Nothing besides the trending "unusuals" to report other than curious looking oscillations in OSC_DB4_PWR that appear to correlate to the temp issues and don't seem to be as prominent in 1-3.

Topped up the crystal chiller with 250 ml.

I have added new features to the element lal_resample used in the gstlal calibration pipeline so that it can perform upsampling for the actuation equal in quality to the old gstreamer (version 1.4.5) resampler. The upgrade to gsteramer-1.10.4 on the clusters introduced a ~2% systematic error in the C01 frames from ~50 Hz to ~1 kHz during the month of August. See, e.g.,
https://ldas-jobs.ligo.caltech.edu/~alexander.urban/O2/calibration/C00_vs_C01/H1/day/20170802/

The change made was the addition of a sinc table filter in the upsampling routine. Several tests were done, and plots are attached:

The first two plots show the filter's response to a series of impulses separated by 4 seconds. This input data was upsampled from 128 Hz to 1024 Hz. The first of these plots shows 30 seconds of data, and the second is a close-up on a single impulse.
The 3rd plot is a 10-second sinusoid upsampled from 8192 Hz to 16384 Hz.
The 4th plot is a 30-second stream of ones upsampled from 128 Hz to 1024 Hz. The apparent thickness of the line indicates the amount of digital error, of order ~10^-8.
The 5th and 7th plots are ASD comparisons between the output produced by the calibration pipeline using this new resampler and the C00 frames from August.
The 6th and 8th plots are ASD comparisons between the output produced by the calibration pipeline using this new resampler and output pruduced with no resampling at all (i.e., all actuation was filtered at 16384 Hz). I suspect the wiggle above 1 kHz is due to a ~2% contribution from the actuation that is lost in downsampling to 2 kHz for the filtering.

For information on filters to be used for C02 production, see
https://alog.ligo-wa.caltech.edu/aLOG/index.php?callRep=39419
https://alog.ligo-la.caltech.edu/aLOG/index.php?callRep=36707

TJ, Sheila

This morning/afternoon we ( Keita, Jeff K, TJ, me, Lisa, Alvaro and Calum) realized that the riser for ZM2 (the squeezing TT in HAM5) is too tall, since it is set to match the height difference between HAM 5+ HAM6 at L1, which is not the same at H1.  

TJ and I went to the chamber to make some sanity checks that the discrepancy is about what we should expect based on drawings. We first set up a laser pointed that was mounted on a small breadboard in the cleanroom by HAM3 in the back of HAM6, pointing towards HAM5.  We leveled this by measuing the distance of the beam off the table near the laser pointer and again at the far edge of HAM6, roughly a meter away.  With our final tin foil adjustment we got the beam parallel to the table to within a mm over this distance.  The laser beam was 99 mm off of the HAM6 table top, and 203 off of the HAM5 table top, so we estimate the difference this way to be 104mm, or 4.1 inches.  

We also measured the height of the center of the ZM2 mirror to be 223 mm off of HAM5, or 8.78 inches.  That means that the center of ZM2 is 8.78-4.1 = 4.69 inches above the level of HAM6, them beam height in HAM6 is 4 inches.  Calum looked at drawings earlier and said that the riser height should be 0.76 inches too tall for H1 if it is based on L1 heights. 

We also attempted to measure the height of the center of the output aperture of the OFI off HAM5, which was a little difficult with our too short ruler and an akward angle.  We measured 3 times and got 207.5 mm, 205 mm, and 209.5mm.  (8.16 inches average, so 0.617 inches below ZM2).  

It seems like these measurements are in agreement with what Calum found in the drawings, within the precision of this measurement technique. 

The attached plot shows no change in PSL PEM signals, accelerometers and microphone, that can be attributed to the new manifold.

S Cooper, J Warner

We've been investigating the reasons into why HEPI tripped during O2 to see if anything could be done to prevent it. To do this we've been looking at both the time series sensor data in the local basis (H1,H2,V1,V2 etc), for the IPS, STS, L4C's and Actuators, the saturation counts and the watchdog status for every chamber (BSC's and HAM's - with the exception of HAM1). The first earthquake we ran this on was the Montana earthquake (GPS 118335800) as this was the only earthquake where HEPI was the first to trip.

When we if we look at the saturation counts, we find that only the L4C's are hitting their saturation threshold and are therefore likely causing watchdog trips, the horizontal L4C's are the first to saturate. 

If we then compare the levels that the watchdogs trip at across all 10 chambers, we find that ETMX trips at 10% the level of the other chambers. I've attached an annotated PNG highlighting the BSC chambers, and a .fig file that contains the saturation data for all 10 chambers. The dashed lines indicate the watchdog level (multiplied by 10,000 for easier comparison) with the solid lines indicating the specific chambers L4C. 

I'm now running the same script for other earthquakes that HEPI tripped in to see if this is an isolated case. 

(Travis S, Gerardo M)

We removed the main septum viewport and set it inside the cleanroom for HAM4 to clean it, first contact was applied on one face and it should be ready for removal on Monday.

For reference before removal, the fiducial line on viewport was at 1 o'clock, see attached photo.

The OFI cage was moved, beam was verified and it appears to be going through the aperture without clipping, for some reason the beam was moving a bit too much on pitch, purge air was dialed down but did not help.  Cage is being held in place by one dog clamp, the rest will be done on Monday.

TITLE: 11/18 Day Shift: 16:00-00:00 UTC (08:00-16:00 PST), all times posted in UTC
STATE of H1: Planned Engineering
INCOMING OPERATOR: None
SHIFT SUMMARY:
LOG:

16:30 Nutsinee to SQZ
17:00 TJ to HAM5, Nutsinee and Terry to SQZ, JeffB to LVEA
17:15 Gerardo to HAM5
17:45 Ed, Fil moving squeezer electronics into LVEA
18:15 Travis to HAM5
18:51 TJ - Out of the LVEA
18:52 Travis - Out of the LVEA
21:15 Travis, Jason to LVEA for elliptical baffles
22:00 Peter out of PSL
22:30 Travis out
23:00 Travis, Gerardo to HAM6, removing main viewport
23:15 TJ, Sheila to HAM5
0:00 Travis out

The new bull's eye detector (S1700201) was placed onto the table where the prototype one was located.
Adjusted the beam using the mirror located in front of the detector.  Between pins 3 and 8 of the DB9,
~ -14V was measured.  Between pins 1 and 6, ~ -26.5V.  Although it was a bit difficult to see with an
IR viewer, the spot on the photodiode looked reasonably well centered.  Powering up the detector caused
the high power shutter to close.  I reset the error condition without any issues.

    Not sure of the exact cause of the signal degradation of the previous detector, as reported by
Sheila, but a cable was close to the beam path.

    To debug some temperature related signals, I placed my hand on/over the temperature sensor closest
to the entrance of the laser room for more than a few seconds.  If I remember what Richard told me,
the signal labelled "north" on the MEDM screens, is in fact the south sensor and vice versa.

   Dust monitor checks showed DM#2 (HAM2) and DM#3 (HAM3) offline. 

   Reset DM2 with no problems. 

   DM3 had been knocked off its perch, pulling out the power plug and leaving it hanging by the network cable. No apparent damage done, as it is back in service. 

Note: I have mounting brackets and adapters on order for mounting these dust monitors on microphone stands. When these are deployed the dust monitors will be much better protected.  
   

Patrick got the new Beckoff controller running at the end station.  The old servo controller is still in place and still running the PID controlled VFD.  The Beckoff is just looking at the pressure transducers on the PS skid which are out of loop for the PID.

The attached plot shows the old and the new (different name) channels on the same plot through amazing wizardry!  Come on, it's about all I have. 

The Beckoff ADC is used much more efficiently than the Athena in the old box.  The Beckoff range is 75mV while the Athena is +-10V for a signal that should never exceed 50mV and nominally run about 20mV.

Clearly the time series is much better for the Beckoff with much tighter range and a lot less hair.  This might further improve when the VFD is controlled by a cleaner channel.  There have been times in the past where the old controller produced cleaner data than it does now but I don't think it was ever any cleaner than this data.

Next, maybe while I change the filters, we'll transition the PID to the Beckoff.

1. IM4 and IOO PZT

Before doing anything the beam was already coming to the OFI. We centered the beam on OFI input aperture using BS just so that it's easier for us to eyeball the beam motion.

We removed the PIT offset just enough so that the output wouldn't rail and IM4 is properly damped. (This happened to be the slider MEDM limit of -25000.)

We used IOO PZT to bring the beam back to the center of OFI input.

2. SR2-SRM-OFI

We used BS to center the beam on SR2, then used SR2 to center the beam on SRM. Centering accuracy claimed by Siddhesh??, Brijesh??, JeffK, and later confirmed by Koji, is a couple of mm on SR2 and less than a mm on SRM.

The beam was off in PIT by negligible but measurable amount (0.5mm) on the input aperture of the OFI, but a couple of mm too low on the output aperture.

Gerardo adjusted the OFI SUS to lower the output side of the OFI and it's within 0.5mm for PIT on both the input and output of OFI, but a mm or so in YAW at the output. This might sound small to you but we'll fix it for the reason explained in 4.

3.  OFI-OM1 craziness of O1/O2 seems to be gone, but this means that the septum window needs to be rotated (again).

We found that the beam on OM1 was about 9mm too high and 12mm off to the side. See Koji's attachment to this log.

Back in the day before O1, the beam was 10.6mm too low on OM1. (alog 13391)

Again back in the day, to fix this craziness the septum window needed to be rotated by 120 degrees clockwise, and this got the beam back to the right height on OM1. (alog 13477)

Changing the Faraday (or maybe something else) should have fixed the root cause of the craziness. But this means that the septum window should be rotated back by 120 degrees counter-clockwise. This will set the septum plate angle as designed, the beam height on OM1 should become about right.

OM1 position should be also shifted sideways to accommodate the yaw change caused by this. Which means that the OM1 and OM2 alignment needs to be mechanically moved in YAW.

4. OFI-ZM2 alignment is totally wrong, OFI YAW alignment should be revisited. 

Right now, when we install the additional HWP, the rejected beam from OFI is blocked by the Faraday rotator. The beam is not centered on the first steering mirror in YAW to start with though it's not clipping, it's close to the edge of the the second steering mirror, and the reflection from the secod mirror directly goes to the rotator.

We could adjust the steering mirror position and angle and be done with it, but Koji felt that we should align the OFI in yaw before moving the steering mirror, especially the position of the mirror mount.

Tomorrow.

I will NOT transition LVEA to laser safe tonight.

S Cooper, S Dwyer, J Warner, 

Attached in this alog are some overviews of the total number of watchdog trips by subsystem (HEPI,ISI,SUS) per chamber over the course of O2 for some known large earthquakes. We get this data by looking at the state of the watchdogs for each subsystem over the course of two hours during an earthquake using minute trended data. To generate these plots we look at the WD_MON State channels and just check for if they ever deviate from their ideal value (1 in most cases) and sum the number of events that these trip. We also pull the interferometer lock state to a) monitor lock losses and b) to veto any trips caused by frontend crashes rather than earthquakes. 

Also attached is some comma delimited text files containing information on which of these watchdogs tripped first for each earthquake, with a list of GPS times listed for each of these events. Note, again because of the minute trended nature of the data its hard to pinpoint what watchdog tripped first, as watchdogs that trip within a minute of each other will appear as having tripped at the same time. With the data being minute trended, it appears as if the suspension trips first, though this is not the case when examining the each of these events using finer timesteps. Note opening these in web browser will screw with the formatting. 

There are two events where HEPI appears to trip first when examining these events in minute trend these only appear to occur on ETMX, please take this with a grain of salt as this is based off minute trended data, so there may be other trips before HEPI on ETMX, that occur within the same minute, or HEPI trips on other chambers first before ISI trips. I plan to investigate these further with second trended/ full data rate  data.  

1) Resonances of the ITMX elliptical baffle match peaks in DARM. Several peaks in DARM, (e.g 70 and 106 Hz), were thought to be due to the elliptical baffles, either or both ITMX and ITMY baffles. This is because, for different vibration injections, the amplitude of these peaks in DARM were best explained by the vibration level at ST0 of BSC2, and these baffles hang from this stage of the ISI ( https://alog.ligo-wa.caltech.edu/aLOG/index.php?callRep=26016, https://alog.ligo-wa.caltech.edu/aLOG/index.php?callRep=31886 ).  To make sure that this identification was correct, I measured the elliptical baffle resonances with an in-chamber accelerometer while tapping on them. Figure 1 shows that resonances of the ITMX elliptical baffle match the DARM peaks, but the ITMY baffle resonances do not. Betsy will check to see if the ITMX baffle down-tube is misaligned when the upgraded baffle is installed. 

2) Possible sources of scattering in BS chamber. Follow-up PEM injections showed that shaking the walls of BSC2 produced noise in DARM (https://alog.ligo-wa.caltech.edu/aLOG/index.php?callRep=39121 ).

One long term concern in BSC2 is the TCS mirror 2 and its support structure that is attached to the BSC2 wall: https://alog.ligo-wa.caltech.edu/aLOG/uploads/9564_20140126161227_Figure2-ITMXcompensationPlate.pdf .

Scattering associated with the mirror could either be from the support region surrounding the mirror or from the nozzle and flange holding the TCS ports, visible in the mirror (see Figure 2 and https://alog.ligo-wa.caltech.edu/aLOG/uploads/9564_20140126161227_Figure2-ITMXcompensationPlate.pdf ). The region around the mirror could be baffled and the port could be baffled. I talked to Steven about the possibility of a port baffle that might also be useful at the P-Cal transmitter and receiver ports.

Another mitigation possibility is to damp the motion of the mirror. I tried a very simple method illustrated in Figure 3 that seems to have reduced the Q’s of some of the resonances by 2 or 3. I think we could do a lot better by wrapping the struts with more material.

A second possible source of the scattering noise are the BS chamber walls themselves, which are nearly normal to wide angle scattering from the beam spot as illustrated in Figure 3. 

A final possibility is the elliptical baffle: I wasn’t able to eliminate this possibility because in the 14-17.5 Hz band we haven’t made strong enough injections with HEPI to exclude ST0 (discussed here: https://alog.ligo-wa.caltech.edu/aLOG/index.php?callRep=39121).

3) Scattering at P-Cal ports. Nothing new to report beyond https://alog.ligo-wa.caltech.edu/aLOG/index.php?callRep=39121, except that I discussed port baffle design with Stephen Appert.

4) HAM1-HAM2 septum not shielded from spot on beam splitter

Figure 4 shows that the HAM1-2 septum is visible from the BS through the structures on HAM2. The septum is a concern because of its high reflectivity and the fact that it is not seismically isolated like the reflectors on HAM2. The baffles are not yet all installed on HAM2, but I think the baffles will leave the upper regions of the HAM1-2 septum exposed to the BS. There is not an equivalent exposure of the HAM5-6 septum because the central part of the MC baffle in front of HAM5 remains in place, while the central part of the MC baffle in front of HAM2 has been removed.

5) “Temporary” floors should probably be removed for scattering reasons. The floors placed inside the chambers for working are sometimes left in place. Figure 5 shows that they may create scattering paths and should probably not be stored in place. I think Betsy was already planning to do this for charge imaging and other reasons.