KAGRA Logbook

[Yokozawa, Dan, Yuzurihara, K. Tanaka]

## Abstract

We performed the phasing for I/Q signals of RFPDs on POP and REFL table (POP_PDA1_RF{17, 45}, REFL_PDA1_RF56) with Frequency excitation and minimized each Q signal.

## What we did

Before the phasing, we aligned PR3 to maximize the normalized GR flash to about 1, and we locked X-arm with both IR and GR beams and engaged each ADS to maximized each transmitted power. Then we recorded the oplev values of IMMT2, PR2, PRM in GOOD_OPLEV values when the transmitted powers seemed maximum.
We lost the lock at once and then relock X-arm with only IR beam. We connected the excitation path( DGCARM EXC -> SUMMING NODE EXC A -> SUMMING NODE SLOW OUT -> CARM CMS IN2). In that time, we decreased the CARM IN2 GAIN from 0 dB to -20 dB before connecting the path. We injected the excitation signal at 1026 Hz from K1:LSC-DGCARM_EXC ch to oscillate the laser frequency.
We could see a beat signal in I and Q of an POP_PDA1_RF17 at 1026 Hz. We measured the ratio between I and Q signal amplitude and obtained the rotation angle from $\arctan(Q/I)$ . This angle indicated the phase difference between the signal and the I quadrature. So we can minimize the Q signal with putting this angle in demod. angle.
We rotated the demod. phase (K1:LSC-POP_PDA1_RF17_PHASE_R) to minimize the Q signal because we used POP17Q as the error signal of MICH so that the frequency oscillation would not be seen in POP17Q. Fig 1 shows the phasing result. As you can see, the Q signal are coverd by the noise level.
We also performed thephasing for RFPDs (POP_PDA1_RF45, REFL_PDA1_RF56) with the same procedure. Fig 2 and 3 show the result. Each Q signal got small enough to be coverd by the noise level.

## ToDo

We need to tune the relative gain and phase between I and Q signal for REFL_PDA3_RF45 and REFL_PDA4_RF17 with AM laser before the phasing because the I and Q signal are not always orthogonal due to the analog phase shifter error.
We need to perform the phasing for the 3f PDs (REFL_PDA2_RF{51, 135}) with MICH or PRCL after we succeeded MICH and PRCL lock because the beat signal between the carrier and 3f sidebands (not +1f and -2f) was too small to see the frequency oscilation.