This afternoon we worked on the tuning on the ITF, foreseeing the Commissioning Run during the weekend:

• the TCS configuration is the one used lately, meaning the NI DAS on at full power, and the WI DAS off;
• at the beginning of the shift we had the same communication issue between Acl processes and/or Metatron that was described in #42069, but now we found out that they may happen after any restart of an Acl process, independently from the ACL objects that got changed (today, only SMS channels were involved);
• we went back to LOW_NOISE_1 without any modification to the lock acquisition;
• after some time there, we moved on in order to restore the DC Readout path; we had some trouble with the lock of the OMCs, so we contacted Romain;
• as it happened already in the past days, we saw the thermal effect of the decrease in the sidebands power which the DAS definitely mitigates, but not completely; since we already use the maximum power on the NI DAS, we explored the possibility to reduce the input power of a small amount in order to compensate; as a test, we did a very small number of steps (10) with the IPC_SIB1 rotator, but this was already enough to make us lose the lock;
• anyway, after about 30 minutes in dark fringe we observed the rise of a slow oscillation of the dark fringe, like it was already observed in the past days; this was pretty visible in TY on the B1p camera, and also on the PR_TY loop.; one suspicion could be that a slight tuning of the alignment of the DAS may cure this effect, but this needs more investigation;
• at this point we decided to reduce the input power already from the start, moving the rotator of 1900 steps (from -1900 to -3810, overall), which should imply a rough decrease of ~1.5 W, meaning that now the power at the input of the ITF is ~ 23.5 W (although this cannot be observed by looking at the INJ_ITF_INPUT channel, as it needs to be rescaled accordingly);
• no change of fundamental lock acquisition parameters was needed due to this little change; the only change was the very initial gain of the PRCL loop (the one in the PRCL_LOCK.ini file), because it was very low (30% of the steady state gain in LOCKED_PRITF) and probably needed some adjustment already;
• however, re-locking the warm ITF was not straightforward, as expected; we decided not to mess with the lock acquisition, as any change would have to be probably re-tuned at each failed attempt between the warm and cold states; in the end, we saw throughout the evening that the mean time after which the ITF relocks is ~35 min, with the lock usually failing later on subsequent trials, and all gains and UGFs progressively going back to good values;
• at the next success, we went along with the DC Readout path and we managed to get back at LOW_NOISE_3; as observed for the last LOW_NOISE_3 attempts last Friday and Tuesday, the DARM gain was way too high (and B1s2_DC and B1_DC need a little better intercalibration), but aside from that we did not have any issue in reaching LOW_NOISE_3:
  • in the first lock (LN3 at 1217271338, duration 3945 s) we saw again, after the usual time of ~ 30 min from dark fringe, the comparison of the slow oscillation, which got progressively worse but with a lesser degradation with respect to the lock before at higher power; the CMRF was quite bad, and the initial range was around ~ 10 Mpc and slowly degraded;
  • for the following lock (LN3 at 1217278024, duration 3550) we left everything as before, but we changed the sign of the PR_TY offset already from the beginning (from 15e-4 to -10e-4); in this case the range started from 15 Mpc, rapidly degraded to around 10 Mpc but then it was much more stable; the oscillation started again, but later and with lower rms, and it was also less visible on the B1p camera; a further in-lock optimization (to -5e-4, start at 1217281644, duration until lockloss) made the sidebands regain stability and the oscillation decrease even more in amplitude, although it is still present; the new offset is left as default in the Automation and was used from the beginning in the LOW_NOISE_3 lock starting at 1217287454, although the initial BNS range is lower in this configuration.

Possible actions tomorrow, if time and ITF allow, could be trying to increase the 8 MHz sideband again, some little adjustment of alignment working points, longitudinal noise injections and/or Alpha retuning.

Figures:

1.  Power drop at dark fringe (time = 0 when ITF_LOCK_index=120, i.e. LOCKED_PRITF_DF)
2.  SIdebands decrease towards steady state (time = 0 when ITF_LOCK_index=123, i.e. LOW_NOISE_1)
3.  CMRF as seen by the SSFS line in DARM (time = 0 when ITF_LOCK_index=160, i.e. LOW_NOISE_3)
4.  BNS Range as in Figure 3.