Matthew Ball, Adrian Helmling-Cornell, David Shoemaker, Cheryl Vorvick, Sharan Banagiri, Robert Schofield
We found that the 48 Hz peak in DARM was associated with the HAM3 chamber doors using a new technique to localize scattering sites on vacuum enclosure walls. We injected from two shaker locations at slightly different frequencies (e.g. 48 and 48.01). Because the injection locations are separated, the relative phase of the two signals varies with location on the vacuum enclosure. As a result, the phase of the beat envelope varies with position, and different sites experience maximum chamber wall motion at different times. The sites with accelerometer signals that have the same beat envelope phase as DARM are candidates for the scattering sites on the vacuum enclosure walls. We get reasonable resolution in part because propagation velocities on the steel membranes are only hundreds of m/s. The technique is illustrated in Figure 1 and there is a bit more detail in my Warsaw talk: https://dcc.ligo.org/LIGO-G1901683 . We first used a variant of this technique, injecting with two separated speakers, which, because it reaches everywhere in the LVEA, may be the best first step (BTW we could clearly hear the spatial variation in beat envelope). The beating shakers seemed better for finer-resolution localization.
We found that there were often a couple of locations with envelope phases indistinguishable from DARM’s, but when we injected at multiple frequencies, we could eliminate the chance phase alignments (e.g. instead of just 48 and 48.01, we also used 50 and 50.01 and 47 and 47.01 as well as beating frequency sweeps).
In the first set of injections, the best match in the LVEA was to the single accelerometer on HAM3, so we began focusing in that area. Previous studies using single shakers and impulses suggested coupling in the vertex area, but had not narrowed the location specifically to the HAM3 area. In order to increase the resolution of our accelerometer array, we mounted temporary accelerometers at various locations on HAM3, the HAM3-BSC2 spool and HAM3 side of the MC tube. Since we only have two spare accelerometers, we had to inject and move accelerometers multiple times. We also checked the GS13s on the table, H1,2,3 and V1,2,3, but their phases did not match DARM, consistent with other evidence pointing to the chamber walls. The accelerometer locations that had the best beat-envelope match to DARM were on central parts of the HAM3 doors, but not the edge.
We opened the viewports and found a lot of scattered light on the inside of the doors, including two “beams” that came right through two different viewports, so we were able to block them with black glass. It turned out that the “beam” that was mainly responsible for the 48 Hz peak was the beam coming through the illuminator port. We put black glass in this port (lower –X port on the –Y side of HAM3) and found that this eliminated the 48 Hz peak. I could easily see a misshapen “beam” on an IR card so I think it might be a milliwatt-scale “beam”. It was surprising that this worked so well since light was scattered all over the doors, but it may be that the illuminator port was the most important location because it has a flat bulb cover that was oriented in such a way that it may have retro-reflected scattered light back to the source (Figure 2).
The source of the “beam” of scattered light (and also the likely recombination point) appears to be the +Y edge of the PR2 scraper baffle aperture (see photos and diagrams in Figure 2).
Figure 3 shows a comparison of summary-page spectra for the day before and the day after we inserted the black glass (https://alog.ligo-wa.caltech.edu/aLOG/index.php?callRep=52136 ). The 48 Hz peak seems completely eliminated and has not re-appeared in a few days. So, fixing the scattered light source in HAM3 does not now seem to be an immediate priority. However, if we arent going to fix it during O3, we would like to make injections to estimate the remaining scattering noise since light is scattered all over the door. Finally, it is possible that there is a second problem on the table that produces the other “beam” that comes out through the lower +X port on the +Y side of HAM3, though Cheryl thinks it might also be from the PR3 scraper baffle. We also put black glass in this port.
