On April 13, Dan ran some modulation depth up-down tests at 2W, 25W before MOVE_SPOTS, and 25W after MOVE_SPOTS.
The analysis is only half-done, but in the interest of getting some answer out there quickly I'll post the output of mod_depth_analysis.py.
This is in conjunction with the April 10 sideband PRGs alog 68628.
Results
| 2 W ENGAGE_ASC_FOR_FULL_IFO | 25 W before MOVE_SPOTS (was not thermalized, can not trust) |
25 W after MOVE_SPOTS | 78 W NOMINAL_LOW_NOISE | |
|---|---|---|---|---|
| 9 MHz PRG | 90.7 | 80.9 | 56.6 | |
| 45 MHz PRG | 23.2 | 24.7 | 22.7 | |
| Carrier PRG | 53.6 | 46.3 | 43.1 | |
| Total PRG | 52.8 | 45.9 | 43.0 | |
| POP path efficiency | 1.068e-5 W/W | 1.055e-5 W/W | 1.054e-5 W/W | |
| 9 MHz REFL ratio | 0.473 | 0.372 | 0.196 | |
| 45 MHz REFL ratio | 0.164 | 0.176 | 0.366 | |
| Carrier REFL ratio | 0.064 | 0.029 | 0.060 | |
| REFL path efficiency | 1.68e-3 W/W | 1.68e-3 W/W | 1.68e-3 W/W | 1.68e-3 W/W |
Method
I had to ignore the 25 W test before MOVE_SPOTS because thermalization was too strong (See PDF 2). We can fit out the thermal transient later on.
Because the LSC POP path pointing is not controlled, I adjusted the PRG results to account for changing POP path efficiency so the total PRG and input power numbers make sense (only a 1% correction). The most significant change is after MOVE_SPOTS.
The REFL path is controlled and therefore the efficiency was not adjusted.
Discussion
As expected, the 9 MHz PRG at low power is much higher than at high power (48% reduction).
This is extreme, far stronger effect than for carrier.
This explains why the PRCL gain is falling as we power up (the exact same situation as O3, just at significantly higher power).
This also supports the idea that ASC using 9 MHz at the REFL port will suffer at higher powers (again, the exact same situation as O3).
Interestingly, the 45 MHz PRG does not really change. It stays around 23 ish, which Koji says might be a little higher than we would naively expect, but it's maybe not super crazy that this is high given SRCL detuning.
The carrier REFL ratio does not seem to follow the same expected progression as the PRG.
The carrier PRG falls consistently as we power up.
If REFL was following IFO losses accurately, then REFL should also fall consistently as we power up. This is not what we observe.
The carrier REFL ratio starts at 6.4% at 2 W, falls to 3% at 25 W, then goes back up to 6%. There are clearly other forces at play for REFL.
This gets brought up time and again, but it really seems like MOVE_SPOTS moves through a better spot than we end up settling on.
If time was not so essential, we could try sitting at MOVE_SPOTS, move around on the TMs, move the mod depths, and optimize all the PRGs.
