[Sheila, Jenne]
We did a first quick test of moving the spot on PR2 during our commssioning time this morning. Our biggest conclusion is that 'wow, we need to move a lot if we're going to center on PR2'. Since it was late in the commissioning window, and I had already broken lock once today when shaking HAM3 (alog 77851), we didn't shake HAM3 while doing these moves. That's something that we'd like to do at some point, to hopefully see that the effect in DARM is lower when the spot on PR2 is closer to center.
Sheila manual-ed the ISC_LOCK guardian to the PR2_spot_move state. This state watches how much PR3's slider moves, and moves the PRM and PR2 sliders to compensate that move. In the first attachment of the slider values for the 3 PRC optics, you can see that PRM and PR2 sliders are following along when Sheila moves the PR3 slider. We didn't see very much response in the ASC loops when we made a step, which means that these 'following along' scales for the PRM and PR2 sliders are still quite good, even though we haven't used this guardian state in several years. First we went a small amount one way, then we went a larger amount the other direction in yaw.
The second attachment is the top mass OSEM witness of the moves, and the third attachment is our cabity buildups. The x-axes of all these plots / screenshots are set to be the same. The bigger yaw movement (almost 3.6 urad in PR3 yaw) a little after -1h clearly shows a bump up in arm circulating power, that then goes away when PR3 is moved back to its nominal position. This bump is clear in the arm circulating powers, even though the IFO was still thermalizing. Interestingly, I don't see a similar bump in POP_LF, the amount of power circulating in the PRC.
While PR3 was at it's extremal yaw position for today, I did a rough Y2L optimization for PR2 with step sizes of 0.05, and found that a Y2L value of -7.40 was about the best I could do to minimize the peak (30 Hz, 1 ct using the ADS oscillator) in PRCL_OUT. After Sheila put PR3 back to its nominal position, I re-checked, and found a Y2L value of -7.50 was the best. This means that we moved the spot on PR2 by about 0.2 mm (out of the 15 mm we think we need to move) closer to center, when we changed the PR3 yaw alignment by 3.5 urad. I'll note that I did not have the 30 Hz notches in the quads' ISCINF pit and yaw filter banks engaged, but I'm not fully sure that those are helpful here (if they are important, we should add them to all of the other optics, not just the quads).
Just as a double check of whether we really should need to move PR3 that far: If I assume the distance between PR2 and PR3 is 1/3 of the PRC length 55m, my lever arm for a very naiive spot movement calculation is about 18m. At 18 meters away, a 3.5 urad angle change will move a beam 6.3e-5 m, or 0.06 mm. So, this very naiive estimate gives a smaller beam spot motion on PR2 for a given PR3 move than we measure using A2L, but it confirms that indeed we probably didn't make a very dramatic move on PR2 today.
