS modes). The FGS's interferometric tracking will occasionally jump into a local minimum in the S-curve, resulting in a pointing offset (0.3 to >1 arcsec) and an abnormally high jitter.
• Focal Plane calibration errors including:
• FGS/FGS alignment (currently time dependent; has been as high as 0.3 arcsec, but after a recent calibration should be < 0.050);
• FGS/SI aperture calibration errors which can depend on the target acquisition method used and the relative errors between target acquisition aperture and science aperture, as well as other internal instrument calibrations. Values range from 0.1 arcsec to 0.5 arcsec;
• PCS calibration errors between the Fixed Head Star Trackers, Rate Gyro Assembly and FGSs
• Gyro drift of 1-5 milliarcsec/sec (single FGS and gyro modes);
• Thermal expansion and contraction, known as breathing (up to 15 milliarcsec).

• If pre-planned or real-time offsets were commanded (all guiding modes). The jitter file reports Ra, Dec and Roll always with respect to an aperture reference whose name and V2V3 position is given in the jitter file header. If an observation consists of a POS TARG offset from, say the PC2 aperture reference, the Obs Log will report the RA,Dec of the PC2 reference pixel position and not the POS TARG pixel position. In other words any motion on the sky is tracked by the Obs Log, but always with respect to the aperture reference position. This is not a source of numerical error but a point of interpretation the user of the Obs Log should be aware of;

• Positional calibration errors at the SI. The Observation Log maps a reference V2V3 position to the sky based on the attitude determination. The accuracy to which that relates to a given detector pixel depends on the calibration of that SI: (position in focal plane, distortions, etc);

• The relative guide star positional error with GSC1 over scales comparable to HST field of view, 0.33" for northern hemisphere plates and 0.43" for southern hemisphere plates (two FGS guiding mode). Since the guidestars' GSC celestial coordinates are fit to the FGS measured positions to find the vehicle attitude, the determination is affected by this error;

• FGS to FGS alignment error (currently time dependent; has been as high as 0.3 arcsec, but after recent calibrations should be < 0.050). This, like the guidestar celestial coordinates, affects the pointing calculation, in this case by introducing error in the FGS XY to V2V3 transformation;

• Fine lock of an unintended or binary guide star (two and single FGS guiding modes). This error is closely related to the first in its affect on the attitude determination;

• "False Lock" of the dominant or subdominant guidestar. (Single and two FGS modes);

• Inherent error of tenths of an arcsec in the telemetered gyro data (gyro and single FGS guiding modes);

• Gyro drift which ranges from 1-5 milliarcsec/sec depending on the gyro bias update and slew activities prior to the observation (gyro and single FGS Guiding modes);

• Thermal expansion, contraction and jitter. The peak-to-peak amplitude of the breathing in the direction of pitch and yaw is less than 15 milliarcsec. Jitter is on average about 2-8 milliarcsec RMS (single and two FGS guiding modes);

• In TLMFORM=FN where the absolute and differential velocity aberration corrections are not applied. The absolute velocity variation is up to 20 arcsec and the differential aberration is up to 50 milliarcsec (all modes);

• Negligible differences in the method of calculation between the observation log software and the operations ground and onboard flight software systems (all modes).

Calculated (jitter file) Positions versus Commanded (science headers)

It is sometimes useful to use pointing information to correlate images taken as part of a proposal for various reasons (dithering, image subtractions). Usually these observations are made with the same guidestars and are often but not always within a visit. When wishing to register or align such a set of observations without convenient point sources allowing an empirical correlation, commanded or calculated pointing information is the alternative. Here the interest is in shifts relative to some initial pointing, and one can use the commanded values (CRVALs, RA / Dec_APER, ORIENTAT) in the science header, or the jitter file's calculated RA / Dec / Roll_AVG to describe these relative motions.

What are the practical differences between the commanded and calculated shifts when used in this application? Assuming the same guidestars are used, and the shifts are correspondingly small compared to the FGS Field of View, within a single visit both the commanded and calculated positions describe well the image shifts relative to an initial point. Agreement to better than 10 mas rms is common. See ACS ISR 06-05. From visit to visit, the differences rise to ~20 mas for the commanded values and ~15 mas for the calculated values. The jitter file's calculated value knows the amount of the roll delta introduced each visit's full-acquisition, since it calculates this based on measurement of the same guidestars each time. This gives better agreement with images taken over multiple visits. But because the jitter file's evaluation of the roll is affected by guidestar catalog error, the commanded roll from the science header is usually closer to what was achieved than the jitter file's calculation (see Roll Repeatability), and the jitter file will exhibit some absolute offset of the calculated target location. However, the jitter files can be used to best describe relative motion, and the roll differences, even across full acquisitions of the same guidestars.

Finally, independently of how accurately the jitter file calculation is performed, even in the case of describing relative shifts between observations with the same guidestars, much of the ~15 mas rms level of disagreement is affected by thermal motions in the focal plane: motions within the SI, within or between FGSs, and from the OTA plate scale, and probably represents close to a limiting accuracy.

PROPOSID

TARGNAME

STARTIME and ENDTIME

SOGSID

INSTRUME

PRIMARY

OPERATE

TLMFORM

CONTENT:

APERTURE

APER_V2

APER_V3

ALTITUDE

LOS_SUN

LOS_MOON

SHADOENT

SHADOEXT

LOS_SCV

LOS_LIMB

ZODMOD

EARTHMOD

MOONMOD

GUIDECMD

GUIDEACT

GSD_ID

GSD_RA

GSD_DEC

GSD_MAG

GSR_ID

GSR_RA

GSR_DEC

GSR_MAG

GSACQ

PREDGSEP

ACTGSSEP

GSSEPRMS

NLOSSES

LOCKLOSS

NRECENT

RECENTR

V2_RMS

V2_P2P

V3_RMS

V3_P2P

RA_AVG

DEC_AVG

ROLL_AVG

Pointing Calculations and Sources of Error