PDS_VERSION_ID = PDS3 RECORD_TYPE = "STREAM" OBJECT = TEXT NOTE = "Known anomalies concerning this EPOXI data set" PUBLICATION_DATE = 2010-09-10 END_OBJECT = TEXT END This file describes known anomalies concerning this EPOXI data set as of 26 November 2010. Keywords in the Primary FITS Header =================================== - The correct unit for NOMPXLSZ and RECPXLSZ is radian, not microradian as specified in the comment. - MSNPHASE was not encoded for the EPOCh Mars observation and is set to 'COMMISSIONING', the default value. It should be 'CRUISE 2'. - RANGE, CTRDIST, and RANGECEN have intentionally been set to the same value: the distance from the spacecraft to the center of the target. Anomalous Target-Spacecraft Velocities ====================================== Summary ------- The way geometry is calculated in the Science Data Center (SDC) pipeline for the EPOCh phase of the EPOXI mission yields some anomalously large values of the TARSCVX, TARSCVY, TARSCVZ, SCSUNRX, SCSUNRY and SCSUNRZ FITS header keywords and their corresponding keywords in the PDS label EPOXI:TARGET_SC_VELOCITY_VECTOR and SC_SUN_POSITION_VECTOR for extra-solar targets. Users are advised to ignore these values for extra-solar targets; note however that these keywords are generally of little use for such targets and little is lost by having them there. The awareness of this issue came about late in the archive process and so can only be documented in this delivery. The pipeline will be modified in the near future to resolve this issue; since integration of updated ephemerides into post-V1.0 data set versions are a normal part of the archive process and all of these data will be re-run through the corrected pipeline, this minor inconvenience to PDS users is anticipated to be temporary. In the meantime, anyone needing resolution of the issue is welcome to contact Brian Carcich, SDC System Manager at Cornell University, for consultation. Details ------- Because of the velocity of the spacecraft relative to the solar system barycenter, there is a shift in the *apparent* position of any target; the correction for this is called stellar aberration. The stellar aberration correction is well documented elsewhere (e.g. see any astronomy text or the SPICE toolkit required reading files at http://naif.jpl.nasa.gov/); it will not be described here other than to say it does not change the length of calculated vectors but it does change their direction by up to about 100 (Vspacecraft / c) microradians for a typical mission's spacecraft velocity (30km/s barycentric relative). The correction was applied in the calculations to determine the apparent vector of the target with respect to the spacecraft. For convenience, the vector from the spacecraft to the target was then inverted to get the vector from the target to the spacecraft, and the corrected spacecraft-target and target-sun vectors were linearly combined to get the sun-spacecraft vector. Because the heritage of the pipeline geometric calculations is for bodies within the solar system (Spacecraft, Tempel 1, Earth, Moon, Sun, etc.), the finite but insignificant errors introduced by these conveniences have been accepted by users of the SDC and the archived DI data sets. Also because of that heritage, extra-solar targets (stars and extra-solar planetary systems such as the objects of the EPOCh phase of the EPOXI mission) have been modeled by the Deep Impact pipeline via fixed-position body proxies at a range of 100 light-years (ly). This simplified the geometric calculation as separate source code paths were not required to process the two types of targets. When that decision was made circa 2005, the stellar aberration correction algorithm used by the SPICE toolkit to calculate apparent velocity was a very rough approximation that did not take spacecraft acceleration into account. This approximation was adequate for most SPICE users' needs. It also did not cause any problems when used by the Deep Impact pipeline. In March 2008, the NAIF group at JPL implemented a better stellar aberration algorithm and distributed it with the SPICE toolkit version N0062. As a result, when this new version of SPICE was installed on a SDC pipeline in early 2009 in preparation for PDS and MAST archiving efforts, the effect of the spacecraft acceleration over the modeled 100 years of light travel time to the proxy bodies for extra-solar targets resulted in numerically correct, but anomalously large, apparent velocities. Note that the issue is with the extra-solar target model used and not with the SPICE toolkit: the SDC has conferred with NAIF to verify that these velocities are being calculated correctly by the SPICE toolkit routines for the target model used. The description of this anomaly was provided by Brian Carcich. Timing Discrepancy ================== Spacecraft clock SPICE kernels (SCLK) prescribe the conversion from spacecraft clock time to UTC. For EPOXI data acquired in 2008, the data pipeline used flyby SCLK versions 35 and 36. However SLCK version 36 has retroactive changes compared to 35 because, while preparing the EPOXI SPICE archive, the NAIF team pushed for a re-fit of clock correlation data for the flyby spacecraft from 6 June 2008 through 21 January 2009. While version 36 of the SCLK has accuracy of no better than 0.5 seconds, 35 has errors of up to 3 seconds for some periods within the date range that was refitted. Because the last self-consistent entry for both clock kernels was effective for 6 June 2008, any EPOXI data acquired on or after this date for which the pipeline applied SLCK version 35 instead of 36 may have time stamp (and thus geometry) inaccuracies of up to 3 seconds. Because timing is critical for photometry of the EPOCh exoplanet targets, the UTC times for data that had version 35 applied were compared to times computed independently using version 36. Comparison of these different time stamps showed the majority of the affected EPOCh exoplanet data have timing inaccuracies under 0.5 seconds while the rest up to 0.75 seconds. This subset of EPOXI data is affected: - Some HRI Visible CCD (HRIV) observations acquired on 24 June 2008 during a standard cruise calibration for EPOXI, - Some HRIV dark and stim lamp frames acquired from 28 June through 31 August 2008 for EPOCh, and - About half of the EPOCh observations of HAT-P-4, HAT-P-7, TRES-2, and WASP-3 acquired from 28 June through 31 August 2008. For this delivery, none of these data was reprocessed using SCLK version 36 because a) the sub-second time errors have minimal effects on calibration-related data and b) the EPOCh team determined sub-seconds timing errors did not effect their photometric analysis of extrasolar transit observations. However, a list providing the recalculated mid- observation Julian date and UTC using SCLK version 36 for the affected EPOCh stellar transit observations is located in the DOCUMENT/ directory as ERRATA_TIMING.TXT.