Took the in-lock charge measurements this afternoon as we wanted the IFO to loose lock so this was a good opportunity to test the SWAP_BACK_TO_ETMX SUS_CHARGE code that we edited last week (67878). We also tried to stay in NLN rather than NLN_CAL_MEAS. The IFO lost lock during the SWAP_BACK_TO_ETMX state.
We the same issue as previously where all the ESD_EXC_{OPTIC} nodes went into error couldn't open thier excitations. We took the guardians to DOWN and restarted them using 'guardctrl restart ESD_EXC_{OPTIC}'. We think the awg.stop() and awg.clear() functions aren't completely clearing the excitations, I'll touch base with Erik.
The in-lock charge measurements excitation frequencies are 12, 13, 14, 15Hz for EY, EX, IX, IY respectively. While I was doing these excitations, Rahul checked on the M0 ETMX stage and saw no 11Hz peak. We can conclude the peak that was seen in alog 63766 wasn't due to the in-lock charge measurements.
Ryan C is analyzing these using new code he added to calculate V_eff ad V_eff2.
Lockloss was at 22:31:13.22 UTC. At that time the swap back from ITMX to ETMX hadn't started. We were just finishing restoring the settings on ETMX after the excitation. Cause: the L3_DRIVEALIGN_L2L output was turned on with a gain of 1 when it should have been had gain of 0.
This was due to guardian restart to clear the awg.py error. During the first SETUP the guardian saved the value of L2L gain as 0 and then changed it to 1 but when the guardian was restarted the gain of 1 was saved to be later restored.
This shouldn't happen if the guardian works nominally. As a extra checker, I've added a line to RUN_ESD_EXC to not turn on the L3_DRIVEALIGN_L2L input if the gain is not zero and the filter input is not zero.
We didn't get far enough to test Georgia's fix to the transition state (alog 67878).
Ryan C analyzed these measurements along with V_effective values in alog 68089. ETMY did not run correctly as the ESD HV was off alog68446.