I would not normally put this here, but the stochastic log was unavailable.

To build our confidence in our understanding of instrumental H1-H2 correlations, we are attempting to identify the features in the H1-H2 coherence. I did this for part of the S5 data back during S5 (here), when I was able to trace many of the features to particular sets of fans, or to low frequency (0-15 Hz) BSC and HAM resonances in the bilinear coupling regions around 60 Hz peaks; the summary from 2006 is repeated here as Figure 1.

Here we check these earlier results with full-S5 data, making sure that the rest of the run is consistent with the earlier partial results by comparing the full-run and partial-run features in coherence and by comparing the full-run PEM spectra to the full-run coherence.

A comparison of Figure 1 and Figure 2 indicates that the clusters of peaks from certain fans and the bilinear coupling features are present in both the partial run and the full run, suggesting that the specific partial-run identifications could be extended to the full run.

We expect that peaks in the H1-H2 coherence are associated with peaks in important PEM channels (sensors at coupling sites or half way between H1 and H2 coupling sites). The full-run PEM spectra of important channels show peaks that correspond with the coherence features, except for the peak at 114 Hz. The 114 Hz peak in Figure 1 is the only peak labeled “uncertain” – there were fans in this frequency band, but the shape of the peak did not quite match the shape in the partial-run spectra. This is again true for the full-run spectra, so I am not confident that we understand the source of the 114 Hz peak in coherence.

We are currently looking into the possibility that a sudden change in coupling when I floated ISCT4 to reduce H1 H2 coupling might be responsible for the inconsistency in coherence and ASDs.