jeffrey.kissel@LIGO.ORG - posted 22:20, Thursday 28 March 2019 - last comment - 09:40, Friday 29 March 2019(48015)
H1 Rides Through 6.1 Mag EQ by demonstrating New(ish) Robust Mode
J. Driggers, S. Dwyer, J. Kissel, Y. Lecoeuche, J. Warner H1 was able to ride through a 6.1 Mag EQ on the pacific coast of Russia this afternoon by switching to the recently developed "robust" IFO-basis common-mode sensor correction system that the SEI team has been creating for the past few months. The only other configuration change was that Jenne increased the SNR of the ~20 Hz Alignment Dither System lines to "ridiculous" (a factor of 1000 higher than they normally are), and compensated for the amplitude increase by decreasing the loop gains by a corresponding amount. This was to avoid the lines being washed out by the scattering arch glitches that were appearing in the performance of the interferometer. The peak BLRMS during the EQ's arrival on site was ~1 um/s in the 0.03-0.1 Hz band, and it spilled into the 0.1-0.3 Hz band at the 0.5 um/s level. Note that the scattering arches were likely not caused by the earthquake, but the degradation in microseismic band performance currently necessary for the common-mode EQ configuration. Jim thinks this can be improved. The EQ occurred at 2019-03-28 22:06 UTC, and arrived on site about 15 minutes after. We received a ~10 minute warning from SEISMON, which successfully warned us to go into the above mentioned robust mode. We returned successfully out of this robust mode into nominal configurations ~1 hour after the first waves hit the site, once the BLRMS in the 0.03 to 0.1 Hz band went below ~0.5 um/s. A great testament to all the hard work the ISC and SEI teams have done between O2 and O3 to improve IFO robustness. A final note that during this EQ, the calibration crew was "measuring" the CAL-CS calibration paths, so the calibrated performance of the detector, DELTAL EXTERNAL and any down-stream products are non-sense. If you need to look at the performance of the IFO, look at less calibrated but perfectly useable DARM loop channels, like H1:LSC-DARM_IN1_DQ, or H1:CAL-DARM_ERR_DBL_DQ. All other IFO channels, including those for other IFO DOF lengths and all IFO angles were perfectly functional and can be used for further investigation. Jenne and Jim may have some plots and further comments to attach later.
A small correction - I increased the ADS excitation value to 1000, from its nominal value of 30. So, this was a 33x increase, not a 1000x increase. We likely could not have used this large an actuation if the quad suspension L2 coil drivers were in their lowest noise state (Acq off, LP on), but we've been in a medium noise state for the last week or so (Acq off, LP off).
To compensate, I had the ADS_MASTER_GAIN set to 0.03, rather than its nominal value of 1.0.
Also, I had the pit and yaw 3 degrees of freedom (PRM loops) set to zero, so that I was only actuating on the arms with the ADS system. I think that leaving the PRM loops on would have been fine, but it didn't seem necessary. I also was turning on and off the ADS for the arms, only having them on when it seemed like the error signals were drifting away from zero. Again, I think it would have been fine to leave them on the whole time.
I have made a new state in ISC_LOCK called EARTHQUAKE. It should be used if the seismic environment is such that SEI should be taken to it's EARTHQUAKE state. I'm not at all sure that we can stay locked with SEI's more severe LARGE_EARTHQUAKE state, so this ISC_LOCK state isn't really designed to handle that (yet). This new ISC_LOCK state so far just sets the ADS lines up, and the ADS master gain low by the same factor. This isn't tested yet, but it should be nearly the same as I did by hand yesterday. Procedure for use: