Overview

Still in the mid way; just for memo.

In the same manner as of 24634 and 24636, the 117 Hz bump in DARM is analyzed in the several silent runs on Apr. 5-6 JST. Recently, there are less coherence between DARM and IMMT1 trans pitch, although there are still a certain amount of bump in the DARM. The averaged bump height (both in DARM and IMMT1 trans) also seeminlgy reduced recently. Anyway, so far, it seems the bumps in DARM (or coherence) are related with the interferometer alignment; the better the alignment the less the coherence at the bumps.

Details

In Fig. 1, DARM and IMMT1 trans spectra are shown with solid and dashed curves, respectively, in the following color assignment. The coherence between them are also shown.

The case (1) would support the theory "the better the alignment the less the coherence at the bumps"; maybe offloading IMMT1 (24687) might keep good initial alignment for the cases (2) and (3), so although there are still remaining especially in (2), the coherence would be also lower than those in the past. What we learn would be

• Averaged bump height, especially in IMMT1 trans, seems reduced recently (see Fig. 1 of 24634 for comparison). In fact, in all of (1), (2), and (3), IMMT1 trans are mostly the same, and lower than the past (interestingly 127.3 Hz peak in IMMT1 trans are so stable in this and the past posts.).
• In these spectra, what we may need to focus on would be the reduction of the bottom of the noise floor; for (1), although the calibration factor (or optical gain maybe) was almost the same, the bottom of the noise floor seems lower than (2) and (3), regardless of the bump coherence. This reduction of the noise floor is, in the frequency range of 40-200 Hz, drastically affecting the BNS range improvement, so it is important. This would suggest that better alignment would resolve most of the things; it may help improvement of the bump(s) in IMMT1 trans, and may also reducing the coupling paths to the DARM of this bump (may not necessarily of real pitch tilt), and the noise floor reduction...
• But still, if we can erase these bumps (as well as the peak at 127.3 Hz, which is so stable...) in DARM completely, the BNS range would drastically improve.

By the way, Figs. 2 and 3 show the timing of PRMI locking for the case of (2) and (3), respectively; for (1), due to the ASC working process, the PRMI part had been far past so omitted here. What we can learn from these figures may be:

• Regardless of that the final DARM would have slight coherence or no coherence, a shock to IMMT1 trans pitch would be observed somehow.
• Maybe the "similar" or "relevant" shock to MCL would be just suprious, as there seems shock in Fig. 3 but the resultant coherence was, as reported in the above, almost zero.

Then, what does the reduction of the noise floor described above suggest? Now it would be better to start investigation more regarding electrical coupling as "confounding" paths among the several relevant channels around REFL.