Summary: 1-10 Hz ground motion was increased by a factor of about 10 by a truck and this resulted in a scattering shelf that reached about 800 Hz. The motion frequency that produced the shelf was about 1.05 Hz, near angular resonances for the output Faraday isolator. A possible mechanism for producing such a high frequency shelf is angular motion of an un-dumped beam on distant surface relief.
When the front loader was picked up, the loaded truck made scattering noise in DARM. The spectrum, during part of this period, Figure 1, shows two interesting features, first, that a scattering shelf reached up to at least 700 Hz, very high, and, second, that there was only a single shelf, not a series of stepped shelves, as would be expected from multiple reflections, which is one potential mechanism for producing shelves that reach such high frequencies (see examples from our recent injections: https://alog.ligo-la.caltech.edu/aLOG/index.php?callRep=31898 ).
Figure 2 shows spectrograms of the seismic spectrum and the scattering in DARM during this period. The seismic spectrum increased mainly in the 1-10 Hz region. Figure 2 shows that the scattering arches actually reached higher than 800 Hz and that their spacing was close to 2.1 arches per second. Since there are two velocity maxima in every cycle, the frequency of motion producing them was about 1.05 Hz. I looked for correlated motion at 1.05 Hz in all of our monitored optics and did not find it. However, the unmonitored output Faraday isolator has pitch and yaw resonances at about 1.05 Hz (https://dcc.ligo.org/DocDB/0124/E1600080/001/E1600080_OFI_Test_Report_LLO.pdf) and is thus a possible culprit. Figure 1, along with PRCL and SRCL spectra support the Faraday isolator possibility because there was very little difference at 1 Hz in these spectra, suggesting that the 1Hz motion did not affect these cavities.
The fringe crossing frequency reached over 800 Hz, and, for a 1 Hz motion, this requires an optical path length variation of about 65 um, huge compared to expectations of any of our 1 Hz motions. While the Faraday isolator is unmonitored, the HAM5 table motion during this period was less than a nm in the 1 Hz band, so it is unlikely that the Faraday isolator would be moving 65um at 1 Hz and it is unlikely that we would have any reflectors in our system that would move this much at 1 Hz. Thus it is likely that the scattering path length variation is somehow amplified. One way is by multiple reflections. But for multiple reflections, I would expect a series of shelves with the higher frequency shelves lower in amplitude than the lower frequency ones, both because the losses in the scattering beam should be greater for more reflections, and also because the power is spread out over a greater frequency band. For example, there should also be a shelf at 400 Hz if the 800 Hz shelf was produced by two passes (and the 400 Hz shelf would be larger by the square root of the inverse of the reflection ratio and by the sqrt(2) for the smaller bandwidth).
A second possibility is that the optical path length variation was dominated by motion of the scattered beam across the reflecting surface. One mechanism to convert motion across a reflector into path length variation would be if the reflector were tilted relative to the beam (and it was not specular). Another mechanism could be surface relief. If the scattering noise were produced by, for example, an un-dumped beam from the Faraday isolator, then for an angular motion of the isolator of 100 nR, the scattered beam would move 10 um at 100 m. If the scattering surface were at 45 degrees to the beam, the reflection at one extreme of the angular beam motion would be 10 um further away than at the other extreme.
The “filled in” nature of the scattering arches suggests a range of optical path length variations below some maximum. This seems consistent with the scanning hypothesis. This would involve path length variation associated with the reflecting surface, rather than path length variation associated with distance from the rotation axis of the scattering optic, which is proposed as a mechanism for “filled in” arches in angular injections here:
https://alog.ligo-la.caltech.edu/aLOG/index.php?callRep=31898 .
I think it would be worth making angular injections at LHO HAM5 to study this more (Anamaria and I tried this Friday at LLO but didn’t find scattering - link just above). The potential scattering from the output Faraday isolator may possibly be a problem normally, because the ground motion didn’t increase by much more than a factor of 10 and so the normal shelf may reach 70 Hz or so.
