BLRMS approaching 1um/s following multiple reports from Azerbaijan area.

21:39 21:11 H1 to Commissioning. TOO as L1 also commissioning

21:39 21:11 Squeezer team out to squeezer area in LVEA

Keita Kawabe, Sheila Dwyer

We investigated the lockloss that Ed and Camilla just had at LOWNOISE_ASC (1257191567), and we believe that the ISS second loop attempted to close while the PSL rotation stage was still moving. We have made some changes to prevent this, although we can't say with certainty that this was the cause of the lockloss, because this seems to have stopped about 4 seconds before the lockloss. The attached screenshot shows that the power input to the IMC was having large gltiches a few seconds before the lockloss, which show up in downstream powers and CARM channels.  

One of the recent changes to the guardian was moving the ISS second loop engagement from happening after DRMI locks to happening at the end of the power increase.  The other guardian change that might be a cause is that TJ recently changed the LASER POWER guardian to fine adjust the input power to get it closer to the requested power, so our input power will be more consistent from lock to lock.  The ISC_LOCK gaurdian was not checking that the LASER_POWER guardian was arrived and done before moving on to the next guardian state.  You can see in the attached screenshot that ISC_LOCK moves from 507 (Maximum power) to 508 (low noise ASC) while the laser power guardian was still changing the power.  

We have made two changes, moving the ISS engagement back to DRMI, and adding a check that the LASER POWER guardian has arrived before we return true from MAXIMUM power.  We still are not sure that we understand the behavoir of the rotation stage.  

18:09 Lockloss

Re-lock attempt -1

18:11 ALSY needed aligning

18:23 FIND_IR starting

• no diff IR- offset in TRY_QPD_B_SUM

18:35 Lockloss - trying to play catch-up with diff slider

18:39 Re-lock attempt - 2

• DRMI was very misaligned; Guardian to PRMI - no lock; 20 minute timeout as per threshold protocol

19:10 Initial Alignment

19:25 IA complete

19:27 Re-Lock attempt - 3

19:52 Lockloss @ LOWNOISE_ASC

19:53 Re-Lock attempt -4

• after 2 minute I realized there was the offset in te Y QPD BSUM again (started FIND-IR @ 19:56)

20:06 lockloss

20:07 Re-Lock attempt- 5

20:37 NLN

TITLE: 11/07 Day Shift: 16:00-00:00 UTC (08:00-16:00 PST), all times posted in UTC
STATE of H1: Observing at 115Mpc
OUTGOING OPERATOR: Patrick
CURRENT ENVIRONMENT:
    SEI_CONF state: WINDY
    Wind: 4mph Gusts, 3mph 5min avg
    Primary useism: 0.01 μm/s
    Secondary useism: 0.18 μm/s
QUICK SUMMARY:

TITLE: 11/07 Owl Shift: 08:00-16:00 UTC (00:00-08:00 PST), all times posted in UTC
STATE of H1: Observing at 117Mpc
INCOMING OPERATOR: Ed
SHIFT SUMMARY: Squeezer took us out of observing twice. There is a user message: SQZ_MANAGER: NOT IN MANAGED MODE.
LOG:

08:00 UTC Squeezer takes us out of observing
08:02 UTC Hit INIT on SQZ_MANAGER. Squeezer relocked.
08:03 UTC Back to observing
11:01 UTC Squeezer takes us out of observing
11:02 UTC Hit INIT on SQZ_MANAGER. Squeezer relocked.
11:04 UTC Back to observing

Squeezer has taken us out of observing twice. No other issues.

TITLE: 11/07 Owl Shift: 08:00-16:00 UTC (00:00-08:00 PST), all times posted in UTC
STATE of H1: Observing at 118Mpc
OUTGOING OPERATOR: Jim
CURRENT ENVIRONMENT:
    SEI_CONF state: WINDY
    Wind: 3mph Gusts, 2mph 5min avg
    Primary useism: 0.01 μm/s
    Secondary useism: 0.10 μm/s 
QUICK SUMMARY: Squeezer dropped out at very start of shift. Hit INIT on SQZ_MANAGER and it relocked. There is a user message: SQZ_MANAGER: NOT IN MANAGED MODE.

Following up on work by Beverly, I ran a short Lasso run on today's range drop. The run spans 2019-11-06 02:00-07:00. Clearly very highly correlated channels are as follows:

Channel "Pearson Coefficient"
0.9799929848627831 H1:SQZ-DCPD_SUM_RMS1_MON.mean
0.9725502574463882 H1:SQZ-DCPD_RATIO_1_MON.rms
0.9312747855964804 H1:SQZ-DCPD_SUM_RMS1_MON.rms

The full page output is at https://ldas-jobs.ligo-wa.caltech.edu/~jrsmith/detchar/O3b/lasso-Nov-6-2019-2/ (updated using the fancy new python 3 version)

Figures below show:
1) The range drop from the summary pages.
2) The range and scaled SQZ DCPD channel
3) A scatter plot of range vs SQZ values
4) The other channels that are duplicate of the SQZ winner

I think not much new from what Beverly reported so far. Happy to follow up with further inquiries. 

Ethan, Adrian, Sheila

A few minutes before Nov 01 2019 20:41:42 UTC, the instrument air compressor was put on isolation springs in order to mitigate scattering in DARM at 35 Hz (see alog 52879). The first attached plot shows when the compressor was placed on springs. The horizontal red line in the figure it set slightly above 35 Hz.

Following up on this a few days later (Nov 6), we see there is still some scattering at 35 Hz that is consistent with the behaviour of the air compressor. This seen in the second attached figure.

Some work could be undertaken to better isolate the compressor and/or locate the coupling site. 

I've written a mid-level library of functions I always use to quickly get and plot data from nds2 in python 3.  It's called nds2utils.  

If you'd like to use it and provide feedback, I'd welcome it at https://git.ligo.org/craig-cahillane/nds2utils.  There are 11 examples.

It's on PyPi, so you can pip install it in your favorite virtual environment.

pip install nds2utils

I just noticed an intermittent timing error on the CDS overview. The RF Amp attached to port 8 of CER fanout-A has been showing a locking error on the oven controlled crystal oscillator. Image shows a 24 hour trend of the error signal alongside the Amp MEDM at a time when the system was in error. First error showed around 08:40 PST this morning, the large error stretches in the plot are roughly between 18:30 to 19:30, and from 22:15 to now (23:00).

I've had a quick look at the duty cycle during owls for O3a. A couple of quick numbers:

Of 176 total owl shifts, there were 96 shifts the IFO was locked when the owl shift started and stayed locked the entire time.

There were 7 shifts where the IFO never got to NLN.

The duty cycle for all O3a owl shifts was 84%, or: we were in some state other than NLN for 16% of the time.

Attached plot is the distribution for the amount of time the IFO was not in nln. The X bins are the number of minutes (with 30 minute bins) the IFO wasn't locked, the Y axis is the number shifts for each bin (i.e if the ifo was out of nln for 130 minutes, that shift gets counted as 1 instance in the 120-150 minute bin). I'll point out the tallest spike by far is the one at 0,  where the IFO was locked the entire shift. 

Not too sure what else to make of this, but the lump around 150 minutes kind of lines up with our normal time required to relock, indicating to me that mostly we lost lock once, then had a normal time relocking. Probably, most of the bins below 100 and above zero are from locklosses at the end of a shift, or a reacquisition that was started before the owl shift.

The first plot on the first page of the figure shows that the seismic signal from the instrument air compressor was reduced today by seismic isolation. But the huge DARM noise, with close to the same periodicity, is still occurring. Our manipulation of the on and off pressures for the compressor, and the new GV5 and 7 pressure settings, see previous entries, did not seem to affect the DARM noise – see end of first page time series. This and the observation that the noise does not occur at exactly the same part of the compressor cycle every time, is evidence that they are causally unrelated, though they have similar periods. On the other hand, the second plot seems to show that the period of the DARM noise changed with the period of the compressor a couple of days ago…. Chandra, Kyle and I were in the LVEA from UTC 23:40 to 00:06 on Nov 1 and 2.

Sheila, Timesh, Robert

There have been several alogs about this and there will be others, but I wanted to try make the situation more clear to detchar.

When the new instrument air compressor turns on, for about 5 minutes approximately every 25 minutes, we see a 35Hz line in DARM, which goes away when it is turned off. Around the time the compressor swtiches off, there are terrible scattering shelves in DARM.  The onset of scattering in DARM seems to happen around the time of the compressor switching off (as seen by PEM seismometers), sometimes it is 16 seconds after, sometimes 24 seconds, sometimes before the ground motion drops off by up to a minute.

The first attached screenshot shows the regular switching of the compressor, and the board band large noise in DARM when the compressor goes off.  The second screenshot shows darm spectra in 3 states:  compressor off, compressor on (there is a 35Hz peak added to DARM) and compressor switching off, shelf.  The PRCL/SRCL/MICH spectra also show fringe wrapping when the compressor is swtiched off.  

These are probably interesting times for people who are interesting in looking for scattered light.  Bubba and Tyler have ordered some parts to better isolate the new equipment, so this should be mitigated within a few days.