Lost lock due to DSOFT instability at about 1.6~1.7Hz in LOWNOISE_ASC three or four times in a row (attached). It's very apparent in AS port video.
There's the following line in ISC_LOCK to change the gain from 3 to 1 during the process,
ezca['ASC-DSOFT_Y_GAIN'] = 1 # Sep17 2016 we had loop instability, so decreased.
I bypassed this and it still started oscillation, wasn't able to manually decrease the number in time to rescue. Currently the script is changed to
ezca['ASC-DSOFT_Y_GAIN'] = 0.7
but it's yet to be seen if this is of any help. I need to go now, giving IFO to Sheila.
I also got a lot of lock losses in FIND_IR, which is annoying. Sheila says that that was the case yesterday.
After Keita left I was able to lock twice before the ALS problems made locking impractical.
In the first lock, I tried the soft offsets that Vaishali and I found last night at 2W input power. I was able to get the recycling gain which had been about 31 at 2W to nearly 33 using an offset of 0.04 in both DSOFT and CSOFT yaw, with the offsets that Vaishali and I found the recycling gain was down to about 29. As I powered up with these offsets in, the recycling gain stayed high and at 27W we had a PR gain of nearly 29. In the second lock I moved these offsets into the TMS QPD offsets, and have saved these in SDF. I also made a simple script for doing this, which is attached.
After powering up I started executing the LOWNOISE_ASC state by hand, and started with the HARD loops because Keita was suspicous of a problem with DSOFT. switching the HARD loops to their low noise configuration caused an instability that looked like DHARD Y which unlocked the IFO. The next lock I lost after sitting at DC readout for about a half hour, as soon as I reduced the RF45 Modulation depth there was an ASC instability (PIT this time).
