At about 5:40pm yesterday, the EtherCAT system stopped communicating with BIO chassis 1&2. Inspecting the chassis in the CER, chassis 1 has an LED in the back indicating that +24V is off, whereas the LED in the front indicates that +24V is on. Things to try:

• Power cycle the chassis
• Check the Ethernet cabling
• Power cycled the EtherCAT chassis
• Replace with spare

I've taken LLO's large DAQ overview MEDM screen and converted it for LHO. I've macrotized the front end blocks and given them roughly the same layout as the main CDS Overview MEDM (attachment shows them side-by-side). The old DAQ detail screen can be opened using the lower-left button.

Matlab 2015b on the workstations will now automatically load the nds2 client.  No more need to do javaaddpath(...).

[Jeff K, Stuart A]

IIET: #11730

ITM QUAD suspensions at LLO have exhibited sharp ring-ups at 3.3 Hz from ASC pitch drive resulting in IFO lock-losses (see for example LLO aLOG entries: 40926, 40953, 40972 and 41501), as a work-around LLO currently employs a "TWO_LEGGED_HARDASC" scheme reducing ASC HARD drive from 4 QUADs to 2 QUADs i.e. just feeding back to the ETMs (see LLO aLOG entry 41502). The 3.3 Hz feature corresponds to the QUAD R3 Mode, which when looking at QUAD L2 P to P Damped TFs (see allquads_2019-01-30_AllSUSQUAD_Phase3b_L2_Damped_ALL_TFs.pdf below) which we have confirmed is also prominent in LHO QUAD suspensions too. However, at present, the 3.3 Hz R3 modes are not being rung-up at LHO.

Given there has been no change since O2 in the LLO QUAD local damping scheme, which demonstrates equally ineffective damping of the 3.3 Hz R3 mode as the LHO QUAD local damping scheme, thus, we believe there has been a change in the global control at LLO. This was not an issue at LLO during O2, so we suspect that changes in the ASC loops to handle higher IFO power have exacerbated the 3.3 Hz ring-up problem.

We found that the 3.3 Hz R3 mode is not well modeled in the QUAD dynamical model (see L1ITMX_Model_L2_PtoP below) which also suggests that the R3 mode should be completely damped n.b. we imported L1 ITMX live damping filters.

Therefore, we are now attempting to adjust QUAD dynamical model parameters to better fit the measured QUAD TF data. Failing this approach, we may seek to resurrect the QUAD model fitting code T1100163.

Here's what the overall noise looks like between the two scheme

This is H1 scheme, with 20dB attenuator on 80MHz LO swapped for 10dB, same CLF power used when we measured 4dB (with PSLLO):

And after we swapped, here's L1/original scheme:

One good news is that I don't think we will see OPO length noise in the IFO anymore as LO is not seeing any of it.

Looking closely at LO phase noise, I don't think this tells the whole story as previous sqz/asqz measurement with L1 scheme appeared to do (just) a little worse. In this plot L1 scheme LO phase noise seems to do much better.

Daniel found CLF correlation < 1kHz with the accelerometer on ISCT6 (alog46712). I'll look at intensity noise next to try to explain noise >1kHz.

See attachment for transfer functions for both noise plots and L1 scheme noise in mrad.

ETM HWS do not show strong evidence of 1064nm leakage but show some systematic variation in intensity related to IFO state (~1% on ETMX-HWS and ~2% on ETMY HWS).

ETMY

ETMX

I wrote a script to compute the correct inspiral range based on a transfer function measurement using Pcal to CAL-DELTAL_EXTERNAL. I used a recent measurement (see LHO aLOG 46530) to test this script. 

In summary, based on this measurement, the correct inspiral range (if Pcal is indeed considered the true value), then the inspiral range is 89.3 Mpc. By comparison, the original computed inspiral range was 85.3 Mpc.

The script is stored here:
$CALSVN/trunk/Common/pyDARM/correct_bns_range.py

The important caveats are:
1) the installed front end inverse sensing filter must be the same as the model file/function that you are using, and you have to point the script to this parameter file (meaning you have to know what was used to install the front end inverse sensing filters)
2) while one can put any GPS time to get a DARM spectrum, the correction is really only valid near the time of the Pcal to DARM sweep
3) the transfer function and coherence data from the DTT template has to be exported to ASCII files

I attach two figures to illustrate this code on the recent measurement link above.
1) The transfer function data between Pcal and CAL-DELTAL_EXTERNAL to measure the correction needed
2) 4-panel plot showing the correction to the DARM spectrum

All plots seem to be nominally ok.

TITLE: 01/31 Day Shift: 16:00-00:00 UTC (08:00-16:00 PST), all times posted in UTC
STATE of H1: Commissioning
OUTGOING OPERATOR: None
CURRENT ENVIRONMENT:
    Wind: 3mph Gusts, 2mph 5min avg
    Primary useism: 0.02 μm/s
    Secondary useism: 0.23 μm/s
QUICK SUMMARY:

Interferometer is unlocked and fundamental violin Mode 1 is rung up rather high. Efforts to dampen will begin shortly.

First I tried implementing the changes Sheila suggested for SRC2 P, then remeasured DHARD P with a gain of -60.  The result is the black line in attachment one.

Jenne and Satoshi rebalanced the arm ASC output matrices in alog 46707.  Unfortunately in DARM_TO_RF we hardcoded the old DHARD outrix (attachment two), and this is likely how we've been operating ever since Wednesday morning.  I was able to change the outrix in lock no problem.  I lost lock when trying to redo the DHARD P measurement from before, but from the one average I got it seems like this did not make a difference to our 1 Hz DHARD P gain margin.

I spent some time gathering TFs related to the DHARD P loop to try and figure out who is the most relevant culprit is at 1 Hz.  This arose from a suspicion about the radiation pressure compensation and regular filters sum, but this cannot be the issue since the RPC digital filters barely affect us with -60 gain on the regular filters.
Attachment three shows the ASC-DHARD_P and ASC-RPC_DHARD_P filters, alongside the measured OLG I took, and a quad sus L2Pitch->L3Pitch TF and L3Pitch->L3Pitch TF stolen from Hang.
The OLG is well-colored by the total digital filter TF, except around 1 Hz.

------------------------------------------------------------------------------------------------------------------------------------------

Twice this evening I lost lock in POWER_30W, and when the IFO went down the IMC could not relock itself.  It just went to from LOCKED->DOWN->ACQUIRE and continually failed to acquire for up to twenty minutes.  This only happens when the IMC is super calm, so I added an automatic toggling of the MC2 M1 L2L DriveAlign offset every 15 seconds in the IMC_LOCK guardian.  Not clear if this helped, but probably did not hurt.

Damped ITMX Mode 2.  Rotated +60 degrees from where the guardian put it.  The violin modes guardian does not tolerate dissent, so I had to pause it.
- The mode bounced back and killed the lock.  I tried damping it for about an hour with no success.
- In the next lock I just reduced the gain from -5 to -1 with the same phase (-60 degs), this seems to have helped, but the mode is damping at a glacial pace...

Calum, Daniel, Rich

An analysis was made of the phase noise content of the VCO serving the as Squeezer CLF actuator.  A comparison is included below of the VCO nominal parameters from E1700010.  Attached is a plot of the single sideband phase noise.

Specification
10Hz -110dBc
100Hz -140dBc
1kHz -160dBc
10kHz -165dBc

Measured
10Hz (not measured)
100Hz -132dBc
1kHz -145dBc
10kHz -160dBc

As can be seen from the measured vs the specification, the noise is a bit higher than expected.  This is still far below the measured CLF noise, but needed to be checked.  The data was compensated for the fact that 2 identical oscillators were beat together during the measurement (3dB correction), and compensated for single sideband (3dB correction).  During the course of the measurement, the slope of the phase detecting mixer was determined to be 1.1268 volts per radian.

Georgia, Adam, Keita, Marie, Sheila, Jenne

This afternoon we found that the SRC2P setting that made DHARD P stable (probably just barely) last night didn't work today.  The first attached screenshot is a measurement of the DHARD P loop with higher resolution than last night, showing that even with the SRC2P gain reduced by a factor of 15 it still has an impact on the DHARD P loop at 1 Hz.  (first screenshot)

Qualitatively, since the AS_C QPD is sensitive to DHARD when there is a DARM offset, because of the interference between the carrier 00 and the carrier 01 mode, the SRC2P closed loop response has an impact on the DHARD P loop (because the ASQ sensor is also sensitive to SR2 motion because of the DARM offset). 

We then spent some time measuring the SRC2 P loop, both in DRMI and in full lock.  With the settings we used last night (green traces in 2nd and 3rd attachment), with a gain of 1 and a compensation filter for the HSTS engaged, the closed loop response of the SRC2P loop has a wiggle in phase at 1 Hz, although the open loop gain is around -20dB at 1 Hz.  We changed the loop to have a ugf around 0.04 Hz, removing the HSTS compensation and adding a low pass at 0.7 Hz to reduce the phase wiggle in the open loop response at 1 Hz.  

We have not put these new settings in the guardian, but we would like to try to engage the SRC2 P loop with the gain of 10, and FM1,3,4,5,7 to see what the impact of this on DHARD is.  

This may not stabilize DHARD, but at least it will remove one cross coupling that is causing the loop shape to change around the zero phase margin frequency.  

Conclusion:

If our goal is to keep the residual angular motion has a roughly constant rms, then we should try to reduce the ASC bandwidth as we power up, which would further allow us to lower the cutoff frequency to improve DARM. 

Detail:

The closed-loop torque to angle transfer function can be written as 

    G_cl = S_rad / (1 + K*S_rad),

where S_rad is the L3 suspension torque to angle transfer function and K is the control filter.

In the attached plot we show S_rad for the hard pitch mode at different input levels (to get arm power, we assumed G_p=42 and G_a=265). We can see that the DC level of S_rad decreases as we power up. 

Now we focus on G_cl at around the microseismic frequency of ~ 0.2 Hz. It dominates the residual rms motion and it is below the sus resonances and the loop UGF. Consider two scenarios:

    1). we keep the overall open-loop gain, (K*S_rad), roughly the same at different power levels. Then this means

        G_cl ~ S_rad, which decreases as the power increases. 

        Since the raw seismic input in torque is independent of input power, as G_cl decreases with S_rad, we are clearly "wasting" control bandwidth by trying to keep (K*S_rad) the same. This cause us inject more sensing noise to the system then we need to. This is not optimal.

    2). We should instead keep the loop suppression the same. Note that G_cl ~ S_rad/(K*S_rad) ~ 1/K at microseismic freq. As we power up, we should not increase the DC gain of K but just keep it roughly constant. This means the UGF of the CH/DH loops should decrease as S_rad decreases. 

TITLE: 01/30 Day Shift: 16:00-00:00 UTC (08:00-16:00 PST), all times posted in UTC
STATE of H1: Commissioning
INCOMING OPERATOR: None
SHIFT SUMMARY:
LOG:

15:50 Chris S out to MY for scaffold setup

18:00 Marc out to MX for LEMI signal testing

18:29 Marc back

18:38 Aiden out to optics lab

18:50 Betsy out to LVEA W Bay to get something

19:29 Betsy back

19:56 - 20:23 Nutsinee out to LVEA - ISCT6

21:00 Nutsinee back

23:39 Chris back

23:00 Weekly Site Meeting

I tried more tests to investigate the 1.1Hz HAM3-ISI line, following up on Jim and Hugh's effort. I am keeping track of the tests done so far in this document. One of the ways to remove the peak is to turn off the Ry isolation loop, at the expense of low-frequency horizontal motion, see the X motion before vs after, or Ry motion before vs after.

With low enough useism, the ifo was locked in this configuration to test if this would change the 1Hz ASC behavior. The two problems look independent since we found ASC to be still marginally stable around 1Hz but with no peak present on the ISI, see attached screenshot.

Will keep investigating.

On January 8th I tried to measure the average arm power by dithering SRCL.  Last night I did it again, this time with better TFs taken simultaneously.

Average Arm Power = 143.0 +- 5.8 kW

This time I directly fit each of the four Arm Trans QPD RINs (H1:ASC-{X,Y}_TR_{A,B}_NSUM_OUT_DQ) to DARM (H1:CAL-DELTAL_EXTERNAL_DQ), then fit a 1/f2 (attachment one).  The DARM calibration used was the stop gap made by Keita.  The average powers for the RIN calculation are as follows, and are good to +- 300 cts:
Arm Trans QPD counts
XA      145658
XB      191409
YA      207723
YB      221769


This time the math is simpler: 
P_arm = DARM/ArmTransRIN Fit × π2 × m × c


If we look at the different in the arm trans RIN response between the arms, and say that the RIN we see is entirely due to radiation pressure in the arms, we get (DARM/Y)/(DARM/X) = X/Y = 1.08, meaning there may be 8% more light in the X arm than the Y arm.  EDIT: It is actually the Y arm with 8% more light than the X arm.

If we calculate the simple power recycling coupled cavity, I find that the estimated average ETM scatter losses are around 92 ppm and the estimated PRG here is 39.  (attachment three)

The measured value of PRG at the time of the measurement is 41.8, and the true input power was 26.5 W (requested was 30W).  If we take these numbers to be true, then our arm gain according to the SRCL dither is 258.

Daniel, Nutsinee

We will later be using this information to calculate the overall phase noise contribution from CLF and what is the best power to operate when injected into the IFO. But for now these plots are just showing noise at various CLF power. At 1X power (0.5 uW CLF transmitted) we are mostly limited by sensing noise. At 10X (4uW transmitted) and 60X (20uW transmitted) power we cleared the sensing noise for the most part. The UGF for 10X and 60X measurement were chosen where the phase margin is ~30deg. The UGF for 1X measurement were chosen simply because we begin to be sensing noise limited right around there. In all case the suppressed CLF signal hangs around 3mrad. 

The RF amplifier (minicircuits ZFL-500LN) was taken out during 60X power measurement. The calibration factor is a bit lower than what we would expected even taken the factor of 25 of the RF amp into account. That part is still unclear why. The manual suggests a gain of ~25 at 3MHz.

The shotnoise was calibrated using a halogen light bulb. By subtracting DN from combined DN SN in quadrature we recover SN which we use to calibrate the RF transimpedance. 

Sometimes it seems like the FSS gets noisy, and I don't think anything else is happening in the IFO.

I caught it happening again, and am posting a quick time series and spectrum to show that it's definitely changing.  Does anyone in PSL-land know what this might be?  Perhaps (as Sheila suggests) someone could take a look at PSL PEM sensors to see if anything changed in there?