KAGRA Logbook

[Tanaka, Ide, Hamayasu, Yanagi (Aogaku), Yokozawa]

We tried to perform to make the summary of the lock loss study during the O4c run, up to the middle of Aug. 2025.
Total number of lock loss was 215, we checked them and we succeeded to apply tags to all lock loss during O4c August.

You can find the previous studies : klog34736, klog34684, klog34685, klog34686, klog34687, klog34688.

From the previous studies, we focused to distinguish following groups.

0. Request
When the locked loss happened by the "request" (ex. maintenance, calibration, commissioning, trouble ) we checked this group.
We didn't care which control became unstable in case of this group, of course.

1. Earthquake
If the lock loss guardian tagged [EQ], we treat the lock loss caused by the large (including the low frequency ground motion) earthquake.
We didn't care which control became unstable in case of this group.

2. 1 Hz ground motion
It turned out the "nearby" earthquake which have the ground motion above 1 Hz.
We checked the BLRMS IXV Z 1HZ3 and 3HZ10, and if there are relatively larger ground motion in 10s, we defined the 1 Hz lock loss. (Threshold is about 0.1 in 1HZ3 BLRMS)
We didn't care which control became unstable in case of this group.

3. ETMY MN oplev glitch
As you know, the sudden jump detected in ETMY MN TILT oplev, we saw several glitches in ETMY suspension.
Some locked loss occurred by this glitch, by checking the control signals of ETMY suspensions, we tagged to this group.

4. IR trans and POP 90 I drop before the lock loss (tagged alignment)
By checking the IR trans power and POP 90 I power, we tagged to this group.
After the Yokozawa's investigation, the most case caused by the large motion of the 0.08 Hz in Type-A suspension.
As you know, there are 0.08 Hz resonant frequency in Type-A suspension, if this motion excess the control range of the ASC, we loose the good alignment,
IR trans and POP 90 I became lower and finally lock loss happened.
We can see 0.08 Hz oscillation in ASC error signals in case this locked loss happened.

5. Quick saturation OMC DCPD (tagged quick)
We defined the OMC DCPD suddenly saturated (order of a few miliseconds), we tagged to this group.
After the Yokozawa's investigation, we may distinguish two groups;
- Excitation of the violin mode
As shown in Fig.1. , the similar glitch happened 2 sec before, the resonant frequency was about 180 Hz.
So I guess some of quick saturation may be caused by the excitation of the violin mode resonant frequency.
ID 16, 34, 37, 112, 115, 144, 147 would be the candidate of the this group
- Unknown behavior
This is quite interesting behavior that when quick saturation happened(Fig.2.),
- the value of the OMMT2 TRANS DC decreased (60 -> 40 cnt)
- the value of the AS PDA1 DC decreased (0.17 -> 0.09 )
- the value of the AS PDA1 RF17 Q ERR decreased (0.014 -> -0.001) AS0CROSS???
- the value of the REFL PDA1 DC increased (4 -> 14)
I discussed with Ushiba-san, Tanaka-san, Takano-san and Fujimoto-san, but we cannot conclude it.
If you notice something, please contact to me.
ID 38, 114, 118, 185, 204 would be the candidate.

6. Slowly saturation OMC DCPD (tagged slow)
If the behavior of the 0. - 5. didn't detect and we can see the several oscillation detected in OMC DC PD and saturation detected here, we tagged to this group.
They would be come from the excitation of the resonant frequency of the cryo-payload (see detail in Ishikawa-san's master thesis JGWDoc17063)

Fig.3. showed the percentage of the lock loss during O4c up to the August.
Fig.4. showed the resonant frequency for the slowly saturation OMC DCPD, by our eye and ndscope cousors
We also generated several summary plots, they will appear soon.

We didn't care much for the EQ and 1Hz ground motion. This would be the next step.

Finally, I uploaded the txt file for lock loss tag.

Thank you very much for great work, Yuzurihara-san and many Aogaku members!