Today the vacuum team implemented a seismic isolation of the dry pump placed at about 10m into the NE tunnel. One steel plate and one layer of sorbothane rubber have been added. Figures 1 and 2 are pictures of the installation.

Figures 3 and 4 show the noise reduction sensed by some seismic sensors in the NEB hall: purple is before the mitigation (pump sitting directly onto the parralepipid support) and blue is after.

The peak at 29.0 Hz from this pump reduces significantly in all sensors. Microphones do not sense and did not sense any peak at 29.0 Hz. The sensor named ENV_NE_DRY_PUMP_ACC is a temporary meggit accelerometer which was placed onto the building floor next to another dry pump located just behind the right tunnel door. The residual seismic peak at 29.15 Hz present in the NEB hall sensors  is produced by this pump.

Today we also investigated a bit the noise of the NEB dry pump responsible of the 29.15 Hz. This pump is located right behind the tunnel separation wall. It sits onto the building floor, and one sorbothane layer which provides already a seismic isolation towards the floor. The other possible paths are the pipe that connects the pump to the NE tower and acoustic emissions.

We could exclude acoustic emissions: we moved the microphone NN_NEB_INF_02 from its assigned location in the hall SW corner close to the separating wall to the other side of the wall behind the pump. There is no evidence of a peak at the pump frequency (Figures 1 and 2).

We also did some test measurements with one accelerometer. The connection pipe (the straight rigid section attached to the floor) vibrates significantly at 29.15 Hz (about 0.1 ms-2/sqrtHz). We placed some heavy mattress onto it and managed to damp it (the pipe vibration reduced to 0.01 ms-2/sqrtHz) yet we did not observe a reduction at the guralp sensor.

Some mitigation attempts will be performed together with the vacum team, possibly next Tuesday.

This Tuesday Riccardo performed two mitigation actions on the NEB "step" dry pump: (1) added a second layer of soft rubber underneath the pump (Figure 1 - before, Figure 2 - after); (2) added a softer section of corrugated pipe (Figure 3). The overall effect was to reduce the seismic line at 29.15 Hz in the NEB hall as recorded by various sensors (Figure 4 - VIM spectrogram, Figure 5 - guralp and accelerometers ASD, Figure 6 - NN array). The line amplitude reduced by about a factor 2 to 3, depending on the sensor. It remains to understand which of the two actions produced the observed effect. This will be done next Tuesday.

Yesterday morning, we performed some investigations to understand which actions between the softer below and the second stage of seismic isolation (steel plate + 4 pieces of Sorbothane rubber) produced the observed reduction of the amplitude of the seismic pump line at 29.15 Hz. The logfile is attached and the analysis will follow.

The spectrogram in Figure 1 illustrates the effect of each single action on the amplitude of the 29Hz line. The sensors in Figure 1 are the vertical guralp, and the accelerometer we placed on the floor next to the pump.  Figure 2 compares in some seismic sensors the amplitude of the 29Hz line during four configurations of the pump: 1) two seismic isolation layers underneath the pump and softer bellow pipe (purple); 2) one layer of seismic isolation and softer corrugated pipe (orange); 3) one layer of seismic isolation and more rigid pipe (green); 4) two layers of seismic isolation and more rigid pipe (blue).

The larger effect (a factor two at most) seems associated to the second layer of seismic isolation (removal: purple to orange) and re-insertion (green to blue). However the softer corrugated pipe seems to have some reduction effect in the accelerometers placed on the NE tower flange and at the CryoTrap (orange to green).

Figure 3 shows the signal of the accelerometer placed onto the straight section of the pipe which connects the dry pump to the NE tower turbo pump. The insertion of the softer corrugated pipe clearly reduces the pipe vibration. Figure 4 shows the reproducibility  of the isolation actions by comparing October 4 and October 5, some 30% change in the line amplutude and a slight shift of the line occurs.

We noticed that the accelerometer on the tower flange (ENV_NE_ACC_Z) from time to time senses noise glitches, to be investigated.