Fig 1 shows the SNR of the lines retrieved in B1, most points in the critical region from 10-70 Hz should be ok, at lower and higher frequencies the lines are not always perfectly visible (which would lead to an overestimate of the projection). Fig 2 shows the obtained transfer functions, multiplied by f^2 for clarity. This removes the effect of the pendulum, so the frequency dependence should be due to shielding effects. Fig 3 shows the total noise for the different buildings and fig 4 shows the resulting magnetic noise projection for the different positions of the coil. Data in quiet condition is taken from the 13th of September, when the horizon was at 9.5 Mpc.
The projected noise from the central building is dominating, both because the TFs for the CB are higher, and since the background magnetic noise is higher than in the terminal buildings. Two of the curves are clearly too high, they even project noise above the current sensitivity. Fig 5, finally, shows the projection in comparison with the design curves. The highest curve was choosen for the end-buildings, while for the CB the highest curve that does not project noise above the sensitivity was choosen.
Possible sources of error:
1) To calculate the transfer function, I divide the line amplitude in B1 by the vector sum of the amplitude in the 3 magnetometers. To do the projection, I again take a vector sum of the 3 magnetometer signals and multiply it by the TF. Note that the direction of the magnetic field for the two vector sums does not have to be the same. It could thus be that the environmental magnetic field is pointed along a direction for which the mirrors are less sensitive. This could thus lead to an overestimate of the projection.
2) Near-field vs far-field: this projection works under the assumption that both the magnetometers and the mirrors see the same field, which is true for sources in the far-field. If a source is very close to the mirrors, the projected noise will be underestimated. If a source is really close to the magnetometers, but further away from the mirrors, this will cause an overestimate.
3) Especially for the end-buildings, some of the magnetometers are close to saturation. If this happens during the injection of the lines, this will again cause an overestimate of the calculated TF. Another problem caused by this saturation is that the magnetometers are positioned along the wall instead of next to the tower. This violates a bit the assumption that the tower and the magnetometers see the same field. We should consider making them less sensitive, as was already done for the ones in the central building.
