Reports 1-1 of 1 Clear search Modify search
MIF (ASC)
shinji.miyoki - 11:05 Sunday 01 March 2020 (13236) Print this report
IFO short locking and recovery

[Yokozawa, Miyoki]

What Happened

IFO continued several minutes observation and lock loss in the morning on 1st March. Relocks were relatively fast and reliable, however, IFO lost lock very soon.

10 time larger 10Hz and some resonances around 20-30hz still exist.

 

Remedy

We took the alignment again and we found OMC injection was not good in Yaw QPD1. We adjusted it using the X arm resonance. We will keep monitoring. BPC? is not used now.

I am afraid of signal saturation of CARM loop or so due to these 10 times larger resonances.

Comments to this report:
shinji.miyoki - 11:34 Sunday 01 March 2020 (13237) Print this report

I found many glitch noises in LSC-MCL_OUT_DQ. At least no such spikes in other channels in "ndscope LSC error and control signals".

It has very characteristic point. It alway forms  a pair with same time interval.

Images attached to this comment
tomotada.akutsu - 13:16 Sunday 01 March 2020 (13239) Print this report

Yokozawa, Akutsu

As "the" 10.1Hz peak would be problematic for long-term stable locking, we added and applied a resG filter in the DARM loop, and it has been working... let's see.

Details

  1. I added resgain(10.125, 10, 50) as FM4 In the filter bank K1_LSC_DARM2. I also applied ramp time of 2sec at "Switching" in the foton.
  2. While the interferometer was being locked, I did "Load Coefficients" for the filter bank.
  3. Turned on FM4; luckly, the interferometer keeps being locked. As far as I measured the 10.1Hz was reduced in the error signal (to be posted by Yokozawa-san; 13241).
  4. Because this is recognized as a change of the SDF parameters(?), a flag showing "KAGRA is transfering obs data" was once redden, so Yokozawa-san operated the SDF screen to accepr this change as an healty state.
  5. At the same time, he modified the Gardian so that FM4 in DARM2 can be turned on at the same timeing with FR3 in DARM2, in which two resG has been already implemented, maybe for dealing with 23Hz and 45Hz peaks. For turnig off the newly added filter, it seemed the "turning off all the filters" would be applied at once according to the Guardian code, so we did not add such a code to turn off specificaly the FM4.
  6. I have lunch.
takaaki.yokozawa - 13:18 Sunday 01 March 2020 (13241) Print this report

I attached teh OMC DCPDA spctrum and RMS

Before and after resG implementation

 

http://klog.icrr.u-tokyo.ac.jp/osl/uploads/13241_20200301051822_screenshotfrom20200301130323.png

Images attached to this comment
shinji.miyoki - 13:18 Sunday 01 March 2020 (13240) Print this report

I closed up a pair of glitch. Then,

1) One glitch actually consists of two opposite side glitches as the
attached file.

2) Yes, same glitch can be seen in IN and OUT.

3) Time  interval is ~ 0.4 btw sec two pair of glitches, and  ~0.06 sec
btw inside a pair of glitches.

4) I cannot judge this is fatal or not for stability. Apparently, no
correlation btw real time sensitivity curve and these glitches.

Images attached to this comment
shinji.miyoki - 21:49 Sunday 01 March 2020 (13248) Print this report

About 10.1 Hz.

Around 21:40, this 10.1 Hz was sometimes damped, sometimes excited. Before, it was always excited after we noticed this resonance. 

Search Help
×

Warning

×