I requested the LSC_LOCK state to DOWN.
The timing of disappearing the GRX is about 2023/05/20 20:17:29 JST(Fig.1.)
No earthquake at this timing.
Even when we turned off the shutter of GRY, some green beam can be seen in IX and EX Telephoto Camera(Fig.2.).
No difference in the input power of GRX (RMS became larger after several minutes)(Fig.3.)
I found the OLDAMP_P of PR3 moved a lot the timing of GRX lost (Small in Y)(Fig.4.).
No significant change in LVDTINF(BF, GAS), IM OSEM and their SUMOUTs.
In the TM oplev, the TILT_HOR, LEN_VER and LEN_HOR moved(Fig.5.).
At least, YamaT-san was working in the control room at that time via webcam.
Although I also noticed that there was no GR flashing in X, I just thought YamaT-san was doing some measurements with Y-arm.
I also checked the error and feedback signals of PR3 when GRX was disappeared.
Figure 1, 2, 3, and 4 show the error and feedback signals of PR3 local controls at TM, IM, and BF, and GAS, respectively.
t=0 represents the time when GRX flash was disappeared.
Only TM stage signals have a significant changes, especially in TMP feedback and TML error signals.
Also I chacked individual QPD signals (fig5 and 6).
Same as fig1 to 4, t=0 represents the time when GRX flash was disappeared.
In the QPD SUM, there is no significant change:that means the reason of the signal change doesn't seem light source.
Also, LEN_VER signals move a lot, which is insesitive to the mirror motion due to its gouy phase.
QPD itself moves accidentially??
Also, I checked PZT of PNC by once turning off bias offsets but FRX is not recovered.
Inaddition, currently AS beam spot is far from the center of gigE camera (fig7), which means IR alignment is also change.
So, it is very likely that PR3 moves somehow.
Discussions in a meeting on the 5/22 morning.
- QPD is pushed by a cable that is not well fixed?
- laser beam mode shape change? -> The sum should be changed at the same time. However, no change.
- QPD/Laser collimator table moving due to the deformation of the floor?
- This time is the Pitch drift this is different from the Yaw drift observed for a long time.
- The QPD stage is not fixed by a screw bolt?
The fact :
DC control signal of P varied a lot, Y varied very small (Fig.4. of original post)
SUM of both TILT and LEN QPD didn't change
TILT VER : not moved
TILT HOR : Moved
LEN VER : Moved
LEN HOR : Moved (Same direction of TILT HOR)
Assumption1 :
LEN QPD moved to vertical direction
- even move the LEN QPD to vertical direction, DC control signal of P didn't change
Assumption2 :
Some optics before the LEN QPD drifted
- Same reason of assumption1, it may difficult to vary the DC control P
Assumption3 :
Moved the injection part to vertical direction
- That is the most natural solution to explain this phenomena
- Check the sign of HOR drift of TILT and LEN QPD
- Check the value of TILT and LEN QPD when injection part moved.
Assumption4 :
TILT QPD moved to vertical direction
- If the LEN QPD position is not focus point, we can relatively explain, but in general, LEN QPD is insensitive to the mirror pitch motion.
So, if LEN QPD is insensitive to the mirror position as expected, the assumption3 is most natural.
But we don't know why it moved. (Touching aluminum somewhere?)