A Collection of Observed WF3 F606W PSFs John Krist & Chris Burrows STScI September 5, 1996 As part of HST program 5849, images of four different stars were obtained in the center of WF3 in filter F606W. The stars range from B-V = -0.073 to +0.34. For each star a 3 x 3 spatial scan was used, forming a grid in which the star images are separated by 0.55 arcsec (5.5 pixels) in both X and Y. The data were processed (dark subtraction, flat fielding, etc) by the standard HST data pipeline and extracted from the archive. A uniform background level has been subtracted from each image. The background was computed using the mean of an empty region of sky near the top edge of the chip. The centers of the stars were determined by eye. A 261 x 261 pixel subimage centered on the star was extracted. In regions outside of about 2.2 arcsec, cosmic rays were replaced by nearby patches of the sky (to preserve the "noisiness"). Cosmic rays inside that radius have not been removed. A single low column was replaced by the average of adjacent columns. Nearby background objects such as stars and galaxies have not been removed. At the end of each visit a series of four images with increasing exposure times were taken at one position. Saturated pixels in an image were replaced with those from the next less-exposed image, scaled for exposure time. All four images are used to produce a PSF which samples the full dynamic range of the star image. Saturated columns which were replaced with pixels from shorter exposure images show increased noise levels. The best results are obtained by choosing the PSF whose coordinates and color are closest to those of your object (typically location is more important than color). The PSF should be shifted at the subpixel level - bilinear interpolation works rather well with saturated PSFs, where you are only interested in the wings and not the PSF core. Line the PSF up with your object based on the residual pattern in the spiders. This should be sensitive to about 0.1 pixels. When using any of these images for PSF subtraction (or any HST PSF, observed or modeled), you should NOT determine normalizations based on the spiders. The ripple pattern in the spiders (the pattern of light-dark-light as radius increases) varies considerably with position, object color, and focus. Normalization is probably best determined by eye if there is material around your object. You will note that in many of the images there is horizontal banding, where the background level appears brighter. This is due to a capacitance effect in the camera amplifiers. The background level in a row is proportional to the number of saturated pixels in that row. This effect varies considerably from exposure to exposure, and may lead to noticeable over- or under- subtractions. As an example of the errors one can expect using these PSFs, we experimented combined MMJ6481.FITS and MMJ6490.FITS PSFs, which are separated by 10 pixels on the detector. MMJ6490 was multiplied by 2.3 and shifted by (0.0,-0.35) pixels to match MMJ6481. The percentage difference was computed, (MMJ6481-MMJ6490)/MMJ6481. Pixels in the spiders, saturated columns, and nearby stars were excluded. Using data from the region where the PSF signal dominates (where MMJ6481 pixels are > 14 DN), the average error is 11% and is normally distributed. The maximum error is 365% (a noisy pixel near the core). The PSFs are provided as FITS files, with the listed dataset name as the rootname. The target is the rootname of the combined PSF for each star. V and B-V are taken from the proposal. Prop Exptime Center Id Dataset Target sec X,Y -------------------------------------------------- 5849 U2SU0101T MMJ6481 40.000 442,411 5849 U2SU0102T MMJ6481 40.000 437,411 5849 U2SU0103T MMJ6481 40.000 431,411 5849 U2SU0104T MMJ6481 40.000 431,417 5849 U2SU0105T MMJ6481 40.000 437,417 5849 U2SU0106T MMJ6481 40.000 442,417 5849 U2SU0107T MMJ6481 40.000 442,423 5849 U2SU0108T MMJ6481 40.000 437,422 5849 U2SU0109T MMJ6481 40.000 431,422 5849 U2SU010AT MMJ6481 0.110 437,417 5849 U2SU010BT MMJ6481 1.000 437,417 5849 U2SU010CT MMJ6481 10.000 436,416 5849 U2SU010DT MMJ6481 100.000 437,417 Combined PSF = MMJ6481.FITS Star is V = 10.03?, B-V = -0.073 5849 U2SU0201T MMJ6490 40.000 438,420 5849 U2SU0202T MMJ6490 40.000 433,421 5849 U2SU0203T MMJ6490 40.000 427,420 5849 U2SU0204T MMJ6490 40.000 427,426 5849 U2SU0205T MMJ6490 40.000 433,426 5849 U2SU0206T MMJ6490 40.000 438,426 5849 U2SU0207T MMJ6490 40.000 439,431 5849 U2SU0208T MMJ6490 40.000 433,432 5849 U2SU0209T MMJ6490 40.000 427,432 5849 U2SU020AT MMJ6490 0.230 433,426 5849 U2SU020BT MMJ6490 2.000 433,426 5849 U2SU020CT MMJ6490 20.000 433,426 5849 U2SU020DT MMJ6490 100.000 433,426 Combined PSF = MMJ6490.FITS Star is V = 10.03?, B-V = 0.11 5849 U2SU0301T MMJ6504 40.000 430,408 5849 U2SU0302T MMJ6504 40.000 424,408 5849 U2SU0303T MMJ6504 40.000 419,408 5849 U2SU0304T MMJ6504 40.000 419,414 5849 U2SU0305T MMJ6504 40.000 425,413 5849 U2SU0306T MMJ6504 40.000 429,414 5849 U2SU0307T MMJ6504 40.000 430,419 5849 U2SU0308T MMJ6504 40.000 424,419 5849 U2SU0309T MMJ6504 40.000 419,419 5849 U2SU030AT MMJ6504 0.230 419,419 5849 U2SU030BT MMJ6504 2.000 419,419 5849 U2SU030CT MMJ6504 20.000 419,419 5849 U2SU030DT MMJ6504 100.000 419,420 Combined PSF = MMJ6504.FITS Star is V = 10.94, B-V = 0.22 5849 U2SU0401T MMJ6511 40.000 442,419 5849 U2SU0402T MMJ6511 40.000 436,419 5849 U2SU0403T MMJ6511 40.000 431,419 5849 U2SU0404T MMJ6511 40.000 431,425 5849 U2SU0405T MMJ6511 40.000 437,425 5849 U2SU0406T MMJ6511 40.000 442,424 5849 U2SU0407T MMJ6511 40.000 442,430 5849 U2SU0408T MMJ6511 40.000 437,430 5849 U2SU0409T MMJ6511 40.000 431,431 5849 U2SU040AT MMJ6511 0.230 431,430 5849 U2SU040BT MMJ6511 2.000 431,430 5849 U2SU040CT MMJ6511 20.000 431,430 5849 U2SU040DT MMJ6511 100.000 431,430 Combined PSF = MMJ6511.FITS Star is V = 10.6, B-V = 0.34