Summary
Based on ACA fiducial light data, the IRU swap (circa July-2003) has caused an apparently permanent alignment shift of about 4 arcsec. Looking at the fid light drift data, we see an aligment shift peaking at about 6 arcsec, associated with the powering up of IRU-2 and possibly the ACA CCD cooldown. This is seen as a rapid drift of the fiducial light positions beginning at the time when IRU-2 was activated. The amplitude is about 5 arcsec in ACA Y angle and 3 arcsec in Z angle. There is some evidence of a smaller shift beginning at the time of the ACA cooldown. The alignment shift is NOT occurring in the FTS periscope, because there has been no degradation in the X-ray celestial location accuracy over this time period. The most likely source of the shift is in the ACA mount itself. Note that a much larger shift (60 arcsec) was seen after thermal cycling during thermal vacuum testing at TRW.Data
The plots below show the dramatic shift in the apparent fid light positions, relative to the nominal position from early in the mission. In this plot the data points cover the entire mission. The different colors and symbols mark different fid lights. Here we show only ACIS-S fid lights, but the same trend is seen for the other three SIs. The Z drift data are inherently noiser because of the OBA flexure which produces a SIM Z alignment shift when exposed to the sun at forward attitudes.
In the plots below we expand to show data covering about 7 months (March through Sep 2003), which includes the IRU swap and ACA CCD cooldown. (The fact that these two events overlap in time is certainly unfortunate.) The beginning and end of the ACA cooldown is marked, as are the times when IRU-2 was powered on and IRU-1 was powered off. It is fairly clear that the steep drop seen in both the ACIS-S and ACIS-I data coincides with the IRU-2 power-up. The data are noisy, but consistent with an exponential thermal equilibration.
In the ACIS-S data, there is some evidence of a drift of about 1 arcsec beginning at the time of the first ACA cooldown. The ACIS-I data are too noisy to discriminate whether this apparent trend is real.
After turning off IRU-1, the alignment recovered only slightly (perhaps by an arcsec), and we see that swapping from IRU-1 to IRU-2 induced a permanent alignment shift of nearly 4 arcsec. It is speculated that IRU-2, because of the different mount configuration, presents a thermal profile which differs from that of IRU-1. Other thermal data analyzed by the FOT support this theory.
Recent X-ray data from observations as late as 2003-July-17 have been processed and checked for any "aspect offset" (an error in the absolute astrometry of processed X-ray data). No such offset is seen, which gives high confidence that the FTS periscope alignment has been largely unaffected by the changing thermal environment. (Any change in periscope alignment translates directly into an error in absolute aspect coordinates).
Operational impact
The shift in alignment seen here is likely a contributor to the IRU-2 calibration offset of 26 arcsec noted by Bill Davis. However, based on these data most of offset appears to be in the IRU itself.It remains to be checked whether the fid light positions are still sufficiently well centered within the 25 arcsec (half-width) search box. No problems have been reported thus far in ground aspect processing, but it may be the case that the CHARACTERISTICS will have to be updated.
Analysis
The analysis for this report was done using data and code in the direoctory /proj/sot/ska/analysis/fid_drift.
Aspect Information main page
Comments or questions: Aspect Help
Last modified:12/27/13