Target acquisition is carried out with the Fine Error Sensor (FES). The FES is
an image dissector with an S-20 photocathode. Its response extends over
4000-7000 Å, with an effective wavelength of about 5200 Å. The FES can be
used in three modes. The first, field camera mode, is a raster scan that
provides an image of the field of view with 8-arc-second resolution. The second,
prime mode, is a tracking mode that measures the brightness and relative
position of an object with a precision of about 0.25 arc seconds. The FES can
track stars of about 13th magnitude or brighter. The third mode, search and
track, is a fast raster scan, where the scan beam stops and tracks the first
star found to be brighter than a specified threshold. This can be useful for
finding very bright stars (m
5). The FES can be used in only one of
these modes at a time.
The FES can view a field of up to 16 arc minutes on a side containing the entire reflective area of the aperture plate, but an FES field of 10.8 arc minutes square is standard for target identification. This field is typically recorded once at the end of a maneuver, enabling the GO to identify the target. A limiting magnitude of 12.0 can be reached with a standard FES image. The detection of fainter stars requires proportionally longer integrations. The FES limiting visual magnitude is about 14.0 to 14.5 in field camera mode, depending on the background sky brightness. Normal practice for viewing such faint targets is to do an offset maneuver from a nearby bright star and then raster scan a smaller (1 or 2 arc minute) square field, known as a "postage stamp". FES images are normally not archived, but can be if the GO so requests.
In the tracking (prime) mode, the FES is used as an image dissector with a cross-shaped pattern. The dimension of the scan pattern is 12 arc seconds for stars fainter than 4.7 mag (overlap mode) and 22 arc seconds for brighter stars (underlap mode). Near the center of the telescope field of view the FES can give relative stellar positions that are accurate to about 0.25 arc seconds. Away from the center of the field distortions reduce the accuracy to about 3 arc seconds.
Since the FES receives the telescope field of view as reflected by the aperture plate, it cannot see a target that is in one of the spectrograph apertures. Therefore, the FES can be commanded to track on a convenient field star for offset guiding during long exposures.
Because the FES is a visual-sensitive instrument, scattered Earth and Moon light can contribute to the measured sky brightness. Sky brightness can hinder the identification of a star field or corrupt the FES automatic tracking mode. The normal sky brightness is about 16th magnitude per pixel (8 x 8 arc seconds), but will be increased to 14th magnitude at a distance of 50 degrees from the Earth and 10th magnitude at a distance of 15 degrees. Viewing at 0.5 degrees from the full Moon gives a sky brightness of 8th magnitude per pixel.
The sky background is slightly enhanced, up to 13th mag, due to the FES
scattered light anomaly(Teays 1991,
Carini 1991). This anomaly (of unknown cause)
is strongly 
dependent, being strongest at high . The additional
background is very uniform, so the FES can lock up and track a star which is
not easily visible on the FES image, providing that we can tell the FES
where to look. This means that using a very faint guide star at high
requires accurate coordinates, or a previous observation using that guide
star, so that its position in the FES can be calculated with TOCC software.
Use of the HST Guide Star Catalog, which is available in the TOCC,
can be helpful in this respect.
The FES has occasionally suffered from errors of up to 5 arc seconds in its positional reference. These errors correlate with opening and closing of the large apertures and the sun shutter, and are referred to as "reference point shifts". Accordingly, to reduce the risks of degradation of spectra and to minimize the overhead time used to measure the error, closing and opening of the large aperture and the sun shutter is done only in the case of a strong scientific need. The acquisition of platinum-neon lamp spectra used for wavelength calibration requires use of the shutter mechanism. If special purpose wavelength-calibration spectra are requested by the GO, they should be taken after, not before, the astronomical data. Normal wavelength calibration and monitoring is provided by monthly wavelength-calibration spectra obtained by the Observatory. This, combined with the Observatory image reduction software, provides more than sufficient accuracy (i.e. internal errors of less than 3 km/sec for high dispersion spectra; Turnrose and Thompson 1984) for virtually all GO programs.
In the prime mode, an object's FES counts can be used to derive an estimate of its visual magnitude (see section 4.7).
As seen in Figure 2.1, there are actually two Fine Error Sensors. FES No.2 has been used continuously since March 1978 because it is 2 magnitudes more sensitive than FES No.1. FES No.1 is available for use in case of the failure of FES No.2.
