The FOS is now routinely used to acquire targets for subsequent spectroscopic observations with the GHRS. These are typically faint targets which are either too faint at all for a direct GHRS acquisition, or which can be acquired more economically with the FOS.
Recently, we had GO's who had relatively stringent requirements on the wavelength scale of the GHRS spectroscopy. This memo clarifies the uncertainties introduced into the GHRS wavelength zero-point when the target acquisition was not done with the GHRS itself (which includes a LOCATE phase) but with the FOS.
If the FOS is used for acquisition, the positional uncertainty at the end of the FOS TA (and before the telescope move to the GHRS aperture) is typically around 0.15". This value results from a binary acquisition or from a 3-stage peak-up (4.3,1.0,0.5), which are the 2 most widely used strategies. It is not worthwhile to improve this by an additional peak-up because of the much larger uncertainty introduced by the move from the FOS to the GHRS. The relative separation of the FOS and GHRS (LSA) apertures is known to about 0.2". As a result, an observer can expect an uncertainty of about 0.35" after moving to the LSA. Note that this uncertainty precludes the possibility to make such an offset into the SSA. FOS-assisted target acquisitions into the SSA are therefore not supported. An additional peak-up in the LSA would be required for an SSA observation but under all likelihood (since the FOS-assisted mode was chosen in the first place) this is not feasible due to flux limitations.
An uncertainty of 0.35" in the LSA (1.7" by 1.7") causes no problem if the prime science goal is to obtain a spectrum without significant light loss. However, the user should be aware of the fact that the spectrum will have a significant wavelength offset. The LSA samples 8 diodes. If a point source is observed and perfectly centered, wavelengths can be measured to a fraction of a diode. The exact value depends on the observational set-up (wavecal, short exposures, no carrousel movement, etc), but precisions up to a fifth of a diode are possible. If the target is not properly centered in the aperture, the positional uncertainty becomes the dominant error. A positional offset of 0.35" translates into 1.6 diodes. If the G140L grating is used, this corresponds to an uncertainty of about 200 km/sec at 1400 A. This is a fundamental limitation which can not be calibrated with a wavelength calibration. Note that specifying extra wavelength calibrations in the proposal in this case is not useful for measuring the wavelength zero-point. One possibility to calibrate for the positional uncertainty would be to obtain a map of the LSA to actually measure the position of the object --- but the cost of spacecraft time would be enormous.
Claus Leitherer18 Mar 1996