I started the lockloss investigation on 2025/02/17-19. The other plots are summarized in wiki.
Oscillation of ETMX with 1Hz
- 2025-02-17 00:59:43.312500 UTC
The label for this lockloss was ISC_WD_L. So, the watch dog was turned on just befoer the lockloss. The cause of the watch dog was the oscillation of ETMX {pitch, yaw}, and ITMX yaw with 1Hz. (Fig) The amplitude seems to increase in time and touched the watch dog threshold. (Fig)
This phenomena didn't appear after this lockloss. So, the tuning of damping filter by Ushiba-san has fixed the issue. [klog32713] [klog32712]
OMC PD started to oscillate with 40 Hz
Just before the lockloss, the oscillation in OMC PD with 40 Hz started. With the chat Kenta-san, the possible cause is the change of the gain for DARM control and it makes the shift of the unity gain frequency. To test this hypothesis, I will check the spectrum before
Similar phenomena happeded in the following lockloss. The frequency of the transient was 52 Hz. (Fig) I will check the spectrum. Although I'm not sure the relationship, there was the coincident transient in the PRM_COIOUTF_H1 and H3. (Fig)
- 2025-02-19 02:52:26.562500 UTC
Other lockloss
There are several locklosses at these days. But, at the lockloss time, the excitatino flag was on, that is K1:GRD-SYS_EXC_OK =1. So, I will leave them.
- 2025-02-18 05:43:39.187500 UTC (the excitation in psliss. ISS control looks to have a trouble. Fig)
- 2025-02-18 08:32:26.062500 UTC (the excitation in visemynb. I'm not sure the reason, but the K1:VIS-ETMX_ISCWD_WD_AC_L_RMSMON increased just before the lockloss. And, the watch dog for ETMX was turned on. Fig)
- 2025-02-18 14:22:44.062500 UTC (the OMC PD detected the transient behavior just before the lockloss. Fig)
- 2025-02-19 04:42:33.687500 UTC (the excitation in visitmyp)
