- Pipeline Calibration Files
- Flat Field Images
- Reseau Positions
- Objective Prism Calibrations
- Linearity Relations
- Point Spread Functions (PSFs)
- Filter Shifts
FOC images are routinely calibrated using three sets of calibration files: flat field images, geometric correction files, and instrument sensitivity tables. Full descriptions and histories of these files and tables can be found in the FOC Instrument Science Report FOC-082 [Postscript, 150 kb]). All reference files can be obtained from the HST Archive.
The FOC DQE curve has recently been updated based on Cycle 4 calibration observations. The new DQE curve corresponds to the curve found in Versions 5-7 of the FOC Instrument Handbook and is coupled with the encircled energy curves found there. An FOC Instrument Science Report FOC-085 [Postscript, 386 kb] describes this new DQE curve and it calibration.
Unsmoothed versions of the flat fields for the f/96 relay and the f/48 relay illustrate the locations of small scale features in FOC images which were not removed in the 'pipeline'. An unsmoothed version of the f/96 flat field using the latest geometric correction has been made available for comparison with post-COSTAR f/96 images.
FOC Reseau positions for the pre-COSTAR f/96 and f/48 relays are available. These files can be used for comparison with FOC images to determine whether a source has fallen onto a reseau. They also provide an indication of the amount of geometric distortion that is corrected in FOC images.
Objective Prism Reduction Techniques and Software:
An Instrument Science Report (FOC-092) [Postscript, 738 kb] has been written to describe the most recent calibrations of the objective prisms. This report also describes the techniques and available software for extracting a spectrum from an image and reducing the spectrum. This tutorial will be useful for anyone interested in using the objective prisms or for anyone with objective prism images already.
Dispersion curves provide the relation between wavelength and offset in pixels from the undispersed image. These curves are available as ASCII tables for the following prisms:
- f/96 Far-UV Objective Prism (FUVOP)
- f/96 Near-UV Objective Prism (NUVOP)
- f/48 Far-UV Objective Prism (FUVOP)
- f/48 Near-UV Objective Prism (NUVOP)
The dispersion curve for the f/96 NUVOP has been re-calibrated based on post-COSTAR observations of a spectro-photometric standard and a wavelength standard. As a result, this curve will wavelength calibrate a NUVOP image to within 0.5% of the wavelength below 3800 Angstroms, and 1% from 3800 to 6000 Angstroms, as long as the position of the undispersed target is known to within a few pixels (less than 4).
Software has been written for the STSDAS package to wavelength calibrate and photometrically reduce objective prism data. The reduction software has been simplified into one routine: objcalib. After extracting a spectrum from an unrotated image using the apextract package, objcalib will produce a flux- and wavelength-calibrated 1-D spectrum. For f/96 NUVOP spectra, the photometry can have errors as low as 10% below 3800 Angstroms with the new dispersion curve, and possibly as low as 20% from 3800 to 6000 Angstroms.
The effects of high count rates on the photometry of sources in FOC images has been characterized and reported in the Section 6.2 of the FOC Instrument Handbook. Some methods can be used to partially correct for the effects of high count rates.
Many PSFs have been taken with the FOC to characterize the shape of point sources in all wavelengths visible to the FOC. A library of observed PSFs for all relays has been established to assist users in analyzing or reducing their FOC data.
A description of the pre-COSTAR FOC PSF along with observational effects that change the PSF can be found in a report by D. Baxter (PostScript version[115 kb]). A summary of the COSTAR-corrected PSFs taken since SMOV is also available.
Calibrations comparing f/96 images of a standard star taken with different filter combinations have measured the image shifts induced by the individual filters. For most UV filters, this image shift is negligible, being less than 1 pixel. However several filters, primarily in the visible, induce shifts of 0.1" (7 pixels) or more. Fortunately, none of the ND filters showed that they induced any significant shifts. These shifts should be taken into account if planning an INT ACQ, as it would introduce extra error in the final position of the target, especially if a small aperture is being used.