Fig 2 shows a zoom of one glitch in the full-bandwidth signals of the suspensions. They appear to be too fast to be real, so they might be caused by some electrical pickup, as we have seen before.
The glitches are also seen in the magnetometers, UPS and IPS sniffers, so they clearly have some electrical origin. They can however also be seen pretty well in the spectrogram of the microphone and slightly less in the seismometer on the flange, hinting on some mechanical component, see fig 3. Listening to the microphone it appears as if someone is hammering 3x per second on the tube, while the glitches in the magnetometer and top-stage appear at half that frequency.
Fig 4 shows a trend of the gliches, the first one occured this night, then another 2 in the morning. In the afternoon they happened every 10 or 20 minutes, making locking impossible. At the moment, they appear to have been gone.
No correlation between the glitches and any signals that monitor the infrastructure has been found, but note that Davide did some work on the chillers of the NE around the time they disappeared. On the vacuum side, the only thing abnormal is that the turbo-pump is switched off since Sunday. The vacuum people had a look at the NE, but did not find anything strange.
