Measured TFs from the IP actuator to the ACCs (or FLDACCs) showed increasing gain toward DC below 70mHz for L and 50mHz for T as reported in klog. It is due to the cradle effect. It is difficult to correct the gain using general servo filters. We are considering whether feedforwarding can reduce the gain at low frequencies. I tried to create ACC L signals compensated by the low-passed signals offline. Plot 1 shows the used low-pass filter. Plot 2 shows time-domain data of LVDT, ACC, and low-passed ACC signals. The IP L was excited with the IP actuator during 300~1500sec. Plot 3 shows the ratio of the LVDT TF and the ACC TF. The gain with compensated signals was smaller than the original, but the phase difference was larger.

[Saito, Hirose, Tanaka, Fujimoto]

Abstract

We found that the intensity noise observed with the OMC REFL PD, which will be used for the SRCL/PRCL measurement, can be improved by aligning SRY.
This result suggests that we may be able to lower the cut-off frequency of the SR560 (high-pass) connected to the OMC REFL PD output from the current value of 300 Hz.
This may allow us to reduce the scan speed of the auxiliary laser.

Details

Improvement of the OMC REFL intensity noise by SRY alignment

In the evening, we measured the intensity fluctuation of the OMC REFL PD while the main laser was locked to SRY.
We found that the spectrum had become worse than the one measured yesterday.

In Fig. 1, the blue trace shows the spectrum measured in the evening, and the green trace shows the spectrum measured yesterday.
In particular, the 60 Hz line noise became prominent.

When we lowered the cutoff frequency of the SR560 (high-pass) attached to the OMC REFL PD from 300 Hz to 100 Hz today, a noise comparable in size to the auxiliary laser flashes appeared.
This noise was likely coming from the 60 Hz line noise, since its frequency was roughly around 70 Hz when we checked on the oscilloscope.

We then realigned SRY so as to maximize the AS DC power with ADS and manual alignment of ITMY.
As a result, the OMC REFL spectrum was reduced, as shown by the brown trace in Fig. 1, and became quieter than yesterday’s measurement.

In addition, the 60 Hz power-line noise was significantly reduced, and the noise appears to have moved to harmonics such as 120 Hz.

Fig. 2 shows the evolution of the OMC REFL spectrum during the process of improving the alignment. It can be seen that the intensity noise was gradually reduced.

Investigation of the origin of the 60 Hz power-line noise

Since the 60 Hz power-line noise was improved by the SRY alignment, mainly in the pitch direction, and the component moved to harmonics, it is suspected that one of the suspensions related to SRY is moving at 60 Hz in pitch.

Therefore, I measured the pitch and yaw spectra of the related suspensions and looked for suspensions showing a 60 Hz line.

The result is shown in Fig. 3. A 60 Hz component can be seen in the pitch spectrum of SR3. Since this component became invisible in the OMC REFL signal after improving the SRY pitch alignment, SR3 pitch motion is likely the cause of the 60 Hz noise seen in OMC REFL.

Plans for tomorrow

We plan to reduce the high-pass cutoff frequency of the SR560 attached to the OMC REFL PD as much as possible while keeping the OMC REFL intensity noise low by using the improved SRY alignment. This is expected to allow us to reduce the scan speed of the auxiliary laser.

And the current UGF of the PLL is about 20 Hz. We will try to increase the control bandwidth, for example by removing the 20 dB attenuator placed after the PFD, so that the auxiliary laser can follow a faster LO scan.

[Tanaka, Fujimoto, Saito]

When the LO frequency was adjusted relative to the beat frequency so that the error signal became close to 0 V, it was found that the beat frequency and LO frequency matched at higher frequencies, whereas an offset appeared at lower frequencies. Next, in order to achieve locking using only the SR560, the SR560 was configured as a first-order low-pass filter with a cutoff frequency of 1 Hz and a gain of 200. Under this condition, locking was successfully achieved. Measurement of the open-loop transfer function showed that the UGF was approximately 17 Hz. Next, we attempted to use the FM modulation function of the LO source (Keysight E8663D) in order to sweep the LO frequency and scan the SRY. However, the FM modulation option was not installed. Therefore, the function generator of the Moku:Lab was used instead. When the frequency modulation was performed at a rate of 1 Hz, no SRY flashes were observed on the OMC REFL PD. This is likely because the frequency modulation was carried out at 1 Hz, whereas the cutoff frequency of the high-pass filter in the SR560 used for the OMC REFL PD signal was 300 Hz. Therefore, the cutoff frequency was lowered to 100 Hz, but the main-laser noise became comparable to the amplitude of the sub-laser flashes, so the cutoff frequency was restored to 300 Hz. In addition, increasing the frequency modulation rate of the LO caused the beat signal waveform to become distorted.
 

• First, the LO frequency was adjusted relative to the beat frequency so that the error signal became close to 0 V. The resulting frequencies were as follows:

  Beat frequency    LO frequency
  33 MHz                    54 MHz
  78 MHz                    85 MHz
  139 MHz                 139 MHz
  179 MHz                 179 MHz

  Therefore, at higher frequencies there appears to be no offset, and the beat frequency matches the LO frequency. Furthermore, at lower frequencies, the beat signal and LO signal were directly observed using the Moku:Lab oscilloscope. The measured beat frequency agreed with the frequency observed on the Moku:Lab spectrum analyzer after the signal had been split by the power splitter, and the LO frequency agreed with the set value. Therefore, the observed offset appears to originate from the PFD.
   

• Next, a DC voltage was applied to the SR560 to determine its allowable input and output voltage ranges before overload occurred. The maximum allowable input voltage was found to be 2 V, while the maximum output voltage was 5 V. When the error signal was near 0 V, observation of the error signal with a 100 kHz low-pass filter showed fluctuations of approximately 4 Vpp. Therefore, a 20 dB attenuator was inserted to ensure that the signal could be safely input to the SR560. Furthermore, when the error signal was 150 mV, the beat frequency and LO frequency differed by approximately 10 MHz, corresponding to a sensing efficiency of approximately 15 nV/Hz. To achieve locking using only the SR560, the SR560 was configured as a first-order low-pass filter with a cutoff frequency of 1 Hz and a gain of 200, and stable locking was achieved. The current control loop from the RFPD output to the sub-laser PZT is as follows:

  RFPD→ 12 MHz high-pass filter→ 20 dB RF amplifier→ 45 dB RF amplifier→ 10 dB attenuator→ PFD→ 20 dB attenuator→ 100 kHz low-pass filter→ SR560 (1 Hz cutoff frequency, gain ×200, first-order low-pass filter)→ Sub-laser PZT

  The open-loop transfer function was then measured, yielding a UGF of approximately 17 Hz (Fig. 1).
   

• Next, in order to scan the SRY by sweeping the LO frequency, we attempted to use the FM modulation function of the LO source (Keysight E8663D). However, this option was not installed. Therefore, a 1 Vpp sinusoidal signal was generated using the Moku:Lab function generator and used as the LO signal. The LO frequency was first matched to the beat frequency and then frequency-modulated by ±2 MHz at a modulation rate of 1 Hz (Fig. 2). However, no SRY flashes were observed on the OMC REFL PD. This is likely because the modulation rate of 1 Hz is below the 300 Hz cutoff frequency of the high-pass filter in the SR560 used for the OMC REFL PD signal. The cutoff frequency was therefore reduced to 100 Hz, but the main-laser noise became comparable in magnitude to the sub-laser flash signal, so the cutoff frequency was returned to 300 Hz. The dominant main-laser noise frequency was approximately 60 Hz. In addition, the frequency modulation rate of the LO was increased, but the shape of the beat signal became distorted.

[Ushiba, Smith, Tanaka, Fujimoto, Saito]

First, the appearance of the sub-laser beam on the OMC REFL camera was examined under the single-pass condition. The mode shape appeared elongated in the horizontal direction. Next, to avoid saturating the OMC REFL PD with the main-laser light, the PD gain was reduced from 40 dB to 0 dB. The PD output was connected to an SR560, which was configured as an AC-coupled second-order high-pass filter with a cutoff frequency of 300 Hz and a gain of 100. When the sub-laser PZT was driven with an 8 Vpp, 500 Hz triangular wave, the noise from the main laser and the amplitude of the sub-laser fringes were comparable, making it difficult to distinguish the fringes. However, when the ISS was turned on, the noise from the main laser was reduced, and by lowering the PZT drive frequency to 170 Hz, the sub-laser fringes became clearly visible. Therefore, it appears feasible to scan the SRY. Next, the ND filter in the sub-laser path of the PLL setup was changed from OD = 0.5 to OD = 0.6. After realigning the PLL path to maximize the beat signal, the beat signal amplitude reached approximately 4.82 dBm. When the LO frequency was tuned to the beat frequency, the PLL successfully locked. Furthermore, when the LO frequency was varied, the beat frequency followed accordingly, confirming that the lock was functioning. However, even while locked, the beat frequency sometimes drifted spontaneously by several MHz. Therefore, the loop filter parameters will need to be adjusted to suppress this motion, and the open-loop transfer function will also be measured. In addition, a piezo driver will be introduced to drive the PZT over a wider frequency range.
 

• First, the main laser was turned off and the SRM was misaligned to observe how the sub-laser beam appeared on the OMC REFL camera in the single-pass configuration. The mode shape appeared horizontally elongated (Fig. 1). This is likely because the beam was hitting the lower part of a lens that had originally been installed in the optical path. Next, when the SRM was aligned, the fringe amplitude was approximately 2 μW (Fig. 2). Since this was comparable to the result obtained previously (klog:37116), the alignment of the sub-laser was not further adjusted.
   

• Next, the main-laser power measured at the OMC REFL PD under the single-pass condition was approximately 1.95 mW, while it increased to approximately 6 mW when the SRM was aligned. To avoid saturating the OMC REFL PD with the main-laser light, the PD gain was reduced from 40 dB to 0 dB. Under this condition, when the main laser was resonant in the SRY, the PD output was approximately 3600 counts, corresponding to approximately 2.1 V, allowing the PD output to be connected directly to the SR560. The SR560 was configured in AC-coupled mode, and the cutoff frequency of the high-pass filter was chosen while observing the OMC REFL PD signal with only the main laser present so as to minimize signal fluctuations. The gain was then increased while ensuring that the SR560 did not overload. Next, only the sub-laser beam was injected and it was confirmed that fringes could be observed when the sub-laser PZT was driven. The main-laser beam was then reintroduced. With the SR560 configured as a second-order high-pass filter with a cutoff frequency of 300 Hz and a gain of 100, and with the sub-laser PZT driven by an 8 Vpp, 500 Hz triangular wave, the noise from the main laser and the sub-laser fringe amplitude were of comparable magnitude, making fringe identification difficult (Fig. 3). However, after turning on the ISS, the noise from the main laser was significantly reduced. Furthermore, by lowering the PZT drive frequency to 170 Hz, the sub-laser fringes became clearly visible (Fig. 4). The power spectra shown in Figs. 3 and 4 correspond to K1:OMC-REFL_DC_IN1. In Fig. 4, the blue trace was taken while the sub-laser PZT was being driven, while the red trace was taken without driving the PZT. Comparing Figs. 3 and 4 clearly shows that the ISS reduced the noise level. Therefore, scanning the SRY appears to be feasible.

• Next, the optical powers measured immediately before the RFPD in the PLL path were 45.5 μW for the main laser and 2.122 mW for the sub-laser. Since the sub-laser power might have been too high, the ND filter in the sub-laser path was changed from OD = 0.5 to OD = 0.6. The sub-laser power was then remeasured and found to be 1.349 mW. The PLL optical path was then realigned to maximize the beat signal (Fig. 5), resulting in a beat signal amplitude of approximately 4.82 dBm. Using a filter similar to that employed in klog:37052 and setting the LO frequency equal to the beat frequency, the PLL successfully locked. Furthermore, changing the LO frequency caused the beat frequency to follow accordingly, confirming proper PLL operation. Next, an attempt was made to scan the SRY by sweeping the LO frequency. However, clean fringes such as those shown in Fig. 4 were not observed. In addition, even while the PLL was locked, the beat frequency occasionally drifted spontaneously by several MHz, indicating that further adjustment of the loop filter is necessary to stabilize the lock. The open-loop transfer function will also need to be measured. Furthermore, a piezo driver will be introduced in order to drive the PZT over a wider frequency range.

I compared the spectra in the IRM damper servo ON/OFF. The IP was excited in yaw with the IP actuators. Plot 1 shows the feedback signal to the IRM actuator (upper) and the TM Oplev yaw signal (lower). Plot 2 shows the spectra calculated from the 0-150 sec data (servo ON) and the 180-330 sec data (servo OFF). The servo gain was 1.5 times larger than the gain at the TF measurement. The peak around 60mHz was damped by the servo.

I modified the LOCKING_SRC_3F state in the VERTEX guardian.
Followings are the summary of the modification.

1. Changed the ramp time from 5 to 1 when engaging SRC lock.
2. Added check functions to ensure that the SRC is properly locked at counters 1 and 2.

#1 is for avoiding lockloss before engaging integrator.
#2 is for avoiding to go DOWN state once when the SRC is locked with wrong resonant conditions.

[Smith, Tanaka, Hirose, Fujimoto, Saito]

The sub-laser beam was aligned to the main laser beam, and after blocking the main laser, SRY fringes were observed using the OMC REFL PD. The alignment was optimized using two mirrors to maximize the fringe amplitude. The measured FSR was approximately 2.3 MHz. Since this value is very close to the expected FSR, we conclude that the observed signal corresponds to SRY fringes. Measurements of the main laser and sub-laser beam profiles showed that the beam waist positions differed by approximately 65 mm. Therefore, a beam profiler was placed at the main laser waist position, and the positions of the two lenses were adjusted so that the sub-laser beam waist would be located there. After blocking the main laser again and observing the SRY fringes with the OMC REFL PD, the alignment was optimized using two mirrors. As a result, the fringe amplitude increased to approximately 1.3 μW. However, the mode shape was not circular. In addition, both the main laser and sub-laser beams were found to be incident on the lower part of a lens that had originally been installed in the optical path. Next time, we plan to further adjust the positions of the two lenses to determine whether the fringe amplitude can be increased. We are also considering adjusting the height of the existing lens so that the beams pass through its center.

• First, the alignment was performed using two irises. After blocking the main-laser beam, SRY fringes were observed using the OMC REFL PD. The alignment was then optimized using two mirrors to maximize the fringe amplitude (Fig. 1). The sub-laser PZT was driven with an 8 Vpp triangular waveform. In Fig. 1, the upper trace of the ndscope display shows the OMC REFL PD signal, while the lower trace shows the signal applied to the PZT. From the plot, an 8 V change corresponds to approximately 6.5 FSR. Since the PZT efficiency of the sub-laser measured at Kashiwa was 1.871 MHz/V, the FSR is estimated as FSR=(1.871 MHz/V)×(8 V)/6.5​=2.3 MHz. Because this value agrees well with the expected FSR, we conclude that the observed signal is indeed the SRY fringe.
   

• Next, the beam profiles of the main laser and sub-laser were measured to evaluate the mode matching. The sub-laser beam profile remained distorted, similar to what was observed previously (klog:37093). When the beam profile was examined from upstream toward the waist, the beam initially had a reasonably clean shape (Fig. 2), then gradually became distorted into a vortex-like pattern (Fig. 3), and finally returned to a cleaner shape near the waist region (Fig. 4). The measured beam profiles were then fitted (Fig. 5). The resulting waist positions and waist radii are listed below. The origin is defined as the position of the mirror immediately after the BS in the path through which the sub-laser enters the interferometer.

  Main laser

  x direction: Waist position = 303.2 ± 8.8 mm, Waist radius = 0.0545 ± 0.0022 mm
  y direction: Waist position = 330.3 ± 10.6 mm, Waist radius = 0.0609 ± 0.0026 mm
  →Average: Waist position = 317 mm, Waist radius = 0.058 mm

  Sub-laser

  x direction: Waist position = 255.8 ± 5.2 mm, Waist radius = 0.0528 ± 0.0014 mm
  y direction: Waist position = 247.8 ± 5.3 mm, Waist radius = 0.0548 ± 0.0015 mm
  →Average: Waist position = 252 mm, Waist radius = 0.054 mm

  From these results, the mode-matching ratio was estimated to be approximately 7.5%. Since the waist positions differed by about 65 mm, the beam profiler was placed at the main laser waist position, and the positions of the 200 mm and 150 mm focal-length lenses were adjusted so that the sub-laser beam waist would coincide with that position. After blocking the main laser and observing the SRY fringes with the OMC REFL PD, the alignment was optimized again using two mirrors. The fringe amplitude became larger than that shown in Fig. 1 (Fig. 6), reaching approximately 1.3 μW. However, as shown in the camera image in the lower-left corner of Fig. 6, the mode shape was not circular. Therefore, next time we plan to further adjust the positions of the two lenses and investigate whether the fringe amplitude can be increased further.
   

• In addition, the position of the sub-laser beam was checked at the location of the lens that had originally been installed downstream of the iris. The beam was not passing through the center of the lens (Fig. 7). The same was true for the main laser beam. It is possible that the fact that the sub-laser beam enters the interferometer while passing through the lower part of the lens is affecting the observed performance.

I measured the open-loop gain of the IRM damper servo in both cases where the loop is open and closed. The maximum gain was about 3, the phase margin at the UGF was about 70 deg, and the gain margin at 160mHz was about 2. The second plot shows each transfer function in the closed loop.

[Kimura and Nakagaki]
 On 22 June, M. Takahashi-san reported the #39 ion pump power supply of X-end had stopped again after displaying an error with "Eroor code 20".
(See Photos 1 and 2)
The pressure gauge shows “7.5×10⁻⁶ Pa,” leading us to conclude that the ion pump power supply did not shut down due to a pressure rise caused by a vacuum leak.
 At around 2:30 p.m., we closed the gate valve between the vacuum duct and the vacuum pumping unit to replace the ion pump power supply.
We then replaced it with a spare power supply and high-voltage cable, and restarted the ion pump power supply.
After startup, the pressure inside the T-tube reached “2.7×10⁻⁶ Pa,” so we opened the previously closed gate valve to resume evacuation by the ion pump.
(See Photos 3 and 4)
The pressure after opening the gate valve was “5.5×10⁻⁶ Pa”.
(See Photos 5 and 6)
  

Abstract:

I implemented SRM GAS modal damping.
The modal damping will be engaged automatically when SRM is at LOCK_ACQUISITION state.

Detail:

I made mistakes in the calculation of actuator matrix, so I calculated it again.
The new modal actuator matrix is shown in fig1.

After fixing the actuator matrix, I measured the suspension transfer functions from M1, M2, and M3 actuators to M1, M2, and M3 sensors, respectively.
Figures 2, 3 and 4 show the measured transfer functions, which appear to be well decoupled.

As discussed in klog36787, the best performance can be obtained by engaging only M1 modal damping and IMV damping.
So, I designed M1 damping filters and implemented it at FM1 of K1:VIS-SRM_GASMODAL_DAMP_M1 filter bank.
Figure 5 shows the OLTF of M1 damping loop.

To implement the modal damping control in the guardian, I also designed filters for K1:VIS-SRM_{F0,F1,BF}_DAMP_GAS aiming to achieve a DC control UGF of 10 mHz.
These filters are implemented at FM3 of each filter bank.

Figure 6 shows the spectrum of GAS filters, IMV OSEM, and OpLevs with modal damping (red) and at current LOCK_ACQUISITION state (blue).
The noise at high frequency becomes better, especially in IMV OSEM, while keeping the spectrum at low frequency.
So, I implemented the new controls in the guardian, and SRM GAS modal damping is now automatically engaged when SRM is at LOCK_ACQUISITION state.

[ Kimura , Oshino,  Nakagaki ]

We have installed a system to monitor the open/closed status of the manual gate valve between SR3 and SRM.
The open/closed status is provided via the following PVs:

K1:VAC-GV_SRM_OPEN
K1:VAC-GV_SRM_CLOSE

These have also been added to the "VAC_OVERVIEW" MEDM screen.  Klog#37109

I offloaded the F0, F1, and F2 GAS filters with the FRs.

[Smith, Tanaka, Fujimoto, Saito]

The sub-laser beam was aligned to the main-laser beam, and after blocking the main laser, flashes from the SRY were observed using the OMC REFL PD. The alignment was further optimized using two mirrors to maximize the fringe amplitude, and the positions of the two lenses were also adjusted. However, the resulting fringe amplitude was smaller than that observed previously (klog:37065). In addition, the signal level in the single-pass configuration was equal to the minimum value of the fringe signal. The mode shape observed with the camera also appeared as if the beam was being clipped somewhere in the optical path. Furthermore, when the beam position was checked in front of the iris located before the sub-laser beam entered the SRY, the beam was found to be offset from the center, even when the fringe amplitude had been maximized. Therefore, it is possible that the alignment was not properly optimized.
 

• First, the alignment was performed using two irises, and the main-laser beam was blocked. When the OMC REFL PD was used to monitor flashes from the SRY, flashes were successfully observed. The alignment was then further optimized using two mirrors to increase the fringe amplitude. The positions of the 200 mm and 150 mm focal-length lenses were also adjusted to maximize the fringe amplitude. However, the resulting fringe amplitude was smaller than that obtained previously (klog:37065) (Fig. 1). Next, the SRM was misaligned to observe the single-pass signal. The signal level was found to be equal to the minimum value of the fringe signal. The HWP located after the FI was then rotated to check whether the signal level could be increased, but the signal was already at its maximum value. The SRM was subsequently realigned, and the spatial mode was observed with a camera. The mode shape appeared as though the beam was being clipped somewhere along the optical path. In addition, after optimizing the alignment to maximize the fringe amplitude, the beam position was checked in front of the iris located before the sub-laser beam entered the SRY. The beam was found to be displaced from the center of the iris. Therefore, it is possible that the alignment was not properly optimized.

Similar work with klog36755, klog36780, klog36781, and so on.

I constructed the sensor and actuator matrices for GAS modal damping of SRM.
Figure 1 and 2 show the sensr and actuator matrices, respectively.
I will test them soon.

Details are summerized in klog #36941.