KAGRA Logbook

ETMX L-to-Y coupling issue

When the input power reached ~10 W, a 0.6 Hz oscillation of ETMX was observed, which likely propagated to other suspensions via ASC and caused lock loss.

To mitigate this, the control path for the ETMX L-to-Y coupling reduction was adjusted. (fig_001.png)

A 0.6 Hz excitation was injected into the ETMX MN LOCK FB (L), and the TM WIT P was monitored.
The MN DRIVEALIGN was tuned to minimize the 0.6 Hz component observed in WIT P.
Finally, a −28 dB gain was added to the coupling suppression path. (L2P_coupling.png)

After adding this coupling suppression path, the interferometer could not be locked (even PRFPMI_RF_LOCKED).

Current status: a band-pass filter (30 mHz–30 Hz) has been added to this path, and further tuning is ongoing.

## ETMX L-to-P (not L-to-Y) coupling issue

As dan-san mentioned, we tried to solved the issue by engaging "-28dB" and "Band pass filter from 0.1 Hz to 30 Hz" filters in ETMX MN DRIVEALIGN L2P.

Since ETMX moved in the P direction as soon as ALS_DARM was engaged if only "-28 dB" filter was engaged, we suspected the DC component of the feedback signal made the DC Pitch alignment worse. So we added the band pass filter ( 0.1 Hz highpass and 30 Hz roll-off) to send the Length feed back signal around 0.6 Hz to MN Pitch actuators. Then we retried the lock acquistion.

However, the situation was not changed.

Second, we performed the coupling reduction method by using a 30 mHz excitation. Ushiba-san assusmed that the 0.6 Hz oscillation is not the motion due to their actuator unbalance but the actual resonace motion. So, it was supposed to we should perform the method not the resonance frequency but the frequency in the DC reasion.

This time, we turned off the DC control with TM oplev during the reduction trial because the control suppressed the excitation. Then, we estimated the gain is -60dB from the amplitude of the pitch oscillation. Fig.1 shows the timeseries of related signals during the trial. At the moment on the left cursor, we turned on the DRIVEARIGN filter. Then the 30 mHz oscillation in pitch oplev (orange line in the figure) seems to be suppressed. We tweaked the DRIVEALIGN_L2P_GAIN to minimize the oplev oscillation. Now, we set the gain value to 1.2.

Then, we retried the lock acquisition with this configuration and PRFPMI RF lock was succeeded in this configuration.

## trial of incrasing laser power

We tried to increase the power. Originally, The power got ~5 W, ETMX started to oscillate at ~0.6 Hz. After the coupling reduction, even if the power got ~7W, the 0.6 Hz oscillation in pitch oplev seems to be still small (fig.2). However, the power was increased more than 7 W, the oscillation started (fig.3).

So we gave up to increase the power to 10 W todaay, we set the power to ~5 W.

Power ramp-up following Guardian

ASC OLTF

ETMX L-to-Y coupling issue