deletions | additions       diff --git a/At_installation_time_the_beam__.tex b/At_installation_time_the_beam__.tex  index 17df85a..d1e68e8 100644  --- a/At_installation_time_the_beam__.tex  +++ b/At_installation_time_the_beam__.tex 
...
The initial alignment of a beam train takes place using the BCA primary fiducial as before (there exist additional options to use the UTM light injector in addition to or instead of the primary fiducial for this task, but these will not be discussed further here).  The start-of-night alignment is changed from the original scheme. Instead of using the primary fiducial, the UTM light injector is switched on. The location of the spot seen  on and its the FTT camera is set to be the fiducial point to which the FTT servoes starlight; as a result all subsequent starlight beams will have the same  tilt and as the injected beam. At the same time the  shear of the injected beam is measured in the BCA using BEASST, while moving the delay line backwards and forwards. The tilts of M4 and M5  are adjusted so that the  measured shear coincides with a fiducial point on BEASST and does not change when the DL is moved: this condition means that the injected beam is aligned with the DL axis.   At this point, the BCA beam injection (MOB) can be switched on and aligned with the UTM injected beam  using BEASST. BEASST to compare the tilt and shear of the two beams. Further alignment of the fringe tracker and piston terms (requirements 4 and 5) can then proceed as in the original system. The difference comes during the night: to maintain alignment in the presence of thermal drift, BEASST can either be used in its original mode with starlight, or the UTM alignment source can be switched on while slewing between stars and used to check for thermal drifts.