During the switch off tests of 20 July, the fans of turbo_PR , turbo_BS-PR, turbo_BS were identified as producing seismic noise on PR B4_ghost flange (we didn't make a distinction between them) so we installed a simple mechanical damping system for each of them (= rubber feet + 6-7 kg metallic flange). The effect is shown in plot 1, purple curve=original, blue curve = damped configuration, as seen by the accelerometer on B4_g flange.
During the maint. break of 25/07 , at around 10h UTC, we removed the shaker device from the B4_g flange, with the effect of changing the local mechanical vibrations of the flange (around 90Hz in particular, see plot 2 ).
Today morning, during the minitower intervention, we tried to add back a weight to the B4_g flange (about 20 kg) fixed in a provisional way (picture 3) from approximately 9h30 UTC to 9h50UTC. In attached some spectrograms . Finally we restored the previous situation = no added weight.