I attempted to restore the IFO to the PRFPMI DC lock state.
Currently, the IFO reaches the **PRFPMI RF lock state**, but both arm cavity normalized transmitted powers only reach 83 and 78, respectively (Fig. 1), even with all ASC loops engaged. Also, the normalized POP90 power peaks at only ~0.8. This suggests that power is being lost somewhere in the interferometer, possibly within the PRMI (e.g., mode mismatch or clipping).
The lock duration is currently limited to ~30 minutes, and the cause of the lock loss remains unclear.
A detailed lock-loss investigation is needed as soon as possible.
---
### CARM UGF Decrease
In this lock attempt, the CARM unity gain frequency (UGF) appears to be too low (~15 kHz; Fig. 2), even though the REFL PD power is 1.1 mW—almost the same as during a previous lock 48 days ago.
Figure 3 shows the REFL PD power (top) and the LSC_LOCK state (bottom) from that time. The cursor marks the current REFL PD power, showing it is nearly equal to that of the previous PRFPMI_RF_LOCKED state.
The reason for the UGF decrease is currently unknown. However, the overall gain is about **–11 dB lower**. (due to an unoptimized 45 MHz demodulation phase?)
Because the lock acquisition procedure was optimized assuming a CARM UGF of 50 kHz, we cannot increase the laser power without modifying the procedure.
---
### DARM UGF Anomaly
Interestingly, the DARM UGF seems to have increased compared to the previous configuration.
Even after reducing the DARM loop gain to ×0.4, the UGF remained unchanged (Fig. 4).
When the loop gain was set to 1, the DARM loop began to oscillate at ~130 Hz shortly after IFO locking, leading to a lock loss (Fig. 5).
The reason for the DARM UGF increase is unclear, especially considering that the DARM signal is normalized using both arm transmission powers.
---
### PRCL and MICH Loops
The UGFs for PRCL and MICH also appear to be slightly higher than before.
Figures 6 and 7 show the respective open-loop transfer functions.
In this test, normalization with POP90 was enabled, which may explain the differences.
---
In any case, further investigation is needed to restore stable IFO operation.
Tonight, I have left the IFO in the **PRFPMI_RF_LOCKED** state to observe its long-term stability.
