Elenna, Jenne, Gabriele, Matt, Trent, Louis, Georgia
- Jenne and I checked the fast shutter and ran the fast_shutter test (SYS -> AS port protection -> run)
- After the initial alignment (alog here) we got through green arms locking at up to CHECK_MICH_FRINGES. When we tried to do PRMI there was a timer that kept taking it back to MICH fringes since the POP90 buildups were so poor. On the AS_AIR camera the alignment looks bad in yaw.
- We edited ISC_LOCK line 1353 to lower the threshold for PRMI flashes on POPAIR_B_RF90 from 20 to 4.
- We ran the dark offsets script with the IMC offline and the ALS shuttered (sdf screenshots here), and started locking again.
- The Y arm was finicky, we struggled to get the flashes above 0.6... We lowered the PD threshold in the ALS Y PDH (H1:ALS-Y_REFL_LOCK_TRANSPDNORMTHRESH) from 0.7 to 0.6.
- We also lowered the threshold in ALS_ARM gen_INCREASE_FLASHES from 0.9 to 0.6.... feels like cheating, maybe the alignment onto ISCT1 is off? The camera alignment for ALS_Y also looks bad (see screenshot)
- Even though we did an initial alignment earlier today, the PRMI flashes are bad and the MICH fringes don't look great. PRMI_ASC caused a couple of locklosses, so we took the guardian to PRMI_LOCKED and tried aligning it by hand. Elenna aligned PRM and BS, but the buildups and camera indicated bad pitch alignment.
- We re-ran initial alignment, and then the guardian took us straight up to PRMI_ASC, DRMI also locked easily! POPAIR_RF18 was a little noisy.
- We started the CARM offset reduction sequence but lost lock at DHARD_WFS, the two arms were very different (higher build up for Y arm), when DHARD came on the arms were pulled closer together but we lost RF9, and we lost lock.
- We decided to stop at DARM_TO_RF, step the X arm (X hard) closer using the move_arm_dev script and fix PRM and SRM alignment to increase the buildups.
- While we stepped the arm and adjusted PRM, the green arms lost lock. We think we might be aligning the whole IFO to some bad ITM alignments (how accurately aligned are they after the baffle PD alignment script?)
We decided to call it a night at this point.
A other things that we found during the inital alignment process:
- Beckhoff reboot from yesterday morning set some settings incorrectly. These show up with a change date in SDF of 1989. The ALSX PLL common compensation was off making the PLL not lock.
- polarization controller box was on, making ALS noisy.
- AS_C whitening was left at 36dB, has been reset to 18dB, this was before the dark offset was rerun.
We changed the threshold in ALS_ARM gen_INCREASE_FLASHES back to 0.9
We are thinking of increasing the guardian timer for INCREASE_FLASHES but have not done so yet.
We also reverted this change, since DRMI flashes are back to their usual level:
We edited ISC_LOCK line 1353 to lower the threshold for PRMI flashes on POPAIR_B_RF90 from 20 to 4.
Sheila, Gabriele, Corey
- Based on the alog from last night, we decided to start with an initial alingment, using the camera set points that were set yesterday based on the baffle PD alignment. This went smoothly, and allowed us to lock to PRE_ASC_FOR_FULL_IFO several times.
- We undid some guardian changes from yesterday, we removed all DRMI ASC except MICH, and put the DHARD gain increase in DHARD_WFS back in.
- We realized that we have been loosing lock because the ALS DIFF guardian was thinking that we lost lock (when we skip shuttering ALS this checker doesn't get disabled). I've added IDLE states to both ALS_DIFF and ALS_COMM, for now we are manually selecting these when we get past the ALS steps of the CARM offset reduction.
- After that, we are able to sit and walk alignments for longer in PREP_ASC_FOR_FULL_IFO. We lost lock as we moved CHARD P in the negative direction which was improving the PRG, but in the last seconds of the lock POP90 was increasing. Our next plan is to turn off MICH ASC while we walk alignment.
Summary of guardian changes that are useful to make when we are recovering from a vent, for future reference:
- bypass shuttering of ALS in CARM offset reduction by changing wieghts of connections, this allows us to reset the green references when we lock.
- Greatly extend the timer for INCREASE_FLASHES
- change connection weights to go through CHECK SDF rather than SDF revert (if this is what we want)
- we are thinking about the check for giving up on DRMI and moving to PRMI, the logic we have worked well during O4a so we are hesititant about changing it. A workaround for if it is switching to DRMI when we don't want that is to pause ISC_LOCK and let the DRMI guardian lock DRMI.
Some guardian changes made yesterday to think about maybe reverting:
- got rid of DHARD gain reduction at DHARD WFS
- added SRC1 back to DRMI ASC
- change of ALS normaliztion to Y arm.