Reports 1-1 of 1 Clear search Modify search
MIF (General)
hirotaka.yuzurihara - 17:51 Tuesday 22 April 2025 (33509) Print this report
lockloss investigation: 2025/04/14~04/21

[Kenta, Guo chin, Yuzurihara]
We performed the lockloss investigation for the recent lockloss from the OBSERVATION state of the LSC_LOCK guardian, between 2025-04-14 01:23:52.562500 UTC and 2025-04-21 22:56:05.437500 UTC. The previous lockloss investigation was posted in klog33395. During this period, there were 38 lockloss.

Lockloss with the interferometer work

16 lockloss occurred in the WORKING_WITH_OBSERVATION_WITHOUT_LINES state. In adittion, 4 lockloss occurred during the excitation.

1Hz seismic motion made the IMC lockloss

For 10 lockloss, we saw the excess of the seismic motion with 1~10 Hz, which made the oscillation of the IMC length control and made the IMC lockloss.
The time and date are summarized below:

As commented by YamaT-san in klog33416, recently, the lockloss labeled by `IMCL` happened very frequently. For example, during this period, 15 lockloss (including the WORKING_WITH_OBSERVATION_WITHOUT_LINES) was labeled by `IMCL`.
It's a timing to investigate more and to model the lockloss mechanism.

I newly found that the noise excess is happring in oplev signals of IMMT1 pitch, IMMT2 pitch, PR2 pitch/yaw and PR3 pitch ( sometimes SR2 pitch, and SR3 pitch ) with the coincidence of IMC length excess. For other oplevs, I couln't find the excess.
Figure 1~4 show the example of the excess. This might be a hint to model the mechanism.

I heard the DQed CARM channels and the details of the CARM (IMC) feedback control from Kenta. Although I checked the feedback signals to EOM, PZT, and TEMP, there was no clear coincident excess. I will extend the analysis with the signals of REFL45.

I checked if the feedback signal to the coils of MCE was saturated or not. it was not saturated. (I forgot that the existance of OVF_MCE label in SYS_LOCKLOSS guardian.)
 

Under the investigation

For 2 lockloss, we need more investigation.

  • 2025-04-19 05:21:55.062500 UTC
    • PRCL error signal is getting large. (Fig)
    • Just before the OMC DCPD was saturated, the OMC DCPD was almost saturated and recovered. (Fig)
    • The DHARD pitch value was -600 which is relatively large value.

  • 2025-04-21 01:18:43.187500 UTC

    • At -12 s, the excess in OMC DCPD, PRCL error signal can be seen. There was no excess in OMMT2 TRANS. (Fig) The alignment to OMC was bad at that time?

1.2 Hz oscillation in IRX and OMMT2 TRANS

  • 2025-04-19 06:55:33.937500 UTC
  • From -3 seconds of the lockloss, 1.2 Hz oscillation in IRX, IRY, and OMMT2 TRANS happened (Fig). The OMMT2 TRANS gradually increased and the lockloss happened.
  • Coincident 1.2 Hz oscillation can be seen in DHARD P error signal and oplev sinal of the ITMY pitch and ETMY pitch.

1Hz and IMC length?

  • 2025-04-21 00:24:38.437500 UTC
  • Just before the lockloss, the IMC length error signal shows the oscillation. Just after the lockloss, the seismic motion started to increase. Although I'm not confident, I guess the time delay for seismic motion data occurred due to the application of the band limited RMS filter.
    • This might have the same lockloss cause of `1Hz seismic motion made the IMC lockloss`.
Images attached to this report
Comments to this report:
hirotaka.yuzurihara - 17:50 Tuesday 22 April 2025 (33516) Print this report

large feedback by DHARD and large seicmic motion

For 4 lockloss, we saw that the DHARD pitch feedback (K1:ASC-DHARD_P_OUT_DQ) shows a large negative value. Their values are -400 ~ -800. which is a relatively large negative value, as shown in Figure 1~4.

For the first two lockloss, we could see clearly the 0.16 Hz oscillation. But for the last two lockloss, the oscillation frequency was unclear to identify.

During 4/14 and 4/15, the seismic motion of 0.1~0.3 Hz was larger than usual. (Fig) (Fig) We are doubting that the large feedback of DHARD is caused by the large seismic motion. We will collect more statistics.

Images attached to this comment
hirotaka.yuzurihara - 15:31 Wednesday 23 April 2025 (33527) Print this report

Here is the additional analysis by Guo Chin.

Preparation

He checked the 24 lockloss on April 2, 12, 13, and 14, excluding the state of WORKING_WITH_OBSERVATION_WITHOUT_LINES (upper half of the plot), and also he checked the 12 lockloss (lower half of the plot), which the LOCKLOSS guardian did not label.

He categorized the lockloss in terms of several conditions: (1) OMMT2 TRANS value > 200, (2) oscillation frequency is ~0.16 Hz, (3) DHARD feedback >= -600

The attached figure summarizes the categorization. Here, the notation indicates that O: OMMT2 over 200, H: with ~ 0.16 Hz oscillation, and D: DHARD over -600. Regarding the condition of the seismic motion, we checked the channel K1:PEM-SEIS_IXV_GND_X_BLRMS_100MHZ300 on the summary page.
 

Summary

  • The seismic motion has some effects on the OMMT2 and DHARD at the time just before the lockloss.
  • For strong seismic motion over 0.3 um/s, OMMT2 TRANS is likely over 200, and the lockless occurred. If the seismic motion reduces to 0.2 um/s, the possibility that the OMMT2 TRANS exceeds 200 is reduced.
  • If the seismic motion is larger than 0.2 um/s, we can see that DHARD oscillates at ~0.16 Hz in about half of the events. In several cases, DHARD feedback signals exceed -600.
  • If the seismic motion is less than 0.1 um/s, there is no clear relationship between OMMT2 TRANS value and DHARD feedback signal.
Images attached to this comment
Search Help
×

Warning

×