Validating Phasing and Geometry of Large Focal Plane Arrays
- Tuesday, 25 October 2011
CCD defects are used here to advantage.
The Kepler Mission is designed to survey
our region of the Milky Way galaxy to
discover hundreds of Earth-sized and
smaller planets in or near the habitable
zone. The Kepler photometer is an array
of 42 CCDs (charge-coupled devices) in
the focal plane of a 95-cm Schmidt camera
onboard the Kepler spacecraft. Each
50×25-mm CCD has 2,200×1,024 pixels.
The CCDs accumulate photons and are
read out every six seconds to prevent saturation.
The data is integrated for 30
minutes, and then the pixel data is transferred
to onboard storage. The data is
subsequently encoded and transmitted
to the ground.
During End-to-End Information
System (EEIS) testing of the Kepler
Mission System (KMS), there was a need
to verify that the pixels requested by the
science team operationally were correctly
collected, encoded, compressed,
stored, and transmitted by the FS, and
subsequently received, decoded, uncompressed,
and displayed by the Ground
Segment (GS) without the outputs of
any CCD modules being flipped, mirrored,
or otherwise corrupted during
the extensive FS and GS processing. This
would normally be done by projecting
an image on the focal plane array (FPA),
collecting the data in a flight-like way,
and making a comparison between the
original data and the data reconstructed
by the science data system.
Projecting a focused image onto the FPA through the telescope would normally involve using a collimator suspended over the telescope opening. There were several problems with this approach: the collimation equipment is elaborate and expensive; as conceived, it could only illuminate a limited section of the FPA (≈25 percent) during a given test; the telescope cover would have to be deployed during testing to allow the image to be projected into the telescope; the equipment was bulky and difficult to situate in temperature- controlled environments; and given all the above, test setup, execution, and repeatability were significant concerns. Instead of using this complicated approach of projecting an optical image on the FPA, the Kepler project developed a method using known defect features in the CCDs to verify proper collection and reassembly of the pixels, thereby avoiding the costs and risks of the optical projection approach.
The CCDs composing the Kepler FPA, as all CCDs, had minor defects. At ambient temperature, some pixels look far brighter than they should. These “hot” pixels have a higher rate of charge leakage than the others due to manufacturing variations. They are usually stable over time, and appear at temperatures above 5ºC. The hot pixels on the Kepler FPA were mapped before photometer assembly during module testing. Selected hot pixels were used as target “stars” for the purposes of EEIS testing. “Dead” pixels are permanently off, producing a permanently black pixel. These can also be used if there is some illumination of the FPA.
During EEIS testing, Dark Current Full Frame Images (FFIs) taken at room temperature were used to create the hot pixel maps for all 84 Kepler photometer CCD channels. Data from two separate nights were used to create two hot pixel maps per channel, which were cross-correlated to remove cosmic ray events which appear to be hot pixels. These hot pixel maps obtained during EEIS testing were compared to the maps made during module testing to verify that the end-to-end data flow was correct.