We are back to full lock at around 9:30 PT. There were a couple of issues in ASC as listed below.

• PRC2 Yaw loop on PR3 was not engaged at all since 31st of August  .
  • The gain of ISCINF filter have been zero in PR3.
• INP1 and PRC1 loops were not engaged in full lock since 24th of August  .
  • ISC_LOCK seemed to have been edited such that it zeroes the gain for them in ASC and does not ramp them back up in full lock.

I took over from Jenne, who stayed all the night, and kept relocking the interferometer. JeffB had a nice hand-written note of what guardian states the interferometer has unlocked. According his note, there were several lock losses in SWITCH_TO_QPD (see also Jenne's alog) and one lockloss in DRMI_ON_POP. I kind of suspected the TRX and TRY signals and therefore took a look at the spectra and dark offsets with IMC intentionally unlocked. They look healthy except for a small offset in TR_Y_B_QPD. They had an dark offset of about roughly 0.3 counts in terms of SUM  which should not be a problem in SWITCH_TO_QPD, but since I just noticed it, I adjusted the dark offsets. One note is that since these QPDs are sensitive not only to red light but also green light, I intentionally kept green light resonant in both arms in order to obtain proper dark offsets.

I did not lose lock in SWITCH_TO_QPD at all for some reason. I am not sure if this is because of the improved dark offsets in TRY or not. I then stopped at CARM_5PM for just checking out how high the bounce and roll modes are. Initially the bounce modes were as high as 10^-12 m/sqrtHz in the darm spectrum with 0.1 Hz BW. I gave several minutes to the interferometer in order to damp them with some addition of extra gain in the DAMP_V filter in all four quad suspensions. At this point, PR3 tilted more than it usually does presumably due to some thermal load on the wires, I manually touched PR2 to counteract it. Without this touch, I think it would not stay locked for more than a minute in CARM_5PM. This maybe an indication of a different spot position on PR3. Though, no idea how different.

Reaching DC_READOUT, I found the interferometer to be more wobbly than it usually is. In particular, POP_LF was clearly modulated by PRM longitudinal actuation. This lead me to look further into the ASC loops, especially PRC and INP loops. As listed above, PRC2 Y, INP1 P/Y and PRC1 P/Y were found to be not engaged in full lock. Sheila looked into ISC_LOCK and removed the sentences which set the ASC gain to zero for INP1 and PRC1 so that by the time when the interferomteer engages full ASC it will engage INP1 and PRC1 as well. Since we did not lose lock afterwords, ISC_LOCK has not been tested. We set the ISCINF_Y  gain of PR3 to 10 according to trend of the past 10 days or so. So far the interferometer stays locked with ASC fully engaged without an obvious or big issue.