KAGRA Logbook

Ushiba, Tanaka

## Abstract

We recoverd PRFPMI DC lock after IFI chamber close and RF lock recovery in klo33190. The jitter peaks at 127 Hz and 220 Hz seem to be disappeared. Then, we performed some trial to improve the stability and the sensitivity. This time, engaging whitening filters for GAS is most effective. Thanks to these engagements, the heights of peaks around 20Hz were significantly reduced. This indicates that GAS DAC noise seems to kick these resonance, (maybe) especially 18 Hz vertical resonance. And OMC initial alignment with RF sideband seems to be valid for improving the stability.

This time, we could not engage the whitening filters for BF V due to disconncetion of BIO by the work in klog33152, unfortunately. Tomorrow, we recoverd the connection and try to engage the whitening filters.

## What we did

After the RF lock recovery in klog33190, we recovered the DC lock just by requesting the LSC_LOCK guardian to lock PRFPMI with DC readout. The 127 Hz and 220 Hz peaks seem to be disappeared by the IFI work in klog33110.
About 18 Hz peak seems to be large. According to Ushiba-san, this resonance is #61 mode in the simulated eigenmode list by Okutomi-san (JGWdoc), which is one of vertical modes of the whole suspension. Therefore, we tried to avoid kicking the mode due to DAC noise by engaging the 2-stage whitening filters for ETMX GAS and BF V. Unfortunately, we found that BIO for the BF V driver was disconnected by the work in klog33152. So we engaged the filters only for ETMX GAS in this time. Fig.1 shows the DARM sensitivity spectrum between 10 - 30 Hz, red curve is the one after engaging the filters. The 18 Hz peak height got lower to about 1-order. Moreover, the peaks around 20 Hz got lower drasitcally. This indicates that DAC noises in ETMX GAS feedback kicked these resonances. We also engaged the filters for GASs, BF Vs and IPs of the other Type-As but the spectrum in this region seems not to be changed. Tomorrow, we connect BIO for ETMX BF V and try to engage the filters whether make the peaks in this region much smaller or not.
On the other hand, the stabillity seems to be still bad, the duration seems to be at most ~20 mins. the main cause of lock loss seems to be the 40 Hz oscillation of DARM. We performed the initial alignment for OMC with TEM00 RF sidebands as reported in klog33034 (Fig.2). After that, DARM open loop gain increased to about 1 dB (fig.3 blue -> brown). This indicates this initial alignment is valid for TEM00 carrier. We also measured the open loop transfer functions of MICH (fig.4) and PRCL(fig.5). Each UGF is almost the same as the designed one.
Furthermore, to improve the stability, we modified the resG filters in the DARM control filter and succeeded in incrasing the phase margin slightly (fig.3 brown -> red). Thanks to this modification, the RMS of DARM residual is also reduced (fig.6).

## Note

Resonances around 40 Hz are still kicked. As 40 Hz peak is raised, the bump appears around 80 Hz. So the nonlinear coupling seems to be large around 80 Hz.
When 40 Hz oscillation occured, the power toward OMC increased maybe due to large alignment fluctuation of PRFPMI. To improve the stability, it might be necessary to enlarge the bandwidth of ASC.