Sheila, Gabriele, Camilla, Jennie
The other day I used Elenna's template to make a PRCL excitation: 77252
We can use this to check on our PRCL to SRCL feedforward element in the LSC input matrix, where we add some of the POP A 9I signal (PRCL) to the POP A 45I signal (SRCL) to cancel the POP45 sensitivity to PRCL.
When the PRCL excitation is on, the transfer function of POP45I/POP9I is -0.028716 cnts/cnt
We want to set out input matrix such that:
5_1 = -5_3 * POP41/POP9I tf (with PRCL excitation on) to cancel the PRCL contribution to the SRCL error signal. This means that now, our POP 9I to SRCL ERR matrix element should be 0.107 but it is now 0.06.
EDITING to Add:
Today I re did this measurement and saw that POP45/POP9I is -0.02551 cnts/cnt, meaning that the input matrix should be 0.0946. Updating the input matrix to this did improve the PRCL to SRCL decoupling, as shown in the second attachment.
I also tried moving the PRCL offset, this did not seem to change the POP45I/POP9I ratio much, so the PRCL to SRCL decoupling doesn't seem to need retuning each time the PRCL offset changes.
The third screenshot shows the PRCL to DARM coupling, there are small changes (2-3dB change) with these changes in PRCL offset and PRCL to SRCL decoupling. So this suggests that the main coupling of PRCL to DARM isn't through SRCL. (And that the PRCL offset sn't helping much).
Gabriele remembered that the PRCL to DARM coupling could also be through MICH, indeed the third screenshot shows that the PRCL signal is a factor of 10 smaller in POP45 Q than in I. In 66151 Anamaria tuned the phasing so that the PRCL signal in 45Q was a factor of 100 smaller than in I, so we can try to improve this by rephasing POP45.
Edited again:
Camilla and I moved the POP45 phase from 8 to 11, this reduced the PRCL to POP45Q transfer function to be about a factor of 100 below PRCL to POP45I (last screenshot blue references show the improvement from a phase of 8 to 11). We then remeasured the SRCL to PRCL input matrix element, which should be 0.087. This made the SRCL coherence in DARM rather bad, although the PRCL to MICH and PRCL to SRCL couplings were improved. We decided to revert these, and wait for next week to make the changes at a time when we could immediately retune FF. As I stepped the phase back I accidentaly moved us from 11 to 6 in one step, which unlocked the IFO.
Getting around to finishing the Locking Survey for overnight. Although h1 was down 4hrs, looks like no phone calls for help---H1 battled on during the latter half of last night's wind storm getting fairly far in ISC LOCK (3x to Transition etmx & 2x to carm150pm). And then once the winds got down to 10mph this is when h1 had a chance.
(All in all, H1 was trying for hours on its own and didn't get stuck anywhere through the wind storm [maybe if H1 Manager tried an Alignment it would have helped, but really once the winds died, H1 automatically made it through without an alignment. It was nice to have the Verbal log history go back to follow the story and complete the Locking Survey entry and see how the night went.)
Microseism: also forgot to mention our low microseism has slowly drifted UP over the last 15hrs (from below the 50th percentile line to above it.
I should mention that I did receive a call last night at 02:00 PDT from H1_MANAGER requiring assistance. It looked like H1 was having trouble locking with two locklosses at TRANSITION_FROM_ETMX (very likely due to high winds). As the winds had died down a bit by that point, I decided to let H1 keep trying, as ALS and DRMI were consistently locking easily. On the next lock acquisition attempt, H1 made it to low noise without issue.
So, while H1 did call for assistance overnight due to high winds causing locking difficulties, I did not need to intervene for H1 to eventually reach NLN (and I don't expect doing so would have sped up the process).