This is Hannah Fair.
I’ve been investigating the FF LSC locklosses from ER7. The following are my findings so far.
About half of the FF LSC locklosses had a 20-25 second oscillation in one of the PRM and SRM channels, and in the PRCL and SRCL channels (and occasionally MICH). The start of this oscillation ranged between around -2000 seconds to only -100 seconds before lockless. All of these ended with one of the channels saturating. This saturation appears to be the cause of loss. This pattern is independent of laser power.
-
1117665437.75 and 1117743020.38: The 20-25 second oscillation appears to be begin approximately 100 seconds before lockloss. In both cases, this ends with the saturation of PRM_UR
-
1117898344.25: The 20-25 sec oscillation begins around -85 seconds. SRM_M3_UL saturates
-
1117946874.25: The oscillation begins at least -500s. There also appears to be beat pattern. Each envelope is about 100s long. PRM_M3_UR saturates.
-
1118022123.88: The oscillation begins at least -500s. A beat pattern with an envelope about 200s long. ETMY_UL saturates.
-
1118088249.75: The oscillation appears to last for several hundred seconds, most evident in PRM and PRCL. Beat pattern has envelope of around 350 seconds. ETMY_UL saturates.
The following are locklosses where I also looked at OPLEV channels (Due to timing, I was unable to look at the data from the OPLEV channels for the locklosses before this):
-
1118207893.12: This was a short lock, so the ~25s oscillation seems to be present throughout. Looking through the OPLEVs, there was a very defined peak around -600s in BS_M3_PIT, ITMX/ITMY/ETMX/ETMY_L3_PIT. PR3_M3_PIT also has a continuous downward progression from lock to lockloss. ETMY_UL saturation causes lockloss.
-
1118213355.125: This is very similar to the previous lockloss. Short lock, about 1000s long. Same oscillation throughout. PR3_M3_PIT has continual decline, and very defined spike in same channels as before around -862 seconds. ETMY_UL saturation causes lockloss.
Some of these are less detailed/precise because of timing and my inability to access data more than two weeks old.
Examples of this pattern are attached as images.
