Hugh, Sheila, Stefan, Elli, Kiwamu,
The mysterious IMC misalignment event happened again. We are speculating that there may be a loose optical component in the IMC refl path on the HAM2 in-chamber table.
Back in the last February, there was a strange misalignment event in IMC (see alog 10335) where the IMC reflection path seemed to have significantly moved for some reason. This happened again this morning. This time, it was associated with the HAM1 activity of taking the door off from the HAM1 chamber. It tripped the HAM2 ISI and HEPI. After untripping them, we noticed that the IMC reflection was misalgined so much that the beam was almost missing the reflection camera and WFS diodes. However, like the previous event, the IMC power build-up was still reasonably high (it was at 820 counts in MC2 trans which is usually 830 counts or so) when it was locked. Sheila and Hugh then restored the HEPI back to the previous positions using the IPSs as a reference. At this point we could already see a reasonable DRMI flash at the dark port and also the IM4 trans seemed to have come back to the previous position. Based on these observations, we determined that the misalignment was only in the IMC refl path which is exactly the same conclusion as the previous event. So we realigned the refl path on IOT2L. Now IMC is locking fine and ASC loops were engaged without an issue.
In addition to those recovery activities, we did a brief test where we steered the HEPI by a big amount (~200 urads or so) in each rotational degree of freedom to see if we can reproduce such a big misalignment in the IMC refl, but we could not move the spot on the camera back to the center. Also, we quickly checked the mirrors in the IMC refl path on IOT2L to see if there is a loose component or something easily movable, but we could not find any. The misalignment was mainly in the horizontal direction on the table. The spot position on all three diodes (REFL, WFS_A and WFS_B) had almost the same amount of displacements horizontally, by half a cm toward the East. The attached is 3 hours trend of various sensors. At one point, both the WFSs got misaligned and at the same time the MC1 witness sensors saw a jump in both pitch and yaw (which I think is me untripping the HEPI and ISI).
Just to be thorough, I had also taken down the entire corner station HEPI to address a fluid leak. This took down all HEPIs and most ISIs--maybe the timing here was poor.
Also, "...realigned to the IPS...," we returned the Cartesian positions back to their pre break positions, not the individual IPS values persay. Still these moves were a few 10s of urads (we ignored the several um of translational shifts.)
Also as a reminder, HEPI only keeps the RZ dof, all others just servo to the free hanging position before the current isolation. Given the lack of impact of moving the HEPI 100s urads, letting it wander about a few 10s of urads each time it breaks isolation may be reasonable.