During the L2Y decoupling work of ETMY MN OpLev signals, I found MN LEN OpLev is so noisy conpared with the one at IX and EX (fig1, I don't compare the one at IY because ITMY has another trouble reported in klog27775).
Since the TILT OpLev doesn't have such problems (fig2), it is not likely to be related to the MN OpLev glitches (see klog27674 and following comments).
We need to investigate the reason for decoupling.
[kTanaka, Ushiba]
Abstract:
We investigated what happened in ETMY MN LEN OpLev but could not confirm the reason.
We need more time for investigation.
What we did:
1. Disconnected AA filter input and measure the spectra: ADC noise level seems fine.
2. Disconnected whitening filter input: whitening filter noise seems fine.
3. Disconnected QPD output: noise level is not changed from #2 except for 60Hz line.
4. Dump the beam in front of QPD: noise level is not changed from #2.
5. Remove beam dump: somehow noise was reduced from the initial condition.
6. Remove thin cable between QPD and nominal KAGRA cable: no significant change
7. Replace QPD to new one: somehow noise was increased.
8. shaking cables: signals of each segment seems so sensitive.
9. Centering the beam to QPD and close the windshield: somehow beam is off-centered even we close the windshield very carefully.
What we would like to do next:
The noise level seems not reproducible, so one possibility is that the cable between QPD and whitening filter is broken.
So, it is worth trying to perform swapping test with healthy OpLevs.
Also, short soft cable might be broken, so it is better to replace it to new one.
Another concern is ghost beam and scattered light.
We can see some ghost beam on QPD board, which seems not entering QPD itself but is better to dump properly.
Also ghost beam from TILT QPD surface is scattered on BS, which is located in front of LEN QPD.
So, it is better to manage the strtay light properly.
In addition, I felt that QPD signals are so sensitive to the QPD position.
That implies that the beam size on QPD is too small.
Since MN OpLevs use the same forcal-length lens, it might be a strange behaviour.
So, it is worth checking the relative position between QPD and lens.