J. Kissel

I've been looking through the data captured about the PMC in the context of the two years of observatory use between July 2022 and July 2024 where we spanned a few "construction, commission, observe" cycles -- see LHO:83020. Remember the end goal is to answer the following question as quantitatively as possible: "does the PMC have a high enough duty cycle in the construction and commissioning phases that the SPI does *not* need to buy an independent laser?"

Conclusions: 
 - Since the marked change in duty cycle after the PMC lock-loss event on 2023-05-16, the duty-cycle of the PMC has been exceptionally high, either 91% during install/commissioning times or 99% during observing times. 
 - Most of the down time is from infrequent planned maintenance. 
 - Recovery time is *very* quick, unless the site loses power or hardware fails. 
 - The PMC does NOT lose lock when the IFO loses lock. 
 - The PMC does NOT lose lock just because we're vented and/or the IMC is unlocked. 
 - To-date, there are no plans to make any major changes to the PSL during the first one or two O4 to O5 breaks.
So, we shouldn't expect to lose the SPI seed light frequently, or even really at all, during the SPI install or during commissioning. And especially not during observing. 

This argues that we should NOT need an independent laser from an "will there even be light?" "won't IFO construction / commissioning mean that we'll be losing light all the time?"  duty-cycle stand point.
Only the pathfinder itself, when fully functional with the IFO, will tell us whether we need the independent laser from a "consistent noise performance" stand-point.

Data and Historical Review
To refresh your memory, the major milestones that happened between 2022 and 2024 (derived from a two year look through all aLOGs with the H1 PSL task):
- By Mar 2022, the PSL team had completed the complete table revamp to install the 2x NeoLase high-power amps, and addressed all the down-stream adaptations.

- 2022-07-01 (Fri): The data set study starts.
- 2022-09-06 (Tue): IO/ISC EOM mount updated, LHO:64882
- 2022-11-08 (Tue): Full IFO Commissioning Resumes after Sep 2022 to Dec 2022 vent to make FDS Filter Cavity Functional (see E2000005, "A+ FC By Week" tab)
- 2023-03-02 (Thu): NPRO fails, LHO:67721
- 2023-03-06 (Mon): NPRO and PSL function recovered LHO:67790
- 2023-04-11 (Tue): PSL Beckhoff Updates LHO:68586
- 2023-05-02 (Tue): ISS AOM realignment LHO:69259
- 2023-05-04 (Thu): ISS Second Loop Guardian fix LHO:69334
- 2023-05-09 (Tue): "Something weird happened to the PMC, then it fixed itself" LHO:69447
- 2023-05-16 (Tue): Marked change in PSL PMC duty-cycle, nothing specific PSL team did with the PMC, but the DC power supplies for the RF & ISC racks we replaced, 69631, while Jason tuned up the FSS path LHO:69637 
- 2023-05-24 : O4, and what we'll later call O4A, starts, we run with 75W requested power from the PSL.
- 2023-06-02 (Fri): PSL ISS AA chassis it was replaced, but PMC stays locked through it LHO:70089
- 2023-06-12 (Sun): PMC PDH Locking PD needs threshold adjustment, LHO:70352, for "never found out why" reason FRS:28260
- 2023-06-19 (Mon): PMC PDH Locking PD needs another threshold adjustment, LHO:70586, added to FRS:28260, but again reasons never found.
- 2023-06-21 (Wed): Decision made to reduce requested power into the IFO to 60W LHO:70648
- 2023-07-12 (Wed): Laser Interlock System maintenance kills PSL LHO:71273
- 2023-07-18 (Tue): Routine PMC / FSS tuneup, with quick PMC recovery LHO:71474
- 2023-08-06 (Sun): Site-wide power glitch takes down PSL LHO:72000
- 2023-09-22 (Fri): Site-wide power gltich takes down PSL LHO:73045
- 2023-10-17 (Tue): Routine PMC / FSS tuneup, with quick PMC recovery LHO:73513
- 2023-10-31 (Tue): Jeff does a mode scan sweeping the PMC FSR LHO:73905
- 2023-11-21 (Tue): Routine PMC / FSS tuneup, with quick PMC recovery LHO:74346
- 2024-01-16 : O4A stops, 3 months, focused on HAM567, no PSL work (see E2000005, "Mid-O4 Break 1" tab)
O4A to O4B break lock losses: 7
       2024-01-17 (Wed): Mid-vent, no IFO, no reported cause.
       2024-01-20 (Sat): Mid-vent, no IFO, no reported cause.
       2024-02-02 (Fri): Mid-vent, no IFO, no reported cause.
       2024-02-08 (Thu): Mid-vent, no IFO, no reported cause. During HAM6 close out, may be related to alarm system
       2024-02-27 (Tue): PSL FSS and PMC On-table Alignment LHO:76002.
       2024-02-29 (Thu): PSL Rotation Stage Calibration LHO:76046.
       2024-04-02 (Tue): PSL Beckhoff Upgrade LHO:76879.
- 2024-04-10 : O4 resumes as O4B start
O4B to 2024-07-01 lock losses: 1
       2024-05-28 (Tue): PSL PMC REFL tune-up LHO:78093.
- 2024-07-01 (Mon): The data set study ends.

- 2024-07-02 (Tue): The PMC was swapped just *after* this data set, LHO:78813, LHO:78814

By the numbers

Duty Cycle (uptime in days / total time in days)
     start_to_O4Astart: 0.8053
    O4Astart_to_O4Aend: 0.9450
    O4Aend_to_O4Bstart: 0.9181
       O4Bstart_to_end: 0.9954
(Uptime in days is the sum on the values of H1:PSL-PMC_RELOCK_DAY just before lock losses [boxed in red] in the attached trend for the given time period)

Lock Losses (number of times "days" goes to zero)
     start_to_O4Astart: 80
    O4Astart_to_O4Aend: 22
    O4Aend_to_O4Bstart: 7
       O4Bstart_to_end: 1
(Number of lock losses is mere the count of red boxes for the given time period)

Lock Losses per calendar days
     start_to_O4Astart: 0.2442
    O4Astart_to_O4Aend: 0.0928
    O4Aend_to_O4Bstart: 0.0824
       O4Bstart_to_end: 0.0123
(In an effort to normalize the locklosses over the duration of the time period to give a more fair assessment.)

I also attach a histogram of lock durations for each duration, as another way to look at how the duty cycle dramatically changed around the start of O4A.