After the removal of SIB1 magnets, the balance of the bench had changed a bit. Before the operation, TX_CORR and TZ_CORR were both about 1 V; after the work, 3 V were needed for TZ_CORR (fig 1).

The motorized movable mass on the marionette has been used, in order to null the corrections: 42500 CW steps in TX and 117000 CCW steps in TZ. The whole operation took more than 1 hour (fig 2). This means that the setting of the motors is wrong: 30 mV/1000steps is a resolution which makes no sense, at least 10 times more precise and slow than really needed. The operation should not require more than 5 minutes.

The other defect of the system is that currently we don't know how far we are from the end of the run. Since the installation, the motors have been used several times, because several times there were interventions on the bench which changed the balancing. A record of the steps performed till now is missing, even because the person who uses to make the operation (me) is not used to fill a notebook. An automatic storage system would be appreciable.