The cryocooler HEAD temp starts steadily increasing again. So we need the same stop-and-go again.

According to the information from Sumitomo Heavy Industry, the high-pressure He gas flow at the cryocooler "head" part and the cryocooler head part operation effectively removes impurities.

So, in our stop-and-go case, only the restart timing at the higher temp at the cryocooler head can effectively push out impurities by using the He gas flow. For more effective removal, the several stop-and-go cycles, for example between 100K (or 110K) and 90K, might be able to push out impurities. In other words, the stop timing can degas impurities, and the go timing can push out impurities. The impurities should be absorbed by filters on the flexible tube route and the oil absorber in the compressor. 

Ushiba-san, YamaT-san, Ikeda

Overview
 Ongoing trouble with HWP with encoder (K-Log#32783, #32475, etc.)
 These have been noted to be out of alignment with the specified angles.
 
 The control scripts for the HWP systems of PSL, IMC_REFL, and IFO_REFL have been updated.
 With this update, the total(deg) value now reflects the encoder readings.

Details

 Changes in Related EPICS Channels

  total(deg): K1:SYS-HWP_{PSL,IMC_REFL,IFO_REFL}_STEP_TOTAL_DEG (Behavior changed)
  step(deg): K1:SYS-HWP_{PSL,IMC_REFL,IFO_REFL}_STEP_DEG (No change)
  base(deg): K1:SYS-HWP_{PSL,IMC_REFL,IFO_REFL}_CURRENT_DEG (Newly added)
  → Reference position upon controller initialization.
  status: K1:SYS-HWP_{PSL,IMC_REFL,IFO_REFL}_STATUS (Newly added)
  → Motor operation status:
   0: Stopped
   1: Running

 Change in Calculation Method of total(deg)

  Before:
   total(deg) was calculated by incrementing (or decrementing) step(deg) each time an operation was performed. This did not reflect the actual motor position.
  After:
   total(deg) is now calculated as:
   total(deg) = Encoder reading + base(deg)
   → The actual motor position is now correctly reflected.

Changes in the Handling of Step Size for step(deg) and total(deg)

 Before: Step size was 1 deg.
 After: Step size is now 0.001 deg.
  Since the motor reading has six decimal places of precision and EPICS supports up to seven significant figures, values up to three decimal places are now handled.
  However, total(deg) currently rounds to the nearest integer due to Guardian's limitation on decimal handling.

  Planned Update:
   Guardian will be updated on Monday to support decimal points values.
   Adjustments can be made by modifying the SCALE and INFLATE parameters in medm_hwp_calib_callback.sh.

Change in DeadBand Value

 Before: Initial value was 0.002.
 After: Updated to 0.003.
  The error tolerance for the stop position has been relaxed.
  It is hoped that this change will correct the problem of PSL HWPs failing to stop or taking too long to stop.
  shift length
   +1deg x 30 times, followed by -30deg x 1 time as a set, 100 times repeated to check the stop position
   0.002: -0.0451725 deg
   0.003: -0.0494520 deg

Change in Stop Processing

　An asynchronous process now monitors the motor movement at 0.5 sec intervals.
　If the motor rotates less than 0.1 deg, it is considered an abnormal state, and the motor is stopped.

Source Code Modifications
Updated Files:

　/opt/rtcds/userapps/sys/k1/scripts/medm_hwp_calib.sh
　Modified SSH call handling.
　EPICS values are now parameterized.
　Callback-enabled function calls added.
　Decimal points support implemented.

　/kagra/apps/agilis/agilis-p_control.py
　→ Replaced with: agilis-p_control_with_callback.py
　　Added motor monitoring functionality.
　　Updated DeadBand value (0.002 → 0.003).

　Newly Added Files:
　/opt/rtcds/userapps/sys/k1/scripts/medm_hwp_calib_callback.sh
　　Handles callback reception from k1hwp to k1script.
　　Writes values to EPICS channels.
　　Initialization Behavior:
    When the reset command is executed or power is cycled (OFF → ON), the motor reading resets to 0.
    total(deg) value is copied to base(deg).
    total(deg) = motor reading + base(deg) (results in no effective change).

 /kagra/apps/agilis/agilis-p_control_reset.py
  Only executable on k1hwpX.
  Sends a reset command to the Agilis-P controller.
  After resetting, the motor initialization process in medm_hwp_calib_callback.sh runs at the next motor activation.
  Note: The reset command does not change the motor position.

Important Notes
 total(deg) Can No Longer Be Manually Overwritten

 Before: Pressing [<>] adjusted total(deg) by adding or subtracting step(deg).
 After: Pressing [<>] overwrites step(deg) with:
  base(deg) + Current HWP motor position
  If you need to manually adjust step(deg), update base(deg) as follows:
  base(deg) = Desired angle - (total(deg) - base(deg))

Next Steps
 After Guardian is updated to support decimal points values, total(deg) will be modified to accept decimal points values.

Compared with the past several stop and go of the cryocooler, this time tendency seems to be same with the past cases. So, the increase of temp will be expected soon.

We can find the successive stop and go showed the temp reduction at around -43 ~ 42 days at the level of ~ 65K.

I checked the oscillation frequency

• from K1:LSC-TRANS_DC_SUM_OUT_DQ:       43.2484 (one sample), 42.559Hz=8.5118*5(5 samples) Fig.1
• from K1:OMC-TRANS_DC_A_IN1_DQ:            43.0879 (one sample), 42.6584Hz=8.53168*5(5 samples)
• from K1:LSC-DARM_IN1_DQ:                        42.8117 (one sample), 42.0315Hz=8.4063*5(5 samples)
• from K1:LSC-ASC_PDA1_REF17_Q_ERR_DQ: 43.2484 (one sample), 42.559Hz=8.5118*5(5 samples)

One and Five sample data show different frequencies. So, can this oscillation be from a mechanical resonance or control instability? 

Vibration noise with more clamps: Fig 001

• There are some improvements around 300Hz a little.

Vibration noise with attaching the small acc (PCB 352c68) on the mirror mount: Fig 002

• There are peak shifts by ~20Hz caused by the attaching acc on the mirror mount.
• The PCB acc could sense the peaks without any tapping.

TF Measurements

• The measurement system: Fig 003
  • Measured TF from "Acc on table (ch1)" to "Laser sensor (ch2)" and from "Acc on table (ch1)" to "PCB acc (ch3)".
• PCB acc at top of the mirror: Fig 004
• PCB acc at side of the mirror: Fig 005
• PCB acc at pico of the mirror: Fig 006

Comment

• The side case was most rigid.
• If the PCB acc holder is thin enough, the attaching risk should be small.

More pictures: link

The cooling speed seemed to be stalled. We hope the temp will be kept to be constant around 76K.

With Onishi-san and Piernicola-san,

We performed 3 kinds of measurement for the mirror mount.
The main target today was to figure out how we can measure the vibration of the real IFI STM.
One candidate it using a very small accelerometer (pcb 352c68 ~ 4mm phi * 10mm length) attaching on the mirror mount.

IFI STM Experiment Summary

1. Effect of Clamp Number on Noise

• Test: Compared noise levels using 3 clamps vs. 7 clamps.
• Result: Minor improvement around 300 Hz, but 3 well-placed clamps were sufficient.

2. Small Accelerometer Test

• Test: Installed PCB 352C68 accelerometer on mirror mount to confirm that its installation does not significantly affect the resonance frequency.
• Result: Peak frequency shifted by 10 Hz, no other significant impact was observed.

3. Transfer Function (TF) Measurements

• Test: Measured TF from table accelerometer to laser sensor and small accelerometer attached on the mirror mount, relocating it to different positions to determine the optimal placement for installation.
• Result: Small accelerometer can be attached to mirror mount with minimal impact on resonance frequency, making it viable for STM measurements. We could see the resonant frequency with/without shaking.

Care must be taken to avoid altering the alignment when attaching the small accelerometer to the mirror mount.
(I will report the measurement result and picture later...)

Good black magic.

The cooling seems to become steady and stable at present. Even the temp of EX_DUCT_SHIELD_BRT1 started decreasing.

[Just a memo]

The operating pressure of the Xer compressor is below.

Before shutdown;

Supply pressure: 24.0 ~ 24.5 bar
Return pressure: 99 ~ 106 psi

(Charging pressure: 16.5 bar)

After restart;

Supply pressure: 24.0 ~ 24.5 bar
Return pressure: 100 ~ 106 psi

We stopped the cryocooler for the BRT at 11:13 (80K at DUCT SHIELD REFBRT4), waited until the DUCT SHIELD REFBRT4 temp reached 100K, and restarted it at 11:51. So it took "38" minutes for heating up. At 13:12, the temp reached 80K again. So it took "87" minutes to cool to the level of the start. (Fig.1)

In the case of EY, it took one hour and 9 minutes from ~78K to 100K at DUCT SHIELD REFARM4 after stopping the cryocooler for the ARM side. So this fact also shows the more heat invasion is expected in the BRT side than ARM side.

During heating up the cryocooler for the EX BRT, CO2 seems to be released according to the mass spectrometer as Fig.2. The max vacuum level is 1.4*10^-5 Pa due to CO2.

I peformed the TCam photo session, I don't know we can take IY image.

DATE : 2025/03/12 23:25 UTC                                                                                                                                            
ARM : Y
TEMP ITM : 91.0
TEMP ETM : 74.1
NUM : 10
IFO REFL HWP : 157.0
PSL HWP : 20.0
IMC POWER : 7.8
VALUE : 1322.4
ERROR : 5.3

Measured data is stored to /users/Commissioning/data/Finesse/Yarm/20250312-232534
 

Is a suspension health check necessary, or should we turn off the cryocooler that shows strange performance??

DATE : 2025/03/12 22:57 UTC
ARM : X
TEMP ITM : 91.0
TEMP ETM : 54.7
NUM : 10
IFO REFL HWP : 160.0
PSL HWP : 20.0
IMC POWER : 7.8
VALUE : 1430.2
ERROR : 7.4

Measured data is stored to /users/Commissioning/data/Finesse/Xarm/20250312-225706

https://gwdet.icrr.u-tokyo.ac.jp/~controls/trendMon/finesse.html

Pictures on 3/12: link

The worming up speed of EX BRT4 becomes higher.

Here are the results of the additional analysis. If the magnitude is larger than 0.5, the curve is colored by magenta.

2/21~23

Figure 1, 3, 4 shows no apparent excitation of resonance.

Figure 2 shows some dumped signals. But, the Q-value (dumping duration) looks different from other dumping signals. The amplitude of the signal looks smaller than other dumping signals. I guess the origin of the dumping signal is different from the others.

3/3~3/5

In Figure 5~7, we can see the dumping signal in 180~185 Hz, which seems to be related to ETMX.  Although the neighbor frequency band has the large amplitude and the timeseries data is colored, it's due to the large tail of the line noise.

Note that on 3/6, the interferometer was not locked.

3/7

In Figure 8, we can see the dumping signal in 165~170 Hz, which seems to be related to ETMY, and 180~185 Hz, which seems to be related to ETMX. It's possible to be a mixture of these suspensions.

3/8~3/11

In Figure 9~12, we can see the dumping signal in 165~170 Hz and 185~190 Hz, which seem to be related to ETMY.
Although the neighbor frequency band has the large amplitude and the timeseries data is colored, it's due to the large tail of the line noise.

With Piernicola-san

Objective

Investigate the impact of clamp configurations and enclosing holders on mirror mount vibrations.

Experimental Summary

1. Clamp Tilting

   • Tested effects of tilting two clamps by adjusting screw heights (from no tilt = 10mm to 6.35mm/5.70mm, then to 0mm).
   • No significant change observed in vibration spectrum.

2. Clamp Angle Dependence

   • Measured with different angles (180°, 90°, 60°) to examine resonance shift.
   • Found that narrower angles lowered the resonance frequency (~50Hz shift).
   • While torque was sufficient, indicating angle alone affects frequency.

3. Enclosing Holder Effects

   • Installed at different heights (80mm, 153.7mm) and measured impact.
   • The spectral changes around 300Hz, but no significant improvement.
   • Removing bottom clamps left only enclosing holders securing the pedestal, leading to instability.
   • In the 2-clamp case, enclosing holders introduced disturbances around 400Hz.
   • A single enclosing holder on the pedestal leg was insufficient; additional screws may improve performance (Virgo doc).

Key Findings

• Clamp Tilting: No effect on vibration.
• Clamp Angle: Alters resonance frequency (~50Hz shift).
• Enclosing Holders: Minor changes observed, but they do not improve stability and may introduce unwanted effects.

Next Steps

• Investigate vibration measurement in the IFI chamber.
  • Consider using PCB 352c68 accelerometer with proper power supply for the measurement of the real STM in the chamber.
  • Test the method with the current in-air setup.