The S functions described above apply only to point source observations acquired through the large aperture. The small-aperture and trailed observing modes are known to have different relative responses as compared to large-aperture point spectra (e.g., Harris and Cassatella 1985, Bohlin 1986, Crenshaw and Park 1989). This effect has been ascribed to several sources, including vignetting effects of the entrance apertures in trailed and small-aperture spectra and spatial inhomogeneities in the UVC efficiency (Cassatella 1990). Application of the large-aperture point source calibration to spectra obtained either in trail mode or through the small aperture will introduce photometric errors of at least 5-7% in the regions near the detector edge and of 1-2% in regions where the camera response is flatter.
Calibration corrections for small-aperture observations and for large-aperture trailed observations have been derived and are applied by the processing system when necessary. IUE standard-star observations from the 1984-1985 epoch have been used to calculate the wavelength-dependent flux ratios of small- and large-aperture point sources, and of large-aperture point source and trail mode observations. Observations from the 1984-1985 epoch were chosen so as to match the large-aperture point source calibration epoch. Also, by limiting the range of time over which the observations were obtained, effects due to sensitivity degradation in the cameras are minimized.
Optimal exposure level observations of the four TD1 standard stars BD+28 4211, HD 93521, HD 60753, and BD+75 325 were used to compute the small-to-large aperture (S/L) ratios and the large-aperture trailed-to-point source (T/L) ratios. The effective exposure times of the trailed observations were determined using a value of 21.48 arcseconds for the SWP major-axis trail path length (Garhart 1992b), 22.55 arcseconds for the LWR major-axis trail path length, and 21.84 arcseconds for the LWP major-axis trail path length. The ratios of pairs of observations of the same object obtained through the small and large apertures, or in point and trail modes were averaged together to determine the mean S/L and T/L spectral ratios. Approximately 20 pairs of spectra were used to determine each of the two response ratios. The mean spectral ratios were resampled into the bin size of the appropriate inverse sensitivity function and a spline fit to the binned ratios was calculated. Tabulated values of the spline fits for each camera are listed in Tables 11.5-11.8.
Because centering errors in the small aperture can lead to large variations in the overall observed flux level for individual spectra, it is impossible to determine an absolute S/L ratio. Therefore, the average of S/L over all wavelengths is normalized to unity. As a result, the small-aperture fluxes are known in a relative sense but not in an absolute one. The relative small- and absolute large-aperture inverse sensitivities are related by
Investigators should be aware that absolute fluxes for small-aperture data are significantly less reliable than those of large-aperture data. For the ratio of trailed response to point sources in the large aperture, the absolute calibrations are related by Only trailed large-aperture spectra are calibrated with the T/L ratio applied. Images processed as extended sources are calibrated as point source observations.