Refactor/station manager calculations

[artistrea]

Considering the same number of stations, the methods changed performance went from:

   ncalls  tottime  percall  cumtime  percall filename:lineno(function)
       11    9.249    0.841   12.380    1.125 station_manager.py:277(get_pointing_vector_to)
       14    1.784    0.127    4.705    0.336 station_manager.py:139(get_3d_distance_to)

To:

   ncalls  tottime  percall  cumtime  percall filename:lineno(function)
       10    0.003    0.000    0.004    0.000 station_manager.py:277(get_pointing_vector_to)
       12    0.001    0.000    0.001    0.000 station_manager.py:139(get_3d_distance_to)

Considering that the number of calls above are the same for a SINGLE SNAPSHOT, and that the dimensions were exactly the same as in a mss dc simulation, the gains are really good.

This change, together with the change in #214 took a simulation file from 35 seconds per drop to 8 seconds per drop (as expected of an approx. saved time of 12 + 18).