Why the ~2.5kHz signals in the PDA1 that is, as it were, an error signal with PDA2 signal dose not appear in the FAST path and appears a slope in the woofer path?
[YokozaWashimi]
We have installed the accelerometers and microphones for the POP, POS, and AS tables.
We will turn on them tomorrow morning.
We located a small hut behind the Analysis building.
Before installing the sensors, the temperature monitor is been performing now.
I changed the setpoints for F1 and F3. I offloaded the F0, F1, and F3 keystones with the FRs.
| EY | count | um |
| F0 | 23100 | 4527 |
| F1 | -2754 | 1230 |
| F2 | 7000 | 1322 |
| F3 | -3561 | 1083 |
| BF | 825 | -700 |
With Sato-san,
We prepared for installing Mid-size baffle SR3 and SRM.
For SR3:
Baffle plates was assembled with the frame. We checked vertical laser leveler position between SR2 and SR3.
For SRM:
We opened cardboard box and put baffle flames and plates into the clean booth. (pic.1)
We assembled targets for SRM. Both HR side and AR side.(pic2, pic3, pic4) It was aligned with laser leveler between SR2 target and SRM recoilmass scratch line.
Today, we performed the characterization of the SR2 oplevs.
Before this work, we performed SR2 suspensions to aligned state.
Then I changed the oplev work state manyally (Mainlhy, turned off the TM OLDAMP)
Then we measured the beam power and those are result;
IN : 0.043 mW
TILT QPD : 0.005 mW
LEN QPD : 0.008 mW
We couldn't measure the beam power after the SR2 reflection and before the beam splitter due to structure.
(And beam power was too low? or Should we change the beam source)
Next, we checked the QPD state (high gain and low gain mode) both are left (high gain mode)
Finally, we measured the beam profiler.
TILT QPD : major=0.072 mm, minor=1.750 mm
LEN QPD : major=0.23 mm, minor=0.21 mm
Who: KIMURA (ICRR), M. TAKAHASHI (ICRR), NAKAJIMA (MIRPRO) and SAWADA (HOKUTO)
Task:
We performed leak repair work on the sealing surface of the EYA side of the gate valve.
The bellows connected to the gate valve was retracted and the seal face of the connection was opened.
As a result, two scratches were found on the seal face, which were contaminated by grease and imagined to be the cause of the leak.
The grease was wiped off with alcohol and the scratches were repaired by lapping.
After repairing the leaks on the seal surface, a new metal gasket was inserted and the flange was tightened with bolts.
Tightening was done in stages at torques of 20, 40, and 60 Nm.
Please see attached photos to confirm the repair work.
After several investigations, we found that F1 GAS seems rubbing, so we moved it down by about 200 um.
Then, we remeasured thee TF from F1 GAS excitation and it becomes healthy (fig1).
We are not sure the exact reason but one possibility is internal rubbing due to tilt of F1 body.
Anyway, we need to silghtly lower the setpoint of F1 GAS, which also requires to reconsider the setpoints of the other GAS filters to avoid changing the height of the mirror.
After Takahashi-san decided the new setpoints, I will remeasure the TFs again and continue health check.
At MN and IM stage, TFs of all 6 dgrees of freedom seems fine (fig1 to fig12).
Abstract:
I checked the mechanical transfer functions of ETMY and found that GAS F1 is not healthy.
I asked Takahashi-san to investigate the problem.
Detail:
Figure 1 and 2 show the TFs when exciting from F1 GAS at present and previous in-air health check (klog20729), respectively.
Obviously, current TFs are strange.
So, I asked Takahashi-san to check the GAS status.
At MN and IM stage, TFs of all 6 dgrees of freedom seems fine (fig1 to fig12).
After several investigations, we found that F1 GAS seems rubbing, so we moved it down by about 200 um.
Then, we remeasured thee TF from F1 GAS excitation and it becomes healthy (fig1).
We are not sure the exact reason but one possibility is internal rubbing due to tilt of F1 body.
Anyway, we need to silghtly lower the setpoint of F1 GAS, which also requires to reconsider the setpoints of the other GAS filters to avoid changing the height of the mirror.
After Takahashi-san decided the new setpoints, I will remeasure the TFs again and continue health check.
I changed the setpoints for F1 and F3. I offloaded the F0, F1, and F3 keystones with the FRs.
| EY | count | um |
| F0 | 23100 | 4527 |
| F1 | -2754 | 1230 |
| F2 | 7000 | 1322 |
| F3 | -3561 | 1083 |
| BF | 825 | -700 |
(Note: alignment situation and mirror position are different from the past.)
Fig.1 and Fig.2 show the current today's picture and one on 5/16, respectively.
They are available also in link below.
https://gwdet.icrr.u-tokyo.ac.jp/~controls/capture/tcam/ETMY/2022/0628/TCam_ETMY_00110_2022_0628_130627.png
https://gwdet.icrr.u-tokyo.ac.jp/~controls/capture/tcam/ETMY/2022/0516/TCam_ETMY_01110_2022_0516_164500.png
Sometimes it is around 5kHz.
But I can't find any process that was stopped.
SummaryPage processes are automatically restarted.
The oscillation frequency seems to be around 2kHz - 2.7kHz. Not a single frequemcy.
When a step down signal is visible in the woofer FB signals (ALS-X_FIB_WOOFER_{OUT/IN1}), we can see a ~ 2.5kHz glowing up chirp signal in ALS-X_FIB_PDA1/2_OUT outputs. Should we inspect some mechanical resonances ?? The resonance fringe width is too large comapred with the Michelson full fringe width.
Or just loop instability due to less fringe width??
With Hirata-san.
After cleaning with First Contact (klog 21281), we released the suspension and found two security stop screws of the recoil mass touching the security structure beneath it (Fig. 1). The two other screws on the other side were not touching.
In order to fix this, we set the suspension in DAMPED state and moved the F0 yaw stepper motor until the oplev came into range. The screws were no longer touching the security structure (Fig. 2). The screws at the opposite side were also free (Fig. 3).
.jpg)
.jpg)
Who: KIMURA (ICRR)
Task:
I replaced the leaked 22 Pin hermetic sealed connector of Yea duct shield cryocooler to new one.
New hermetic sealed connector have already leak tested, and confirmed leak level was less than 10^-13 Pam^3/s.
Ushiba, Dan, Nishino
Alignment improvement
We adjusted the alignment of Gr M4 (see POP layout) to maximize the error signal amplitude (Michelson fringe) and it was improved from ~1000 to ~2000 counts.
The lock was stable for ~1 hour in this state (without thermal feedback).
But after that, the lock became unstable and the error signal amplitude has become ~1000 counts again.
GRX DCPD calibration
Count: 1772 counts (bacground=5.5 counts)
Power right after the GR fiber: 9.93(1) mW -> calibration factor = 5.62e-3 [mW/count]
Power in front of the GRX DCPD: 0.71(1) mW -> calibration factor = 4.02e-4 [mW/counts]
Open-loop function of FIB servo
The measured open-loop function of the FIBX servo is below.
UGF is ~7.5 kHz and the phase margin is ~ 43 degrees.
We could see the tweeter's resonance at 58.912 kHz.
Small change; PDA1 the gain
We noticed that the gain knob of PDA1 in the PSL room was halfway between 30dB to 40dB. We fixed it to 40dB and did not see any significant change in the output signal.
When a step down signal is visible in the woofer FB signals (ALS-X_FIB_WOOFER_{OUT/IN1}), we can see a ~ 2.5kHz glowing up chirp signal in ALS-X_FIB_PDA1/2_OUT outputs. Should we inspect some mechanical resonances ?? The resonance fringe width is too large comapred with the Michelson full fringe width.
Or just loop instability due to less fringe width??
The oscillation frequency seems to be around 2kHz - 2.7kHz. Not a single frequemcy.
Sometimes it is around 5kHz.
Why the ~2.5kHz signals in the PDA1 that is, as it were, an error signal with PDA2 signal dose not appear in the FAST path and appears a slope in the woofer path?
with Hirata-san, Fabian-san and checked by Miyoki-san
We removed four parts of FC on the HR surface and inspected it using green light.
Some FC stains could be cleaned successfully but unfortunately, some FC residues were still on the circumference line of the first FC.
It was not perfect, but we decided to finish this cleaning work.
Please confirm each picture.
https://www.dropbox.com/work/Subsystems/MIR/mirror%20inspection/SR3/20220627
DAQ Error Log Analysis
DAQ Timing Error.
Check the results of countermeasures against this error.
Analyze the DAQ log and check against the DAQ error.
Data
Fig.1 Comparison of Ask for retransmission error and DAQ_OK in 2022
Fig.2 Ask for retransmission error in 2022
Fig.3 K1:GRD-SYS_DAQ_OK in 2022
Fig.4 Comparison of Ask for retransmission error and DAQ_OK in 2021
Fig.5 Ask for retransmission error in 2021
Fig.6 K1:GRD-SYS_DAQ_OK in 2021
Figs. 1 and 4 are composites of 2, 3, 5, and 6, respectively.
Figs. 2 and 5 are graphs of daily totals of the number of "Ask for retransmission XX packet port 7097" errors and the number of packets in the DAQ log file, compiled by a script.
/users/DAQ/daqd_log_analize/daqdlogana.py
Fig.3 and 6 were obtained by specifying the period in ndscope.
Analysis
Fig.1-1
DAQ and DGS LAN cables were exchanged.
DAQ_OK seems to improve gradually during this period.
Fig.1-2
The percentage of DAQ_OK errors increased around the time I removed the VIS dummy channel in K-Log#19981 on 3/3.
Around 4/9, this was just a test that increased in frequency.
Fig.1-3
DAQ server setting was changed to FIFO, but "Ask For..." errors occurred frequently.
Later, the parent of daqd was also changed to FIFO, and the frequency decreased.
Fig.1-4
The test was performed with DAQ's Swap disabled.
"Ask For..." Fewer errors, but more error Packets at a time.
DAQ errors appear to be decreasing.
Result.
What we know so far.
1. "Ask For..." The frequency of errors and the frequency of DAQ_OK errors do not necessarily match.
However, "Ask For..." The frequency of errors and the frequency of DAQ_OK errors are similar.
The number of packets has nothing to do with the error frequency of DAQ_OK.
2. If the parent process of daqd is delayed, the frequency of "Ask For..." errors in k1tw0 and k1tw1 increases.
This also does not necessarily increase the frequency of DAQ_OK errors.
3. after changing all daqd to FIFO in K-Log#20717+21102.
"Ask For..." The error situation changed.
Before the change, fw0 always had an average of 100 errors.
However, fw0's "Ask For..." errors decreased.
And fw1 became worse.
However, the number of packets per time has increased.
We will see how it goes for about a week.
I investigated that that may come from type of fiber collimator.
Now, we used LPC type OZ optifcs fiber collimator, but the beam path of PR2 is too long to use it (~2,400 mm)
So, I started the investigation to use LPF type fiber collimator (which is used for PF oplev of Type-A, 3500 mm in beam waist)
There are several spare of PCF type fiber collimator, so I checked the beam profile at the point of 2,400 mm
As the result, No16 one is reasonable one by replacing.
Forced interval as the 3 times longer than the exposure time is inserted.
For example, when we take a picture with 60s exposure time, we always need to wait 180s to take a next picture by this modification.
-----
In the case that the picture is taken with changing exposure time, 1-3 times longer duration is required to revert to the movie mode.
Spent time depends on the timing of the guardian process, CPU load etc. In other words, it depends on your luck.
If we request to take a next picture before reverting, camera servers hang up.
In originally, I inserted an interval time same long as exposure time and requested users to wait to revert completely.
But I learned users cannot wait at all. So I changed interval time as the maximum duration in order to avoid the hang up.
Date: 2022/6/27
I checked YPcal beam positions on ETMY at 7am today.
The both beams are still on the ETMY with a few cm shift from the designed positions.
DGS Regular maintenance day(6/24)
Stepper motor
[Checked]
Check if the power supply can be controlled by BIO
Check if the script can be started
Checked that the limit switches installed are green.
[Check result]
No problem
ITMX is suspended for maintenance at other locations.
Release Sensor
[Checked the Release Sensor.]
Checked that Release Sensor is 0, 1: Seated.
[Check result]
ETMX, ETMY, ITMX, ITMY : No problem
Picomotor
[Confirmation]
PingTest is performed.
[Result]
Only known
NG:
BS(IM)
=> Hang. Not Recovery.
MCI,STM1,
AS_WFS,POP_POM,POS_POM,POM1
MCF(MCO,MCI)
=> Confirmation only, no recovery operation was performed.
OK:
ETMX,ETMY,ITMX,ITMY
BS(BF),SRM(IM,BF),SR2(IM,BF),SR3(IM,BF,STM)
PRM(IM,BF),PR2(IM,BF),PR3(IM,BF)
MCE,IMMT1,IMMT2,OMMT1,OMMT2,OSTM
REFL_WFS,POP,POP2,POS,POS2
PCAL(EX1,EX2,EY1,EY2):Not checked because it is ON only when used
Temperature controller
[Checked]
Is the temperature acquisition working properly?
[Check result]
Is the temperature acquisition working properly?
=> No problem
Precision air Processer
[Check Contents]
Is the temperature acquisition working properly?
[Result]
No problem
FunctionGenerator
(Confirmation)
Is there a ping response?
Are all EPICS channels alive?
[sitemap]-[Commissioning]-[ALS FG]-[Xarm],[Yarm].
[Check result]
X and Y PLLs are working properly.
HWP control PC
(Confirmation)
Check if it is possible to login to the control PC.
(Result)
Both k1hwp0 and 1 are OK.
with Hirata-san, Fabian-san
Before removing the FC, we cleaned the floor again, and then we peeled off both FC and inspected the mirror again.
It looks pretty clean around the center of each side, but unfortunately, residues on the FC circumference remain on both sides respectively.
I don't know why this happened because I thought I had applied it the same way I always do...
Furthermore, some line stains, or scratches, on the upper right side of HR still remain. FC couldn't remove them.
https://www.dropbox.com/work/Subsystems/MIR/mirror%20inspection/SR3/20220624
After some discussion, we started to re-apply partially to remove these residues only on HR.
We are going to remove them and inspect the mirror again next Monday.
.jpg){kind=link}
.jpg){kind=link}
