Modification to observed pointing wobble during staring observations

Issued: 21 October 2010

As of 17 Oct 2010 07:25:00 UTC, the pointing wobble observed in long staring 
observations with IRAC was significantly reduced in amplitude and the period 
decreased from ~60 minutes to ~40 minutes. The cyclic pointing wobble and the 
response variations across a given detector pixel (the pixel-phase effect) for 
the IRAC arrays combine to produce a variation in photometry for long staring 
observations. Observers will still need to correct the photometric variation for 
science (e.g. exoplanet transits) requiring high precision. The expectation is 
that by reducing the amplitude and increasing the frequency of the wobble the 
resulting photometry will have better precision and residual photometric 
fluctuations will have a period which is much shorter than the periods of 
astronomically interesting (transits and secondary eclipses) events.

From a comprehensive analysis, the Spitzer engineering team was able to 
correlate the pointing wobble with the cycling of a heater used to keep a 
battery within its operating temperature range. The heater cycling results in a 
1 degree Celsius temperature change to the battery with a period of one hour. 
After extensive testing and review, it was determined that the heater cycling 
could be safely reduced to 0.5 degrees Celsius. It was hypothesized that the 
thermal cycling led to a small flexing between the telescope boresight and star 
trackers used to keep the spacecraft attitude constant, and that reducing the 
cycling would reduce the associated flexing and pointing wobble. Please note 
that the pointing wobble is much smaller than the Spitzer pointing stability 
requirement and that the pointing performance of the observatory far exceeds its 
design specification and requirement.

The efficacy of this change was tested by performing identical staring 
observations of the standard calibration star, HD 158460, which has exhibited no 
measured variability in the infrared. The observations are identical three-hour 
stares of the star using 0.4 second subarray frames at 4.5 microns. The 
observations were executed back-to-back with the heater cycling changed in 
between. The pre-change AOR is 40837888 and the post-change AOR is 40837888. The 
data are publicly available via Leopard and the Spitzer Heritage Archive. We 
encourage interested observers to download the data and compare the resulting 
light-curves.

The SSC has performed a preliminary analysis of the data and finds that the 
photometric variations associated with the pointing wobble are reduced. The
period of the pointing wobble after the heater cycling change is estimated to
be ~36 minutes from a periodogram analysis of the measured stellar centroids. 
It is possible that this period will change (decrease) as a new equilibrium is 
reached with the implementation of the heater cycling change.

A figure at

http://ssc.spitzer.caltech.edu/warmmission/news/relphoto_sm.png

shows a comparison of the relative flux variation with time for two staring
observations of HD 158460. The black curve is the data from the pre-change AOR
(40837888). The green curve is the data from the post-change AOR (40837888).
Each data set has been boxcar smoothed using a window of 101 samples.

No additional structure is noted in the light curve. After fitting for the 
pixel-phase effect using a 2nd order polynomial in deltaX, deltaY (relative 
centroid position), the signal-to-noise of the pre-change data is 444.3 per
exposure, the signal-to-noise for the post-change data is 463.2, while the
theoretical signal-to-noise for the observations should be 533.6. That is, the
pre-change SNR is 80% of the ideal, and the post-change is 84% of ideal. This
initial analysis suggests that the precision of the binned data after the 
change is even better. A more in-depth analysis will be presented in an
upcoming memo.