[Kimura]

 At approximately 19:38 on January 11, continuous residual gas measurements using five Q-mass instruments installed in BS, IXC, IYC, EXC, and EYC were stopped.

I turned off the heater at the top of the inner ratioation shield.

I turned off the heater at the top of the inner ratioation shield.

I turned off the heater at the top of the inner ratioation shield.

I turned off the heater at the top of the inner ratioation shield.

I turned off IM heater.

I turned off IM heater.

I turned off IM heater.

I turned off IM heater.

Background

We are working toward the open release of the O4a calibrated reconstruction data.
In parallel with the production of the main dataset, I am also developing verification scripts for the product data.

Guideline

The verification scripts are intended to check the following points:

• Check 1: Verify that there are no missing data in the product data and that the data structure is correct.
• Check 2: Compare appropriate state vectors between the production data and the raw data, and verify that they are consistent with each other.
• Check 3: (Possibly applicable only to the O4a case) Construct the transfer function between the production h(t) and the online h(t), and verify that it behaves as expected.

Current Status

I have completed the scripts for Check 1 and Check 2. For both checks, I have also finished tests using mock data, and the results were as expected.

Detailed Working Logs:

• Work log top page
• Working log for the check 1 script: link1, link2
• Working log for the check 2 script: link1, link2

Next Steps

• Apply the verification scripts (check 1 and 2) to the actual O4a data.
• Develop the script for check 3

Abstract

As reported in klog#35983, I applied whitening filters (S1808683) for PMC monitor channels.
Because expected improvement around kHz band was able to be available for K1:PSL-PMC_PZT_HV_MON_OUT, I finally engaged 1 stage of z1:p10 whitening filter and set z10:p1 de-whitening on K1:PSL_PMC_PZT_HV_MON.
On the other hand, there was no improvement below a few Herz for K1:PSL-PMC_FAST_OUT1_OUT.
And also, whitening gain enhanced glitches during the lock acquisition, so I decided not to engage de-whitening gain for K1:PSL-PMC_FAST_OUT1_OUT today.
I will calculate an effect of whitening filter for K1:PSL-PMC_FAST_OUT1_OUT again and if improvement is really expected, I'll try it again.
Anyway, K1:CAL-CS_PROC_PMC_FREQUENCY_DQ became better estimation of PMC frequency noise than before.
 

Details

klog#35983 suggested that K1:PSL-PMC_PZT_HV_MON_OUT and K1:PSL-PMC_FAST_OUT1_OUT are contaminated by the ADC noise around kHz band and below a few Herz, respectively. I tried to apply whitening filters for these channels. Whitening filter chassis (S1808683) had been already installed at U32 of IOO0 rack for the fast DAQ channels of IMC-CMS (K1:IMC-SERVO_{MIXER,SLOW,FAST}_DAQ_OUT_DQ). IMC related channels were connected to the Board-2 of S1808683 and the Board-1 wasn't used. So I used Board-1 of S1808683 for PMC monitor channels. Before this work, PMC monitor channels were sent from the BNC-Dsub converter (S1605850) at U29 of PSL rack to AA chassis (S1504687) at U15 of IOO0 rack. So I reconnected the cable coming from the PSL room from the AA chassis to the whitening chassis, and added a cable from the whitening chassis to the AA chassis.
 
Fig.1 shows the spectra of K1:PSL-PMC_PZT_HV_MON_OUT_DQ and K1:CAL-CS_PROC_PMC_FREQUENCY_DQ (it's better to use K1:PSL-PMC_FAST_OUT1_OUT, but it's not DQ-ed) before and after engaging whitening filter and gain. Both two signals shows some improvement (Pink curve -> Red curve and Light blue curve -> Blue curve) around kHz band due to the whitening filter for K1:PSL-PMC_PZT_HV_MON_OUT_DQ. But there is no improvement around a few Herz and this fact suggests that whitening gain for K1:PSL-PMC_FAST_OUT1_OUT doesn't work effectively. It may be a mistake in my past estimation. So I will check the noise contamination again.

Even if whitening gain for K1:PSL-PMC_FAST_OUT1_OUT is engaged, there was no problem on the lock acquisition of PMC (and also IMC). But this gain enhanced a glitch during the lock acquisition shown in Fig.2 and large whitening gain made a short-time saturation (see also Fig.3). So I decided to engage only whitening filter for K1:PSL-PMC_PZT_HV_MON_OUT_DQ as the final configuration.

ADC noise projection in the final configuration is also shown in Fig.4. Now PMC control signals were released from ADC noise contamination around kHz band and coherence betweeen the error signal and HV PZT output was drastically improved (Black curve -> Brown curve in Fig.5). Contamination can be seen above 20kHz but it's no problem for seeing signal with 16kHz DQ channels. If we want, we need additional whitening filter chassis for K1:PSL-PMC_MIXER_MON_OUT_DQ which is taken by another Dsub cable.

This is an additional measurement of the laser PZT frequency response, obtained by monitoring the error signal of the locked PMC while injecting an excitation signal to the laser PZT by Moku:Lab.

The figure shows two transfer function measurements. The darker reference line was measured with the excitation amplitude of 2mV, and the bright line was measured with 20mV.

Below a few kHz, that is the UGF of the PMC loop, the TF is suppressed by the control since it was measured at the error point, so the data is available only above several kHz. The flat area between 4kHz and 40kHz corresponds to the former measurement in klog:35884 as 1.54 MHz/V.

One clear resonance can be seen just below 100kHz, but a very complex structure can be seen above 100kHz. First, I suspected that it might be noise, so I compared the different excitation amplitudes. Noise was improved at low frequency with the large excitation, but the structure at high frequency looks the same, so this structure does not come from noise.

This shape is very similar to the previous measurement using IMC, as in this original thread klog:35917. Measurement using PMC as a monitor is a kind of very direct measurement of PZT response using an optical cavity as a monitor. So, this structure above 100kHz is unlikely to originate from measurement noise. 

Figure 1 shows the time series data of laser power and PMC transmission power.
The left and right T cursors indicate times around 2 p.m. and 4 p.m., respectively, which likely correspond to the start and end times of work today.

Since we planned to characterize the laser with low power, the reduction in laser power during the activity is not unusual.
However, the laser power was reduced by 2 W just before finishing the activity, even though the power right before the reduction was almost the same as before the work.

I'm not sure what happened, but was this power reduction intentional?

I offloaded the BF GAS with the FR.

I changed the setpoint for the F2 GAS from 1010 to 1160 so that the MN V Oplev can be zero. The F2 GAS was offloaded with the FR.

I changed the setpoint for the F2 GAS from -1330 to -1258 so that the MN V Oplev can be zero.