The relative positions of the various entrance apertures and detectors need to be known to ensure that a single grating wavelength calibration can be applied to all combinations of entrance slits and detectors. The determination of the Focal Plane Geometry (FPG) was performed during the ISO Satellite Commissioning Phase Phase (on revolutions 12, 14, 17, 18, 19 and 20).
The offsets between each of the three SWS apertures and the ISO OSS (optical support structure) optical axis were calibrated by making raster maps around point-like sources, with the two SWS grating scanners set at fixed positions, corresponding to selected key wavelengths as defined in the flux calibration programme. In order to ensure the most accurate pointing, five guide stars were used (normally less were used when making observations).
The first measurements to be performed were square rasters around
Dra, one of the main SWS photometric calibrators, in revolution 17.
These were used to calibrate apertures 2, 3 and 1. The calibration
derived was verified with a finer square raster on aperture 2 and by
cross-like rasters on the other two apertures on the following
revolution. In order to avoid possible effects specific to Dra and
its associated guide stars for that day, another six point-like sources
were used for further verifications in revolutions 19 and 20.
The verification and fine tuning rasters were cross-like in order to optimise the use of time. The cross-like axis were parallel to the y-axis (cross-dispersion) and the z-axis of the spacecraft respectively. Each part of the cross consisted in a pair of scans, back and forward, in order to counteract possible detector hysteresis effects.
At the end of revolution 20, SWS02 line scans were
performed at each side of the slits. As of revolution 21, the nominal
pointing mode with one guide star was used, and the FPG was verified on
Dra by means of a smaller square raster around it.
The FPG calibration from revolution 19 proved to be accurate to the
arcsecond level, and no further updates were requested. The analysis of
the revolution 20 raster however (done with one guide star) showed a
2
offset in the z direction on Dra. This was
judged not worth a further update of the FPG offsets
(See Salama et al. 2001, [33] and [16] for
more details).
In the direction perpendicular to the dispersion, the slits are oversized. There the fields of view are determined by the dimensions of the detectors. The cross-dispersion dimensions are different for almost all detector bands. Since the imaging of the slits onto the detectors (or vice versa) is imperfect due to aberrations, the short sides of the fields of view (detector edges) are more fuzzy than the long sides (the slit jaws).
Small offsets in the fields of view perpendicular to the dispersion direction are caused by alignment errors in the instrument. The internal alignment specification allowed misalignments up to 10% of the detector width and such misalignments have been established.
The monochromatic images of the grating detectors fill about 55% of the slit widths. This means that the spectral resolution for point sources is higher than for extended sources, in a ratio that is affected, of course, by diffraction. See Section 4.5 for more information on this. For extended sources the dispersed detector image fills all of the slit width.
The monochromatic detector images of each Fabry-Pérot detector fills about 50% of the slit width. The FP etalon between the slit and the detector transmits in very narrow wavelength bands at a mutual distance of the order of the spectral resolution offered by the grating spectrometer. The detector image seen in the slit is dispersed by the grating and has fringes parallel to the grating due to the Fabry-Pérot resonances. Thus, the FP detector field of view arises from a dispersed detector image, modulated by the FP fringe pattern.
If the FP is properly tuned, an FP transmission fringe runs through the centre of the slit. The width of this fringe is determined only by the monochromatic detector image. This then is the effective beam width for FP observations on a point source, certainly applicable to the response to a narrow spectral line.
The fringe separation is on the order of the detector width so that there will always be several fringes at least partly visible in the slit. For an extended source the detector is effectively seen to fill the entire slit width, as is the case for all the grating detectors. But the detailed shape of the beam profile should vary rapidly with wavelength, as the FP fringes in the dispersed detector image move across the slit.
Extended sources observed with the FP will have more leakage in unwanted FP orders than point sources. But spatial extent does not affect the spectral resolution of the FP's.