There is coherence between HAM2 and DARM, it seems possible that this is because of clipping on the ISS 2nd loop array, as solved in 46412. The attached plot shows this.
Since I don't know how to distinguish between clipping on the ISS path after transmitting through IM4 and clipping through the input Faraday without picoing the ISS path, I instead moved IM1 and IM2 a bit until these peaks went away and we were more centered on the ISS QPD (at least in pitch.... we might need to pico for yaw if this IM1, IM2 position is better overall for the IFO).
Since we lost lock, I took the opportunity to move things around. I had IMC-only, with 30W injected, and closed the ISS second loop by requesting ISS_ON of the IMC guardian. The blue spectrum is the RIN of the out of loop PD before I moved anything, and brown is after. The peaks around 400Hz are gone, as is a bit of noise at low frequency, and another peak around 575Hz. The ndscope shows the ISS QPD spot, and the IM1 and IM2 sliders.
Gavin and I are running through an initial alignment before trying to relock.
It occured to me that I ought to have used IM3 and IM4 to compensate for the IM1,IM2 move earlier, rather than using PR2 and IM4. So, since I'm not succeeding at acuiring lock (although as this earthquake rings down, I'm getting further each attempt), I undid PR2's moves from the earlier initial alignment and zeroed the INP error signals instead using IM3 and IM4 with Xarm IR locked. Now I'm finishing up the rest of initial alignment, and hopefully the earth will have quited down enough for locking.
Hmmm. No good. We can't get through the CARM offset reduction sequence with this alignment. We've had similar problems in the past, where if we move the IMs, we can't quite compensate and get back the exact same input alignment, and then can't get through CARM reduction. The reliable solution in the past has been to revert the IM alignment, which I have now done. We'll need to pico the ISS to get rid of the peaks.
I've moved picomotor PM1 on HAM2 to reduce the peaks in the ISS out of loop diodes. In the attached spectrum, blue is before the move (after reverting the IMs), and brown is after moving the pico. I didn't try moving PM5, since the January alog from Georgia indicated that PM1 is the one that was most effective last time. Centering the beam on the QPD did not reduce the peaks - notice in the ndscope top right that the yaw is (as was true after the IM move) now farther from center. But, this is what makes the peaks go away, so it's what we'll try.
Nope, we've lost lock at the same DARM_TO_RF area again. Georgia and folks are going to keep having a look.
