There is a comb in h(t) at 56.84044 Hz and multiples, identified in the Oct 11 low-noise data, with extreme amplitude modulation.  In alog 44495 and follow up, it is found to be in many EX channels, including unused ADCs in the PEM system. The time when this line began has now been tracked to August 28 15:20 UTC.

The attached spectrogram shows the lines appearing. The most likely cause that I can see in the alog is an update of the TwinCAT system manager on h1ecatx1. This is logged as starting at 14:54 UTC (alog 43706, and alog 43697 seems to indicate that the restart happened at 15:20 UTC.

Gabriele, Jenne, Shiela, Hang, Craig and Danny

• Today and yesterday there have been quite a few lock losses so Gabriele, Sheila, and myself looked at the trend data of the past day and noticed that there was a significant change in spherical power in the Hartmanns while transitioning from 2 watts to 20 watts that correlated to a decrease of the PRG and POP RF18 (please see first attached figure). 
• Because ITMX HWS was so noisy, Gabriele wrote a python script that smooths out and overlaps all the changes in HWS ITMX and HWS ITMY right after transitioning to 20 watts and before losing lock (please see second attached figure).
• The spherical power should be held relatively constant throughout this transition which means that the TCS settings needed to be adjusted. 
  • To compensate for the approximately 15 microdiopter increase on ITMY, the ring heater power was increased by .7895 watts. The inverted ring heater step response was used to estimate this (please see last two figures). To compensate for the change in spherical power due to this ring heater change, CO2Y preheating needed to be adjusted and was increased by .6 watts.  
    • Guardian was getting angry when I changed the CO2Y preheating settings because the measured CO2 power was far off from the requested power. This is due to bad calibration in the rotation stage that manages the CO2 power delivered to ITM. I pressed the "search for home" button and it seems to have fixed this issue temporarily.  
    • The inverted RH step response method is being used to adjust the ring heater settings on work station zotws17 (should be running till 10-16-2018 6:00 pm PST). 
  • To compensate for the approximately 10 microdiopter decrease on ITMX, CO2X will only be reduced by .1 watts (from .5 watts to .4 watts) as we power up to 20 watts. 
  •  Both CO2 power adjustments were determined by this CO2 actuation calibration
• A large earthquake in New Caledonia at 2018-10-16 1:31:25 UTC prevented us from acquiring lock for the rest of the evening. 

The RMS calculated from suppressed LO phase noise yields 20 mrad. Daniel thinks there's something wrong with the measured transfer function so it couldn't really be well fitted with known parameters. Since fast path out of LO common mode board goes to TTFSS, it could be that the funny behavior of the TTFSS is to blame. The cross over frequency between fast and slow path is 1kHz. 220 Ohm that we doodly clipped last time was also taken off.

The measurement was taken after PFD board and TTFSS modification (alog44490)

I looked at a recent lock loss event from the CHARD blend guardian state (gps time 1223672219), 

https://ldas-jobs.ligo-wa.caltech.edu/~lockloss/?event=1223672219

I saw an interesting glitch in ETMX_L2_MASTER_OUT about a second before the lock loss (see saturations figure), and followed this up with Sheila on the lockloss tool to look at various CHARD, DHARD channels surrounding the glitch (see CHARD figure).

Sheila suspected that because the CHARD pitch/yaw gains were not occurring simultaneously (happened about 0.3 seconds apart), that the waits for the ASC-CHARD channels in the CHARD_BLEND guardian state needed to be removed. In this stage, the low pass in A (ASC-CHARD_Y_A_SWSTAT) needs to happen at the same time that the gain is added for B (ASC-CHARD_Y_B_GAIN).

After adding the "jel" packs last week, it looks like the lower frequency (10mhz-50mhz) noise is better on BRSY. In the first attached plot I compare some 3000 second spectra from Sept 26 2018 ~10:45 utc (before adding the packs) and Oct 15 2018 ~8:00 utc (after all the work done last week). Both of these windows are low wind times. The top plot shows that the RX out asd (ground tilt in nrad/rthz) is lower now with the packs. The second plot compares the tilt subtracted sts and the uncorrected sts asds. The unsubtracted asds are very similar betweeen 10mhz and 50 mhz for both time periods, but the supersensor is lower with the packs.

The .03-.1hz blrms channels for the BRS are also a quick way to evaluate the noise during low wind times. The second attached plot shows zoomed in 20 day time series of the blrms minute trends for both BRSX & Y, and during quiet times BRSY blrms are lower now than they were before adding the packs, BRSX is unchanged. It bears watching, but I saw a similar difference over 50 day trends. 

Here's a scan of the dark port content at 20 W, from the OSA, with the main fields highlighted.

TITLE: 10/15 Day Shift: 15:00-23:00 UTC (08:00-16:00 PST), all times posted in UTC
STATE of H1: Commissioning
INCOMING OPERATOR: None
SHIFT SUMMARY: More work on angular controls. Initial alignment was done at 1400 PDT.
LOG: See attached

Yesterday Sheila and Danny retuned the OSA: now there is only one scan per second, with better resolution. Therefore the OSA analysis script did not work properly anymore.

Now the script does not average anymore multiple scans (since there's only one scan per second). Since the 9 MHz are well resolved, I removed the complicated fitting procedure, so that now each scan is analyzed as follows

1. identify the largest peaks and find the two spaced just right to be the 45 MHz upper and lower sideband
2. calibrate the scan x axis in frequency, based on the 45 MHz sideband positions
3. find the largest peaks in a small range around the nominal positions for carrier, 9 and 45 MHz sideband
4. compute the power in each field by integrating the whole peak

The plotting is similar to the previous version: there is only one time race now, corresponding to the integrated power in each peak. The same is true for the data saved to file: I removed the fit peak values, since there is no more need for complicated fitting.

Also, the scan turns out to be slightly longer than one second, and out of sinc with the LIGO timing. So now the script uses two seconds of data to find the trigger and the peaks.

As before, the script is availbale in /opt/rtcds/userapps/release/isc/h1/scripts/osa.py

Sometimes the peak finding fails, especially if one of the sidebands falls close to the edge. I'm sure we'll soon find another failure mode...

Coherence between DARM and the ISS second loop error signals is at a 0.1 level.

So not big enough to be worried right now, but not far enough to forget about it.

H1:HPI-PUMP_EX_CTRL_OUTPUT trended up slightly, but all seems good.

Closing FAMIS 7724

ETMY Y, ITMX Y, and BS P are getting closer to 10urad.

BS and ETMY sums are trending down.

Closing FAMIS 11187

Laser Status:
Front End Power is 33.34W (should be around 30 W)
70W Output Power is 74.13W
Front End Watch is GREEN
70W Watch is GREEN

PMC:
It has been locked 0 days, 2 hr 15 minutes (should be days/weeks)
Reflected power = 17.35Watts
Transmitted power = 43.57Watts
PowerSum = 60.92Watts.

FSS:
It has been locked for 0 days 1 hr and 9 min (should be days/weeks)
TPD[V] = 2.624V (min 0.9V)

ISS:
The diffracted power is around 1.4%
Last saturation event was 0 days 2 hours and 15 minutes ago (should be days/weeks)

All seems good. Closing FAMIS 10982

• Lockloss around 13:17:00 UTC was me - A large step in PRC2 Pit (100 counts input offset) was okay for a short time, but slowly things started to ring up.  Didn't take out the offset soon enough.
• Lockloss around 15:50 UTC unknown.  This was the first lock with PRC2 on REFL WFS (alog 44553)
• Lockloss around 17:22 UTC unknown.  We were doing PRCL injection while changing POP_9_I->SRCL LSC input matrix element, but shouldn't have caused a lockloss.
• Lockloss around 18:33 unknown.  Some violin damping ongoing, but otherwise nothing was changing right at that time.

Beverly, Andy, Josh, TJ

In the low noise (and 22W into IMC) locks of 10/13 we saw a handful of whistles around 4 UTC and 8-9 UTC. They are the classic old-style whistles we also see at LLO: they couple in prcl/srcl, but not iss. They show up strongest in REFL A RF9 Q (see attached, along with strain). These were picked up by hveto which found them at the same time in the SRM noisemon, which is seeing the srcl feedback, almost certainly (so that noisemon is working right). The whistles aren't very numerous, but we wanted to give a heads up as this suggests a change, since whistles were rare to non-existent at LHO in O2.

Here are some scans for your perusal. 

• 03:53:01 scan: https://ldas-jobs.ligo-wa.caltech.edu/~detchar/hveto/day/20181013/1223424018-1223452818/scans/1223437999.2/
• 04:13:25 scan: https://ldas-jobs.ligo-wa.caltech.edu/~detchar/hveto/day/20181013/1223424018-1223510418/scans/1223439222.55/
• 08:56:09 scan: https://ldas-jobs.ligo-wa.caltech.edu/~detchar/hveto/day/20181013/1223424018-1223510418/scans/1223456187.02/
• 08:25:22 scan: https://ldas-jobs.ligo-wa.caltech.edu/~detchar/hveto/day/20181013/1223424018-1223510418/scans/1223454340.04/

Hang, Dave:

at 12:37 we restarted h1asc to fix a mis-connected IPC receive part. No DAQ restart was needed.

While locked at 20 W, tuned POP_9 and POP_45 phases to maximize PRCL signal in I phase:

POP_RF9 phase changed from -20.5 to -11
POP_RF45 phase changed from -4.9 to 14

Early this morning we got cell phone alarms due to a high EY chiller supply water temperature. Plot shows 5 hours minute trend, max values in degF. Alarm level is 49degF.

During the low noise lock at 4:50 UTC today (Oct 11), there is a comb of lines in h(t) at 56.84 and multiples. They are quite narrow lines. The most interesting characteristic is that they have extreme amplitude modulation.

The first attachment is the h(t) spectrum with the lines marked. The second shows narrow BLRMS around each of the lines. The fundamental at 56.84 Hz has a half-period of about 24 minutes. The amplitude goes very close to zero, and looks very periodic. Each multiple n has zeros in the same place, but the period is n times shorter. Maybe this could be all generated from the fundamental with the right kind of nonlinearity. I'm not sure what would give such a slow but extreme modulation - maybe some centering servo?

The BRUCO results show a lot of channels at End-X coherent. Accelerometers on BSC9 seem to see it, but without amplitude modulation. PCal X sees it in the TX and RX PDs, but stronger in RX; but it doesn't seem anywhere near the level to get into h(t).