Precision of the resolvers

The accuracy of the two Artus resolver is of primary importance when tracking a target. We recently ran an experiment to observe the evolution of AZ and EL on two very simple trajectories.

The following figure at the top shows an evolution of AZ with EL constant. The lower figure is an estimation of the speed using a Savitzky–Golay derivative filter:


Same experiment now for EL (AZ is constant):

expe-20150326-el_velocity_1We note that the behaviour of the these two trajectories are not exactly the same. EL exhibits an oscillating random noise that is acceptable. Some parts of the AZ trajectory shows odd irregularities, for example a speeding zone between 10 and 40ms, and showing a plateau right after 40ms.

We do not have good explanations yet to figure out what’s going on AZ (Is PB8 really stopping periodically ? Is there something wrong on the return chain (mechanical gears)   ?