18:28 Lockloss - Unknown

18:29 Richard out to TCS to check on Interlock

18:31 Richard back

18:40 Trouble getting Green Xarm back. PDH VCO error - Reset

19:26 NLN 

There are two diffs in SDF regarding CO2 LASER power/setpoint. Typical SDF wanting accuracy wanting e-15 resolution. Jason was contacted and gave the OK to accept the diffs and move to Observing.

19:32 Intention Bit: Undisturbed

- No extraordinary features were found in the bullseye signals at HF/VHF frequencies.

[Summary]

During the commissioning call in the past Friday, there was a suggestion to check the HF/VHF (~MHz region) behavior of the bullseye sensor in order to make sure that the sensor is detecting what it should detect. In response to it, I have measured high frequency part of the signals coming out from the bullseye segments using an RF analyzer (an HP 4395A). I didn't find extraordinary features. I am concluding that the HF/VHF contents are not an issue.

[Some details]

The measurement was done by inserting a 9-pin Dsub breakout board right in front of the whitening board (34154) during the maintenance period yesterday. This measurement was performed after Peter and Jason had finished their flow sensor task (34811). The PSL laser was running in a nominal condition. The IMC was locked with a low power (~2 W). The interferometer was kept unlocked.

Each segment was read out before they go into the whitening board. I hooked up an HP 4395A to the breakout board using a pair of clips and a BNC cable. Unfortunately, this cheap setup was so good that it picked up lots of RF lines. Nonetheless, I didn't see an outrageous line or suspicious feature in the bullseye signals. Most of the lines are as small as -60 dBm and many are less than that except for the one at 160 MHz which I think is a pick up from the measurement cable (as high as -50 dBm). See the attached for the spectrum. I also attach the raw data just in case.

Since all four segments showed almost identical spectrum, I saved the data only for the center segment. For comparison, I also saved the background spectrum with the clips left open. Note that the background spectrum wasn't stationary. After the measurement, I removed the breakout board and completely cleaned up the measurement setup.

Starting CP3 fill. LLCV enabled. LLCV set to manual control. LLCV set to 50% open. Fill completed in 1903 seconds. LLCV set back to 14.0% open.
Starting CP4 fill. LLCV enabled. LLCV set to manual control. LLCV set to 70% open. Fill completed in 1416 seconds. TC A did not register fill. LLCV set back to 42.0% open.

WP6522

During the maintenance period yesterday, I have measured the spectra of IMC WFS DCs with and without the IMC locked. The purpose of this measurement was to cross-check the calibration factor of 1/3 which previously estimated by Daniel (31631, Nov 2016).

A conclusion:

I agree with Daniel that the IMC WFS DC signals should be scaled by a factor of 1/3 when the IMC is locked.

[Some plots]

• The 1st attachment: raw spectra of the IMC WFS DCs
  • (orange) when IMC was unlocked with the MC2 significantly misaligned. Pin = 2 W. (It had to be low power in order not to saturate the ADCs.)
  • (red) when IMC was locked. Pin = 20 W. The main interferometer was left unlocked.
• The 2nd attachment: calibrated spectra of the IMC WFS DCs.
  • (blue) when unlocked.
  • (red) when locked.
  • Also, I have included an extra factor of sqrt( 8 / pi) to get all the spectra calibrated in the unit of fractional amplitude defined as |E_{10}/ E_{00}| or |E_{01} / E_{00}|.

A side note: depending on whether the IMC is locked or not, the spectral shapes vary. This is not so surprising because when the IMC is locked, the reflection port must be dominated by non-00 modes. So the signals should contain relatively large amount of cross terms from various HOMs. So for the reason, I don't have a good explanation why they have to be a factor of 1/3. Even so, it is interesting to see some peaks come and go.

a2l measurement shows YAW out up to .85. Livingston is down.

TITLE: 03/15 Day Shift: 15:00-23:00 UTC (08:00-16:00 PST), all times posted in UTC
STATE of H1: Observing at 68Mpc
OUTGOING OPERATOR: Jim
CURRENT ENVIRONMENT:
    Wind: 15mph Gusts, 13mph 5min avg
    Primary useism: 0.04 μm/s
    Secondary useism: 0.24 μm/s 
QUICK SUMMARY:
 

TITLE: 03/15 Owl Shift: 07:00-15:00 UTC (00:00-08:00 PST), all times posted in UTC
STATE of H1: Observing at 67Mpc
INCOMING OPERATOR: Ed
SHIFT SUMMARY:
LOG:

Quiet shift. Nothing to report.

Flow rates for head 1-4 for ~a day after the sensor replacement.  The signals look relatively clean.
Head 3 looks a little ragged compared to the others.

Power meter flow and front end flow rates look okay.  The front end flow rate looks a little noisy
in its maximum recorded value.

TITLE: 03/15 Eve Shift: 23:00-07:00 UTC (16:00-00:00 PST), all times posted in UTC
STATE of H1: Observing at 71Mpc
INCOMING OPERATOR: Jim
SHIFT SUMMARY:

Recovery from Maintenance Day for the first 1-2hrs.  And most of the shift we have had sustains winds around 10-20mph (with gusts above 40mph).
LOG:

• 23:08 INJ NODE transitioned to INJ_SUCCESS state 
• 23:11-23:26:  LVEA Sweep (Corey
• 0:32 - 0:46:  OBSERVING
  • Lockloss due to PI Mode 27
• 2:14 - 2:34 OBSERVING
  • Lockloss due to PI Mode 27
• 3:32 - 6:14 OBSERVING
  • Lockloss unknown
• 3:35 TCSy Chiller Flow Low Alarm
• 6:43 OBSERVING & handing off to Jim (Have not dealt with PI Modes so far.)

Terra, Corey

PI Mode #27 was a bit touchy this evening.  After two locklosses, talked with Terra.  It's frequency was 14039.5Hz.  The original filter being used was 14038.5Hz.  We ended up switching to one a little closer:  14040Hz.  

So for this current lock, we

• Stayed at DC Readout for atleast 15min after the PI Mode lockloss.
• Once PI Mode#27 rung up, we adjusted the phase.  Could clearly saw positive and negative phase values.  
• From 90degrees went as far as -30 & ultimately settled on 60deg & we now look good.

Note For Operators After a lockloss due to a PI Mode:

1. Confirm the filter frequencies match with the actual mode seen on DTT
2. After the lockloss, stay at DC Readout for atleast 15 min to allow the PI Mode to settle down. 
3. Also check that the mode is not visible in DTT, before moving on from DTT.k

OK, off to cook supper!

After Maintenance had a few SDF diffs and went through them with Kiwamu and Nutsinee.  All of them were ACCEPTED (namely for ASC, OMCPI, PSLFSS, etc.).  

After the last few locklosses, there have been some repeated TCS diffs (attached is a snapshot of the channel).  Nutsinee said we can ACCEPT them (they are tiny changes 10^-15), and can probably NOT MONITOR them eventually...but we should wait for call from TCS person.

Terra & I also switched to a different filter for PI Mode#27, so this was also ACCEPTED.

After all the hub bub of a Maintenance Day, the site is somewhat quieter (no one here and the winds are finally mostly under 20mph).  

Made a couple of lock attempts starting just before my shift.  First two locks did not last long:

• Lock#1 ended due to PI Mode 27 (it had a setting of -70deg phase & +7000 gain).  Range was an ugly 60Mpc
• Lock#2 broke lock after a few min at NLN.  No obvious reason.  Range was a much better 70+Mpc
• Lock#3 (also at a decent 70+Mpc) ended due to PI Mode27 (same settings as above.  Tried various changes to no avail and it broke lock within 3min while also ringing up neighboring modes).  Gave Terra a call for help!  She had me check the Mode27 freq (it is 14039.5Hz.  BP Filter is 18038.5 with PLL set frequency of 238.5.....Terra thought this should be wide enough for that mode.)  

As noted earlier we ran the gamut of rung up modes (roll, some bounce, violin & now noticing PI modes).  So I am going to stay at DC Readout for a while (Terra suggested ~15min) & then we'll go from there.

Taking a screenshot for Hugh.....

TITLE: 03/15 Eve Shift: 23:00-07:00 UTC (16:00-00:00 PST), all times posted in UTC
STATE of H1: Lock Acquisition
OUTGOING OPERATOR: Cheryl
CURRENT ENVIRONMENT:
    Wind: 31mph Gusts, 25mph 5min avg
    Primary useism: 0.19 μm/s
    Secondary useism: 0.28 μm/s

It's become windy in the last few hours with gusts over 40mph (might be subsiding now).
 

QUICK SUMMARY:

Cheryl handed over an H1 which was in a LOCKING state.  Kiwamu noticed that we had rung up Roll modes & so he addressed them a bit & then we went to for NOMINAL LOW NOISE. 

Locked up but the range was down at a paltry 60Mpc.  Now it was still windy.  And we also have the BRSy in an odd state after Maintenance Day.  (Ultimately lost lock after a pesky PI Mode 27.)  

Jason, PeterK, Kiwamu (WP 6522)

We powered up the diagnostic bread board (DBB) this afternoon during the maintenance period and compared the pointing jitter signals against those from the bullseye sensor.

They show good agreement in their spectral magnitudes and showed high coherence. So pointing jitter noises seen by the bullseye are consistent with what the DBB sees.

* * * * * *

The attached are the plots showing the calibrated jitter noises and their coherence. As we know from the past measurements (e.g. 31631), the QPD2 spectra of the DBB show lower magnitudes than that of QPD1 by a factor of 5 or so across the entire frequency band. The calibrated bullseye jitters are found to be roughly at the identical magnitude to those seen by QPD1 of the DBB. Also, looking at the coherence, we found that the bullseye sensor seems to be almost at the same Gouy phase as QPD1.

After the measurement, we powered off the DBB by following the standard procedure (31717).

I also changed the input matrix for the pitch jitter mode of the bullseye in order to obtain zero offset when the beam is well centered. Now the input matrix is set such that (lower segment, upper right segment, upper left segment) = (-2, +1, +1) which had been (-1, +1, +1) since the week before the last week (34460). The calibration coefficient for it was then changed to reflect this change by a factor of 0.67. I updated the SDFs.

Head flow rates for heads 1-4 for the past 6 hours.  The hiccups for the first 40 or so minutes
are due to starting and restarting the crystal chiller as we checked for leaks.

Also shown are the flow rates for the front end laser and the power meter circuit.  These were
previously disconnected when we were trying to debug what was causing the chiller to trip so often.
Now that all four sensors in the heads have been replaced, they are not likely to be a cause of
a trip unless there are still stray bubbles working their way through the system.