Jenne, TJ, Sheila, Naoki,
We looked at some trends together of what's been happening with H1's range, PRG, and various sensors related to the PRC over the last few months.
The first screenshot here shows some of these trends, with the first horizontal cursor showing, Oct 24th a day when we had a particularly good BNS range just above 160 Mpc. The vertical cursors are set to show a reference of what the various signals were on that good day. Some things to notice:
- The PR3 yaw moves from Friday (74535) and Monday (74578) seem to have restored our PRG to nearly 50, with a similar stability to what we had in late October. These moves improved our range and glitch rate, but the range is still lower than our best times in October and early November.
- The top mass osems aren't a reliable way to track PR3 yaw (neither is the optical lever): We had previously seen that PR3's optical lever wasn't tracking long term drifts well at all, (70008) so on Nov 7th we used the top mass osems to reset PR3 yaw to previous times (74065), as marked by the red arrow. You can see that from this time forward the PRG dropped, became more noisy, and seems to drift more. So it seems that this move started to introduce clipping probably in the LSC POP path.
- The POP QPDs also can't be used to track this shift: POP A is the QPD used in DRMI ASC, which has large offsets applied. Both POP A and POP B yaw would indicate that our recent moves Friday and Monday went in the wrong direction, but we observe that these moves reduced clipping in the POP LSC path (74586 and 74582). Also, looking at their trends over the last few months they don't seem related to changes in PRG, (for example, POP A YAW was the same as at the good time in our worst PRG times 14-7 days ago).
- Jenne suggested looking at trends of IM1-3, which are uncontrolled and could be drifting. The second attachment shows their trends, they seem to be drfiting in a way that might correlate with outdoor temp, like many things do. I believe these are fairly small shifts of the IMs (I'm not sure that these are calibrated into urad, for historical reasons some channels which are calibrated on other suspensions aren't calibrated in the same way on the IMs). I'm not sure that we can tell without taking some steps on some of these if they are responsible for the PRG change.
I do not believe we have any sensors that are more stable than just relying on the mechanical stability of the suspensions themselves. In the past, all our attempts to stabilize the uncontrolled dof in the IMC have produced worse results than relying on the stability of its suspensions.
Naoki, Sheila
We looked at some spectra comparisons. In both attached screenshots the references are the 160+ mpc time from Oct 24th. The first one shows the good time from Oct 24th compared to a time when the DARM noise was high due to clipping from Nov 29th. This shows features that have been aloged else where, the yaw ASC loops have excess noise 74502, the coherence of LSC and ASC POP is high and the ASC pop signals have extra noise (74580). There isn't as dramatic a difference in the LSC signals, which are measured in the path that was clipped, which is interesting.
The second attachment is a comparison to after our recent PR3 moves, motived by wondering if some small amount of remaining clipping could explain why our sensitivity isn't as good as it was in October. The noise of the POP sensors and the coherence of LSC and ASC POP are similar now to the time of our best range, so there isn't any obvious indication here that we are still clipping.
