Ryan S, Sheila, Syracuse Squad
After Sheila and Ryan turned up the BS oplev laser current more (and no one is climbing on HAM1), we seem to be requiring more successfully.
We got through DRMI_to_POP. To check it was safe, Sheila loaded the LSC POP_A input matrices into the unused XARM filter, and we checked the transfer function from PRCL, MICH, SRCL to the XARM for each input matrix. The PRCL TF looked flat and had 0 phase: good! The other DOFs had a bit of slope and phase but were close enough to 1 that we powered through, and DRMI_to_POP didn't cause any locklosses. It's on our to do list to check the LSC open loop transfer functions.
Then we sat at PREP_ASC_FOR_FULL_IFO and tried checking the phase of the ASC REFL diodes. Sheila++ plugged the LSC AO 2 into the refl servo so we could drive a line in frequency noise, which we want to tune the phasing so it shows up in the I quadrature for each QPD quadrant. I found some handy templates in /opt/rtcds/userapps/release/asc/h1/templates/phasing, complete with excitation. I turned the excitation down to 0.1 and that seemed good. The line was predominatnly in I for each quadrant in REFL_A_RF9 and REFL_B_RF9 EXCEPT Q3 of REFL_B. I imagined I would just tune the phase and see the line go down in Q, but this wasn't the case. It bounced around a bit and was hard to tell if I made progress. Attached screenshot (#1) is with the new phasing.
I accepted the changed phase in SDF, screenshot attached (#2) in case this was a bad move that needs to be reverted.
It seems like we have a 15 minute death (lockloss) timer while we sit at PREP_ASC_FOR_FULL_IFO right now, which is making it hard to get the ASC working, but we'll keep trying...
- We're able to consistently get to PREP_ASC_FOR_FULL_IFO - yay!
- Since we're not using DRMI ASC (except MICH) Ryan was touching up the PRM alignment between locks.
- We had one longer lock
- Ryan and Craig tried walking PRC2_P, INP1_P, CHARD_P trying to zero the error signals and get the buildups better. Zeroing the error signals tended to make the buildups worse
- They closed PRC2_P
- We saw a big glitch in the PRG and lost lock a few seconds later
- After lockloss we reverted the ETM and ITM sliders, but kept IM4 and PR2 where they were.
- In another lock I moved SRM to get the SRC2 error signal close to 0
- This made the buildups quieter
- INP1_P seemed close to zero so we close the loop
- The error signal and control signals went off to lala land and we lost lock
- We're thinking back to similar situations in the past and wondering if we should just temporarily not close INP1...
Ryan S, Sheila
Ryan stepped aligments using move_arm_dev and IM4 and PR2 to improve the build ups. This did not bring the error signals close to zero, and the x arm lost the green lock as he stepped it. Next time we would like to run the servo that updates the QPD offsets to keep the green locked to the arm as we move things, so we can capture the alignment with the green references.
One the build ups looked about as good as Ryan could get them with CHARD, IM4 and PR2, he stepped each of those DOFs in yaw, the attached ndscope shows a rough sensing function measurement based on this. POPX seems to have a nice error signal and zero crossing for PR2, we tried to close this loop with an input matrix of -1 and gain of 0.2, this was too small a gain and we lost lock when I made a typo trying to increase the gain in the input matrix.
REFL 45 I has clear response to both CHARD and IM4, based on the slope of these it looks like we would want to use the difference of the 45s for CHARD and the sum for IM4, this would not produce a good zero crossing for CHARD however. In the past we used the sum of the 45s (plus some 9s) for CHARD, which would still produce a zero crossing at the good alignment, but doesn't seems very sensitive.
REFL 9 signals show up more clearly in Q than in I, which seems strange since Georiga and company just phased them to have the frequency line show up in I.
Candidate input matrix: PR2 on POPX with a negative sign, CHARD on 3* REFL A 45 I + 2.5 REFL B 45 I and IM4 on REFL B 9I.
Next lock, Ryan made a nicce set of steps that we can use to check the input matrix for IM4, PR2, and CHARD p+ Y from 5:42 UTC until 5:57 UTC, improving the build ups as he went. POP X looks like a nice error signal for PR2 P +Y. We reset the green references here.
We closed PRC2 P + Y loops with a gain of 250 and -1 in the input matrix. We also closed INP1 Y with the suggested input matrix from above, which was fine. Trying to close CHARD Y with this input matrix misaligned the IFO, so we have left it off.The last screen shot shows some CHARD Y steps while the INP1 + PR2 loops were closed, the time cursor shows where the build ups looked best.
