The first step is to slew the telescope to the coordinates of the object to be observed, followed by placing the object in the slit so that maximum signal is received. The latter operation is called peaking-up. These two operations are performed by the TO. By running an EXEC the relevant CONFIG is loaded and data can be taken.
All data presented in this work were obtained using the following strategy:
a given astronomical source was observed for n cycles of four
observations (OBJECT-SKY-SKY-OBJECT) until the desired integration time
(or signal-to-noise ratio) was achieved
. A 
sampling was used, i.e.
each
observation consists of four integrations : a first integration
of t seconds was done, after which the integration was repeated
without moving the telescope but shifting the detector array by half a
pixel to fully sample the spectrum; two more integrations followed using
the same sampling procedure but now shifting the detector array by one
pixel (relative to the pixel position during the first
integration) in order to compensate for any bad neighbouring
pixels. Such a sampling procedure is indicated in the CONFIG file by a
sampling of 4 occuring over a range of 2 pixels. For the SKY
observations, the telescope was nodded along
the slit direction by 10 pixels so that the astronomical source remained
in the slit. These 10 pixels correspond roughly to 20''
in the
sky. The total elapsed time for the observation of one
astronomical source (excluding overheads) was therefore t x 2 x 2 x
4 x n seconds whereas the integration time per spectral point was
half the above time. t is the exposure time, 
corresponds to the sampling procedure, 4 corresponds to the four
observations OBJECT-SKY-SKY-OBJECT and n to the number of times
the cycle of four observations is carried out.