After 10 days taking data in this configuration, it was clear that there was no improvement at any frequency due to the shortening of the optical path in air (120cm -> 55cm).
There was however a worsening of the sensitivity around 90Hz. These bumps disappeared when the setup has been moved back to the EDB (plot1).
On the same plot, one can see that another bump around 65Hz, which has always been present, disappeared after this morning intervention (?).
Looking at the coherence with the seismometer attached to the detection tower, there is no clear evidence but some indication that these bumps may be due to backscattered light from the B1 setup.
