[Yokozawa, Komori]

Abstract:

We found a favorable beam spot position with an IX drive alignment value of -3.12 and optimized the INP2 and PRC2 offsets.
As a result, we recovered the previous best sensitivity—or even slightly improved upon it.

Detail:

As reported in klog:33603, we were unable to improve the sensitivity by centering the beam spot on IX.
Therefore, we attempted to find a better spot by shifting the beam toward the edge.

Before adjusting the beam spot, we turned off the beacon excitation to eliminate additional noise sources, including a sharp peak around 300 Hz.
This peak, commonly observed with beacon excitation, likely results from down-conversion of a 300-Hz distant elastic mode previously identified during drumhead mode investigations.
After disabling the beacon, both the 300-Hz peak and smaller associated bumps disappeared.

Next, we searched for an optimal IX drive align value by monitoring the DARM sensitivity in real time.
To accelerate BPC response, we increased the BPC gain by a factor of three.
This gain increase did not cause any oscillations in any state (RF lock, 10 W input, or DC readout), so we applied the gain increase (×3, or +10 dB) to FM1 in {ETMX, ETMY, IMTX, RPM}_PIT_OUTF.

We found that an IX drive alignment of -3.12, corresponding to a PRM pitch oplev value of approximately -11, yielded the best sensitivity.
We then fine-tuned the INP2 and PRC2 offsets, and the optimal values at that time were INP2 = -1.0 and PRC2 = 0.12.
Under this configuration, the frequency noise level at high frequencies recovered to its best-known level, and the arm transmission showed good symmetry (TX ≈ 120, TY ≈ 116).
The DARM open-loop transfer function is shown in Fig. 1, with a gain difference of only 0.3 dB compared to conventional measurements.

The resulting DARM spectrum is shown in Fig. 2.
The noise level in the 100–200 Hz band is the best achieved to date.
Because both the open-loop gain and the DARM sensitivity were reproducible, we maintained this configuration.

Additionally, we tested a 100 Ω transimpedance for the OMC DC photodiodes to evaluate whether the slightly higher dark noise of the 400 Ω setting could be contributing to the DARM sensitivity currently.
However, as shown in Fig. 3, the shot noise floor significantly worsened with the 100 Ω transimpedance due to the worse dark noise, so we reverted to the original 400 Ω setting.