The guardian and cdsutils (ezca) libraries have been upgraded:
Mostly just bug fixes and some performance improvements.
All SUS and ISI guardian nodes were restarted. ISC nodes were not restarted but should be restarted soon.
David Hosken, Thomas Vo, Greg Grabeel Ran the chiller pumps starting from around 2pm until 5pm. Only one small leak on the loop back hose at BSC 3 was found. BSC 1 had a flow rate of 7 lpm when running in a partial bypass pressure of 24 psi. BSC 3 had a flow rate of 7.5 lpm when running at a partial bypass pressure of 21 psi. Both were running at 20* C and had a very minimal fluid level drop. Hopefully a longer test tomorrow will yield the same positive results. Good job Apollo.
dust monitors 1 and 15 going in and out of communication prep work for opening HAM5 doors 10:02 - 11:02 Jeff B. installing dust monitor 4 in clean room over LVEA test stand 10:19 - 11:23 Filiberto checking ESD cable connections at end Y 10:47 Betsy, Travis staging equipment in LVEA, opening rollup door 11:00 BS ISI left with Start blend filters (Patrick, Corey, Jeff K.) 11:03 Andres working on SR3 11:05 - 13:11 Gerardo working on OFI in H2 PSL enclosure (no laser) 12:58 - 15:01 Karen cleaning at end Y 13:16 Greg, David H., Thomas going to test TCS chiller lines in LVEA 13:46 Cris to end X VEA to get Acetone 14:16 Gerardo working on OFI in H2 PSL enclosure (no laser) 15:17 Kyle done testing valve in of temporary RGA at Y1 16:20 end Y going to laser hazard
(Craig, Doug, Travis, Jason) We set the camera pointing path by mimicking the camera view using a laser pointer. An existing fixture was modified and worked well and mounted up well. We aligned the north side camera view without much issue. The south side required a bit more tweaking of the internal steering mirrors. The procedure we used involved retro-reflecting a laser pointer through a 1/8" pin hole centered on the viewport. We steered the beam with a couple of turning optics on the fixture. We then propagated the beam through 2 steering optics permanently mounted to the periscope inside of the beam tube and then onto the P-cal target mounted in front of ETMy core optic. All P-cal hardware is removed from the vacuum chamber and repackage for storage.
(Craig, Doug, Travis, Jason) We set the camera pointing path by mimicking the camera view using a laser pointer. An existing fixture was modified and worked well and mounted up well. We aligned the north side camera view without much issue. The south side required a bit more tweaking of the internal steering mirrors. The procedure we used involved retro-reflecting a laser pointer through a 1/8" pin hole centered on the viewport. We steered the beam with a couple of turning optics on the fixture. We then propagated the beam through 2 steering optics permanently mounted to the periscope inside of the beam tube and then onto the P-cal target mounted in front of ETMy core optic. All P-cal hardware is removed from the vacuum chamber and repackage for storage.
Mitchell, Jim, Gerardo, With help at key times, while I was on the forklift, the bottom half of the monolithic was able to be completely heli-coiled and parked onto the test stand. The top half has been parked onto the granite table for it's turn at a heli-coil makeover.
Verifying predicted response in preparation for pressure accumulation data collection on Y1 this week.
The ETMX decoupling filters for top mass drive to pitch motion at the test mass are now installed in the appropriate L2P bank (split in two filters in FM1 and FM2), and using a gain of -1 (cf 2nd attached screenshot L2P_medm)
It reduces the pitch motion from length drive by a factor of ~7 @ .44Hz and ~5 @ 1Hz (cf first screenshot ETMX_L2P_testing), comparing the L2P transfer function with/without the filter.
Method :
1) Measured L2P and P2P transfer functions using following templates
/ligo/svncommon/SusSVN/sus/trunk/QUAD/H1/ETMX/SAGM0/Data/2014-03-17_H1SUSETMX_M0_LtoPY_WhiteNoise_0p1to10Hz.xml
/ligo/svncommon/SusSVN/sus/trunk/QUAD/H1/ETMX/SAGM0/Data/2014-03-03_H1SUSETMX_M0_PtoPY_WhiteNoise_0p1to10Hz.xml
2) Compared the measurement with model, fitted the measurements, inverted P2P and calculated L2P/P2P using the script LP_decoupling.m located in
/ligo/svncommon/SusSVN/sus/trunk/QUAD/H1/ETMX/SAGM0/Scripts
The fitting and inverting process is summarized in the attached pdf.
Note : There's a factor of 11 between fit and model at DC
J. Kissel We're really getting pummelled by ground motion, especially the past few days. I attach plots of the band-limited RMS in the Z direction of PEM ground inertial sensors in the corner station and the X end station over the past 48 hours. We've been hit by earthquakes (in the 30-100 [mHz] band), high-micro seism from windstorms and spring ocean activity (in the 100-300 [mHz] band), and gusts of wind and anthropogenic noise (in the 1-3 and 3-10 [Hz] bands). I think this is an excellent representation of why we've had so much trouble trying to find a configuration of the isolation systems that "just works" -- the characteristics of the input motion radically change constantly on a ~3-hour timescale. Not to mention there are times (like in the middle of the plot, 2014-03-16 21:50 UTC) where the earth is just down right angry with all three bad things at the same time. I apologize for the ambiguity in calibration, I have an email out to Robert asking the proper calibration into physical units. The FOMs with these channels on the wall suggest that the channels are calibrated into "[decaum/s]" in the frames (which I think is 10 * 1e-6 [m/s]) but with no calibrations in the DMT template (and none on the data viewer trace, obviously), I get different numerical values. pem.ligo.org suggests that both the corner-station and endx should be 0.0076e-6 [m/s]. I think whatever units they are, they're at least self-consistent. I'll take spectra at choice times to gather the diversity; stay tuned.
Here're the promised, associated spectra. I attach an annotated version of the BLRMS too, with vertical lines demarcating the UTC time in the same colors as the traces on the spectra. As indicated by the legend, I chose representative times that reflected the variety of ground motion: 2014-03-16 15:45 UTC - High micro-seism 2014-03-16 18:05 UTC - High micro-seism + Gusty Wind 2014-03-16 21:45 UTC - High micro-seism + Large Earthquake, in Chile + Gusty Wind 2014-03-17 05:50 UTC - Medium micro-seism + Cluster of Small Earthquakes (Chile, Again in Chile, Philippines, Puerto Rico, and Mexico, at magnitudes 6.2, 5.0, 5.1, 3.0, and 2.8 respectively). 2014-03-17 18:45 UTC - Low micro-seism + Gusty Wind Notes: - While we don't necessarily expect to be able to stay up during large earthquakes, it's good to see a sort of upper bound, and the contrasting color of the ground motion as it fills in the 30 to 100 [mHz] band. - It's interesting to see that during periods of high microseism, the peak at 55 [mHz] seems to be present regardless of whether the winds are gusting or not (comparing BLUE and RED) . - Further, there exists times (i.e. MAGENTA) where there is lots of gusting winds, but the microseism is relatively low. - Most importantly we see that ground is sort of oscillating between requiring more isolation in the 30 to 100 [mHz] band and in the 1 to 10 [Hz] band. - The variations in the 30 to 100 [mHz] band can be as much as a factor of 100 during earthquakes, but not too much change between windy and not windy, maybe at most a factor 5 in a small range between 40 and 80 [mHz]. - There's very little to no change in the region of the lowest SUS resonances between 0.2 and 0.6 [Hz] between all of these seismic bad times - Naturally, gusty wind times are the worst at high frequency, but again the variation is only a factor of 3 to 4. We should be able to ride this out. Let's see what the HEPI L4Cs say, maybe the gusts have an interesting affect playing with the pier flag poles.
J. Kissel Here're the HEPI L4Cs for each of the times described above. While they show the same sort of variety at the microseism and below (where they can sense the ground motion above their self-noise), the most interesting is how much more varied the tilt spectra are (in this case RY) in the 0.3 to 1 [Hz] band -- exactly where the translation degrees of freedom claim there's very little difference. And naturally the ITM reacts differently than the ETM, just to make things difficult. For ITMX, both high microseism conditions and gusty wind show elevated tilt in the 0.3-1 [Hz] band, by about a factor of 10. Only the during the medium microseism + small earthquakes did the L4C hit it's (tilt) noise floor. For ETMX, there's quite a variety, with gusty winds alone being the worst-case scenario. But the moral of the story is that, where translation degrees of freedom get tossed around by an order of magnitude between 0.03-0.3 [Hz] and 1-10 [Hz] during these excurisions, the tilt fills in the 0.3-1 [Hz] gap with its own order of magnitude variation, and as we know *both* tilt and translation of the table affects the optics just about equally. So -- if we want to operate the isolation systems during these 90th percentile times, we're gunna have a lot of automation and blend tailoring to do, but just increasing the blend frequency won't necessarily win. Further, when we *do* #movetheblendup, the green still can't handle the loss in performance.
The data for the above measurements can be found in the following templates:
/ligo/svncommon/SeiSVN/seismic/BSC-ISI/H1/Common/Data/
2014-03-16to17_GroundMotion_ASDs.xml
2014-03-16to17_HEPIPierMotion_ASDs.xml
There's way to many curves for me to export, but all curves shown are stored as references so you can export them at your leisure and use them in any models.
Ground Motion Pier Motion
2014-03-16 15:45 UTC Ref. 0-11 Ref. 0-23
2014-03-16 18:05 UTC Ref. 12-23 Ref. 24-47
2014-03-16 21:45 UTC Ref. 24-35 Ref. 48-71
2014-03-17 05:50 UTC Ref. 36-47 Ref. 72-95
2014-03-17 18:45 UTC Ref. 48-59 Ref. 96-119
All plots and templates are committed to the repository as of this entry.
After getting TF's Thursday night that look much better than the previous ones, I've been able to make pretty good progress on ETMY. Friday afternoon I was able to install damping and blend filters. This morning/afternoon I installed a level one controller, then copied it to the level two and three slots, much like we did at HAM4 for the guardian testing. I still have a couple of tests to run on the ISI (no more than a couple of hours total) and write the report for approval, but we have most of what we need and I think it looks pretty good. The attached screenshots are two sets of seismometer transfer functions from thursday night and a set sensor spectra from this morning.
I dug up some data for ETMX for comparison. Too lazy to put it all up, but I'm posting the GS-13's from ETMX. Looks pretty good, considering the ETMX tf was taken in vacuum.
After Travis had notified me of a problem on ETMy where the copper clamps on the ring heater were touching when it was moved into final position I went in chamber and made some adjustments to keep the upper and lower clamps separated. Decided to check if the same problem existed on ITMx and ITMy. Unfortunately, ITMy had part of its macor break while I was adjusting the copper clamps. ITMx had the glass former break sometime after it's installation onto the lower quad. Both lower ring heaters have been removed.
[ FYI ... There is a specially designed Ring Heater Segment Replacement Fixture (D1101253), which is to be used ANY time a segment needs to be removed from a QUAD, if a dummy mass or a TM is also on the QUAD ]
Note, the ITMx unit has already been stripped of it's lowest dummy mass in prep to load the new glass mass later this week. So, the ITMx unit had extremely easy access for this RH work and no fixturing was required.
It's worse than I originally thought. The glass former broke along with the macor on the ITMy lower ring heater.
Apparently, damage was done (also) to the very tip-end of the glass former when the adjustment was made (14-March) to the copper clamp plates of the lower RH segment (assembly D1001895-v8 SN-210) on the ITM-Y quad. This damage was not revealed until the lower segment was dis-assembled. Photos are attached.
Regarding the lower ITM-Y assembly issues, see the attached (PDF) package of images.
The following feedback (attached) has been received, as guidance, from SYS
The aLOG will be down for maintenance tomorrow (Tuesday 18 Mar 2014 starting at 12:15pm pacific). Please save or publish all log entries prior to the outage. A reminder entry will be posted to the logbook tomorrow.
Today we have a different type of interesting seismic time, the low frequency ground motion due to the earth quake has subsided, but we still have microseism well above the 90th percentile (the wind has also died down). While the microseism is high, it seems like these might be conditions we want to be able to ride out in the long run. ETMX and ITMX are both isolated with Tcrappy (no sensor correction), both have pitch fluctuations of about 0.3urad, but yaw fluctuations of more than 1 urad. Right now it is UTC 1:28.
A screenshot of the ground motion and Op Levs are attached, the .xml file which also includes SEI sensors is available here
Is this ground motion just too much for us to ride out, or is there a configuration for the ISIs that would be better under these conditions?
Though they haven't been confirmed to work well under these conditions, you should try switching to the "Start" filters (which have a high blend frequency) on ST2, and switching on some modest ST1 to ST2 sensor correction. You lose performance at 1 [Hz] but improve microseism by a factor of 10ish. You can see the variation in performance from LHO aLOG 10728, specifically, the attachment. Note, we can also try the T750 [mHz] blend filters on ST1, but I don't have a good aLOG of their performance. My guess is that Seb didn't have time to characterize this configuration. #movetheblendUP
