TITLE: 12/20 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:

Rough day with big earthquake which kept us down for first half of the day.  Then we had an EX timing crash which has kept us down for a good chunk of the afternoon.
LOG:

• DRMI & PRMI didn't look good (surprising after quick locking rest of night).  Will need to run alignment.
• 16:20 Cleaning in LVEA & YARM (Vanessa & Karen)
  • 16:50 Yarm complete (Karen)
• 16:28 HAM2/3 ISI transfer functions
• 17:10 EQ alarm for 7.4 at Russia.  Jim took SEI_CONF to EARTHQUAKE + manually requested SC_OFF for ITMy, ITMx, ETMy Stage 2's (trying to avoid WD trip).
• 17:12 Praxair arriving for CP1 fill (planned & purely coincidence with EQ)
• 17:14 Bug fixing by Jenne (since earth is noisy)
• 17:32 - 18:10 NDscope update (Jonathan)
• 17:48 Fiber measurements near HAM1/PSL & out at EX (Fil)
• 18:29 SUS ETMx Model & DAQ restart (Dave, Jenne)
• 18:41 Checking Power supplies (Richard)
• 18:44 Visiting LVEA for VAC computers (Gerardo)
• 18:52 Checking HAM2 Camera access (Cheryl, Charlie, Jason)
• 19:04 Xmid new ion pump (Gerardo)
• 19:00 PR2 Measurement (Hang)
• 19:25 EX ESD grounding toggle off/on (Richard)
• 19:45 EY ESD grounding (Fil)
• 19:45-22:45 Checking HAM2/3 cameras (Cheryl)
• 21:00 Running Alignment
• 21:36 EX went down "timing glitch/error"---ETMx had frozen M0 channels, & this tripped the ISI.
  • No body was down there when this happened.  (I was finishing up alignment....did get connection errors for SRC OFFLOAD
  • Kill models
  • "Restart the world" by Dave
  • Check Alignment sliders for ETMx & TMSx
  • 21:50 Dave out to EX
  • 22:30 Dave discovered we are missing a DAQ at EX.  Will need to go out to power cycle again.
• 22:00 Removing transformer (Gerardo)
• 22:02 CER hole plugging (Fil)
• 22:05 ISCT6 work (Nutsinee)

Handing off to Niko for last hour (have an appt in town).

After disconnecting some extension cords that were not being used, the unit was powered off at 22:00 UTC.

WP8022

Danni and Dave:

h1tcscs epics parts recording the inputs and outputs of the ring heater inverse filters had name issues which were fixed. After the model and DAQ were restarted, the data concentrator would not start reporting a name duplication for the EpicsOut part which was named <filtername>_OUT. My check_daq_channels_validity script reported no duplicated channels.

Turns out daqd was correct and my script was in error because we had chosen an EPICS channel name which coincided with a test point channel name. My script was just checking for duplications within the INI and PAR file groups separately, I'll fix it to check the full channel list.

Following Robert's post about fan lines in the Darm spectrum from EX Fil and I went down and investigated cabling.  I did not like the grounding scheme on the new +-18/48VDC power supplies so we changed the connection.  The supplies are now hard wired together through the returns and grounded.  Other supplies are done this way but they only supply 2 voltages not 4.  Will look for improvement or not.

While poking around looking at PR SUS channels this morning, I see that the M3 stage of PR2 needs it's Sat amp "rebooted" (aka cable unplugged and replugged back in has sometimes worked).  (See blue and green UL and LR which have gone noisy at high frequencies since May sometime.)  We suspect that if we look around we will find more of these which do not pose a threat until it is determined that they are a problem.  No action for now, maybe add it to a Tues morning task to try the power cycle.

Using Rana's excitations last night, we measured the REFL port ASC sensing matrix for YAW at three different times (each measurement last 512 sec ~ 10 mins):

1). gps start: 1229337258; input power: 20 W; POP18: ~62 ct

2). gps start: 1229338980; input power: 23 W; POP18: 56 -> 55 ct (approaching thermal steady state)

3). gps start: 1229316688; input power: 23 W; POP18: ~54 ct (reached thermal steady state)

The response matrix corresponding to the three times are (in order listed above):

The response matrix can be inverted to get an input matrix. For the two interferometric DOFs PRC2 and CHARD, the input sensing matrix:

The input matrices for both PRC2 and CHARD have some large cancelation between the signal in RF9_A and (RF45_A+RF9_B), and use the residual as the input signal. This indicates a high condition number for the original response matrix. Not sure we could use those matrices as its variation might make the residual signal changes sign...

0.7.2 is a bug fix release for ndscope.  It had been rolled out successfully at LLO, and control room users report that it fixes some bugs.  I rolled it out to the CDS workstations this morning under WP 8020.

Heard Verbal Alarm (thought it was from "Japan").  I went to Jim and we took it to the EARTHQUAKE state in thought it was from Japan; in hindsight, now that we see it's closer and from Russia, might have been preferred to use the "Very Large EQ" Red button for a big & close EQ such as this.

• Watchdogs tripped for: 
  • SUS:  TMSx
  • SEI:  ETMx (Stage 1/2), BS (Stage 1/2), ITMx (Stage 1/2), ITMy (Stage 1/2), ETMy (Stage 1/2), HEPI HAM1
• Now waiting for seismic signals to come back down to the 10um/s range (currently up to 50-60 um/s) & then can start bringing back ISIs. 
• Sensor Correction can be turned on around ~1um/sec.

TITLE: 12/20 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: 0.04 μm/s
    Secondary useism: 0.94 μm/s
QUICK SUMMARY:

Commissioners left just before shift.  Have been trying to acquire DRMI for last 25min, so switching to acquire PRMI.  Microseism looked like it had been trending down slightly, but has been picking back up the last few hours (still above the 90th Percentile).

I had the CO2 guardians paused on the 20W setting for several locks tonight. During this time we had several lock losses and were back up again without any issues - DRMI was locking within 2 minutes. Given that it didn't seem to make a difference I have remove the 2W CO2 powers and now it's just set for 20W to make things simpler (CO2X=0.6W and CO2Y=0.3W).

We tried common ITM ring heater changes this evening, 0.25W on each segment. RF18 leveled off around 58, down from 61, so we stopped and dropped the RH levels back down.

Currently ITMY ring heater is set to 1.8W, turning it down did not help, eventually lost lock to a 0.45 Hz instability, DHARD, CHARD, DSOFT, PRC1 pitch were oscillating a lot. As soon as we start approaching RF18 ~ 50 again the ASC starts becoming more unstable. At this point PRCL gain had dropped to about 60% of its value on power up to 20W. Increasing CO2X also didn't help. This time the 3.5Hz oscillation in the SOFT loops. DRMI is now currently struggling to lock, some big bullseye modes can be seen. Hopefully the CO2 and RH changes will settle down soon for the day shift...

Plotted the REFL_A_9_I and REFL_B_9_I analog outputs.  Units are uncalibrated RF9 I monitor Volts/rtHz.
REFL_A is still being used for CARM control, REFL_B is just an out of loop sensor.
For reference, here's the last calibrated CARM noisebudget I made.

Looks like sensing noise dominates at 1kHz and below, and the seismic/scatter noise rises much faster below 10 Hz.  What could this excess low frequency noise imposed on the laser mean for the rest of our control loops?

According to the digital LF readouts, there's 5.2 mW on REFL A, and 4.8 mW on REFL B at the time of these measurements.

I phased LSC REFL B using an excitation at 1234.3 Hz into the Common Mode Board EXC1 and reading out the REFL_B_RF9 I/Q demod monitor and an SR785.  

Initially, the LO delay was set to Internal, using the physical switches for delaying (JPG 3).  All switches were on, giving 32ns of delay, or 105 degs of phase.
I switched the LO delay to External, and found a minimum response in Q at phase of 3 degs.  I am guessing that the change from Internal and External gave us ~100 degrees of extra phase.

I am only now appreciating Keita's previous work on this, where it's possible that REFL B has some frequency dependent demod phase.  I'll take some transfer function of Q/I and make sure that I didn't mess up Keita's phasing at low frequency.

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

In the meantime, I am tracking down the CARM analog/digital signal chain.  One highlight are Pic 4, where I found a TNC splitter connected to the output of the 2 channel demod board I-output.  The one end of the splitter goes to the common mode board, while the other goes to an 8 channel whitening chassis.
This trackdown is to think about whether our I and Q responses are balanced, and what can our Q phases tell us about the IFO thermal state, sideband imbalance, etc. 

Useful DCC of the PSL electronics: D1200666

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

We are wondering if the REFL_A_9_{I,Q} digital readouts can tell us about the CARM loop, because the digital channel REFL_A_9_I_ERR goes to 27.2 counts when the IFO is locked at 20 W, and not zero, even though the REFL A 9 digital phase is zero.  
It's clear REFL_A_9_Q_ERR is responding to thermal changes of the IFO (like the differential lensing Dan tuned last night). However, REFL_A_9_I_ERR behaves like a suppressed error signal with a bias that increases with input power and CARM loop gain, and does not change with IFO temperature (see pic 7).
I checked the dark offsets, they are not responsible for the bias we see.  

------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
 
I phased from H1:LSC-REFL_A_RF9_PHASE_PHASEDEG = 84.80 degs to 70.01 degs.  I did this using the digital REFL_A channels, so this may not be perfect. 

EDIT: I rephased REFL_A_9 to 94.67 degrees during the last lock using the analog demod board monitors (pics 9 and 10), the phasing was significantly different than the digital phasing.

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

I also tried an failed to see PCAL in the REFL_A and REFL_B spectra.  Naively, if we believe this plot and the calibrated PcalY spectrum response at 8 Hz being ~1e-14 m/rtHz, we should see equivalent compensation by CARM amounting to 0.7 mHz/rtHz, meaning the 8 Hz line ought to be visible.  The super massive CARM loop gain crushes the line at 8 Hz of course, but the displacement signal should be there at REFL.
I forgot about the massive additional loop gain of MC2 at 8 Hz, there will be nothing at 8 Hz to see.  And things don't look great at higher frequencies because the PCAL strength falls as 1/f^2 and the CARM optical gain falls like 1/f.

• 28 Hz instability:
  • We found that the PRC2Y ASC loop was changing the loop shape of the PRCL loop, only after the interferometer sat at 20W for several minutes.  This was the cause of the 28 Hz instabilities we have seen several times (4062).
  • Hang moved the ASC gain changes and cut offs for SRC2, PRC1 and PRC2 to the end of 2W ASC engage state, so that these are on during the power up. We haven't seen the 28 Hz instability since.
  • Rana measured the PRC2 loop, which has a very funny shape.  We switched PRC2 actuation from PR2 to PR3 a few months ago.  Rana suggested that there could be something wrong with the actuator (like a broken coil) that is causing large cross couplings.
• 3.2 Hz instability
  • Since fixing the 28 Hz instability, we have been having trouble with an instability at 3.2 Hz which seems to be an ASC problem.
  • We measured many things, including PRC2P which has a suspicious feature around 3Hz (but this loop should be stable).
  • Rana edited the cut off filter for PRC2.
  • Georgia saw that the 3.2 Hz instability seems to be more in pitch than yaw, it is especially visible in MICH P, PRC1 Y, SOFT P+Y, HARD P+Y
• Moving PRC2 to POP:
  • When we powered up to 20W we lost lock when the PRC2Y error signal drifted away from zero and there was a large excursion in SRC1+ SRC2 Y (POP 90 increased, OPO18 dropped). 
  • Next lock we moved the PRC2 ASC loop to POPX, this is now in the guardian, this seems to have helped us to be more stable.
• In one of the locklosses, 1229307644 we saturated the ADC for MC2 trans before the lockloss.  I lowered the whitening gain by 9dB, and added 9dB filters to keep the calibration the same as in 9716.
  • We are also saturating the ADC for PWR_IN_AC_INMON on ASCIMC, which I think is on a diode in transmission of the bottom PSL periscope mirror. 

If you want to monitor the PRCL and SRCL optical gain, you may run 

python /ligo/home/craig.cahillane/Git/IFO/LSC/scripts/PRCL_132HzLine_and_OscillatorMon.py
python /ligo/home/craig.cahillane/Git/IFO/LSC/scripts/PRCL_152HzLine_and_OscillatorMon.py
or
python /ligo/home/craig.cahillane/Git/IFO/LSC/scripts/SRCL_138HzLine_and_OscillatorMon.py
 
to start lines in the LSC oscillator matrices 1 and 2 respectively.  
The lines are demodulated at the correct phases for 20 watts, and the script sets the filter modules with 10 Hz bandpasses about the exc frequency, and a 0.1 lowpass on the I and Q signals.

An example PRCL optical gain monitor is shown below.  The POP 9 I signal is sent to H1:LSC-LOCKIN_1_DEMOD_1_SIG, demodulated by the excitation, and the results lowpassed and reported in H1:LSC-LOCKIN_1_DEMOD_1_{I,Q}_OUTPUT.

EDIT:
I made a MICH equivalent script in the same directory, and a bash script which launches all the OSCs: 

bash /opt/rtcds/userapps/release/lsc/h1/scripts/set_up_LSC_dither_lines.sh

and a code which shuts down the dither lines:

bash /opt/rtcds/userapps/release/lsc/h1/scripts/take_down_LSC_dither_lines.sh

Nutsinee, Sheila

Summary: We went back through the data from last night's squeezing injection.  We see that the noise added to DARM doesn't depend directly on the CLF power, but it is suppressed by the LO locking loop (which is feeding back to the OPO length).  This suggests that OPO length noise is coupling to DARM but not through the CLF or seeding; back scatter is one possibility.   

Details:

Our goal last night was to change the CLF power to see how it impacts the noise added to DARM by the squeezer.  We did not change the demod phase at all to try to adjust the squeezing angle. Since the locks were short last night, we were only able to get one or two injections in each lock, and our alignment was changing from lock to lock.  We looked back at the data to find the CLF fiber launch power, which we use to estimate how much CLF was injected into the interferometer, the RF3 demod power, and a few other things.  Since squeezing injections were done over a couple of locks, and the squeezer alignment was adjusted after the first lock, the alignment of the squeezer to the OMC was different for different times.  The number in parenthesis in the legends of the attached plots are an estimate how well aligned the squeezer was.  We took the ratio ((10^(RF3/10)/sqrt(CLF launch power)))^2, and normalized by the best alignment from last night to make this estimate. 

The first observation (second attachment) is that when the squeezer is badly misaligned, we add less noise to DARM (which is what you would expect).  The second attachment shows that we have more noise introduced into DARM with better alignment, the same injected CLF power, and the same analog gain in the LO loop.  This means that we had lower optical gain in the LO loop, but also that we have less light from the squeezer reaching the DCPDs. 

The next more interesting observation from is that increasing the analog gain for the LO locking loop reduced the noise added to DARM for a similar CLF power and alignment efficiency (first attachment). 

One explanation could be that increasing the LO loop gain (which is currently feeding back to the OPO PZT) decreased the OPO length noise.  Nutsinee and Daniel today found that we were causing a lot of OPO length noise (46058), which is not suppressed by our OPO locking loop since it feeds back to the laser.  Since the OPO length noise was the dominant noise in the LO error signal at the time of these injections, the OPO length noise could be decreased by increasing the LO gain.   

The OPO length noise could directly couple to DARM through intensity noise on the CLF.  Since the CLF is off resonance in the OPO, the OPO length fluctuations cause intensity noise on the CLF, which is attenuated by the OMC but could still add noise if large enough.  We also have the comparison in the third attachment (below), where the analog gain stayed the same, the CLF was increased and the alignment was improved, but the noise in DARM improved.  The DCPD 3MHz demod shows that there is 3dB more 3MHz signal at the time of the yellow trace (3:46:59 Dec 18th), which means the LO loop gain was increased by sqrt(2), so the OPO length noise and the CLF RIN could have been decreased by a factor of sqrt(2).  The CLF power reaching the DCPDs was a factor of 4 higher, so if the CLF RIN was the coupling mechanism we would expect the noise to be larger by 4/sqrt(2) ~3 in the yellow trace, but the noise is actually lower, suggesting that the coupling isn't through the CLF RIN.  We are planning to revisit 45858, since that projection of CLF RIN to DCPDs was based on a RIN measurement made on ISCT6, before the OPO length noise would have been imposed on the CLF intensity.  '

This comparison also suggests that the noise isn't due to seeding that comes through the AOMs in the CLF path, since there would be more seeding when we inject more CLF into the OPO, and when it is better aligned.

The fourth attachment is perhaps a more convincing comparison to see that CLF intensity noise and seeding aren't the cause of noise in DARM.  At the time of the mint colored reference (4:33:40 UTC) there was 6dBm more 3MHz signal than at the time of the black reference (4:30:26 UTC), but the LO loop gain would have been kept the same since Nutsinne increased the analog gain by 6dB for the lower CLF power.  (The shelf at around 100 Hz is a transient, we were having these kinds of transients which could be fringe wrapping during all of our locks last night).  

Here are the times that we used. Some of these were very short measurements.  

M. Wade

There was an issue at LHO where the OBSERVATION_READY bit of the ODC-MASTER channel was not reporting properly.  Since this is the channel used by the GDS calibration code to determine h(t)-OK in the GDS-CALIB_STATE_VECTOR, in consultation with Keita I made a configuration change so that h(t)-OK is now determined based on OBSERVATION_INTENT instead of OBSERVATION_READY.  This change was made and the pipeline was restarted on the DMT machines at GPS time 1228956497.