The Apollo crew changed the existing C-3 soft roof (one zipper) to one with four zippers.
The PSL ISS out-of-loop photodiode PSL-ISS_PDA is extremely glitchy. This channel was used during ER3 as the fake h(t) channel and was clean except for unexplained glitches below 55 Hz. When checking to see if the low-frequency glitches had changed, I found that this channel is now extremely glitchy at all frequencies, and the spectrum is much higher than previously. I used a 9-hour long lock from May 16 5:23 UTC to 14:37 UTC, where the ODC showed that the PSL should be working correctly. The first two plots below are Omicron searches from ER3 and from this new time. The channel is constantly glitchy. The second two plots are the spectra from the in-loop and out-of-loop photodiode first from ER3 and then this new time. The colors were switched by the plotting program, but it's obvious which one is in-loop.
HAM 1 is now in flushing. HAM 2 is in operations, HAM 1, 3 and BSC 1, 2, 3 are in flushing. The pressure has been holding steady at 31.5 psi with a motor speed ~1983.
The IPS shift on HAM 1 was:
Pre flushing | Post Flushing | ||||
H | V | H | V | ||
1 | 392 | -3,537 | 154 | -4,018 | |
2 | -6,445 | -16,215 | -6,689 | -16,335 | |
3 | -3,409 | -23,234 | -3,640 | -23,359 | |
4 | -2,846 | -1,257 | -3,046 | -1,385 |
This report is for Friday, 17 May. The Apollo crew spent the morning assembling the A-frames for the ISI install. Hugh and the crew also spent some time going through the Hazard Analysis so that that is out of the way.
Scattering from beam tube baffles is more likely to be a noise source in aLIGO than it was in iLIGO. With this in mind, the motion of the beam tube was studied. Figure 1 shows beam-line, horizontal and vertical motion for accelerometers mounted midway up the side of the beam tube near one of the fixed supports. This spectrum was taken at 988m Y from the beam splitter, when the wind was less than 5 MPH. Figure 2 shows beam-line motion spectra for different locations along the beam tube and at different wind velocities. These results are similar to results at LLO here.
The beam tube consists of a series of modular segments, with the great majority being 39 m long and weighing about 4000 kg (A+B segments). The segments are supported close to the middle with a fixed support, and each end hangs from a guided support. The beam tube segments are connected at the guided supports to the next segment in line by bellows that have a measured spring constant of 8.2e5 N/m. Figures 3a and b show that, in the 5 – 20 Hz band, the segments act as rigid bodies, with beam-line motion at the middle and at the ¾ point of a segment being virtually identical, highly coherent and with zero phase difference. Figure 3b also shows that, in contrast, the motion of adjacent segments is not very coherent. This is likely better news than if adjacent sections moved coherently.
The calculated beam-line resonance of a beam tube segment, considering only the bellows springs, would be about 3 Hz, while the actual lowest resonance in the figures is at about 8 Hz. This demonstrates that the central support is the stiffest spring. Near the stations, some of the segments are shorter; Figure 3 shows that the lowest beam-line resonant frequency of a 33.5 m segment is about 9.3 Hz instead of 8 Hz. The frequency goes approximately as mass rather than the square root of mass, suggesting, as one might expect, that there is a torsional component about the central support. A modal explanation of the 8 and 14 Hz resonances awaits modeling.
Since the beam tube segments are moving as rigid bodies in the 10 Hz region, the insulation may not damp the motion much at these frequencies. We may get a chance to test this as John is going to remove some insulation.
Attached are plots of dust counts requested from 5 PM May 16 to 5 PM May 17.
At the request of Hugo, the following software was installed: opsws0, opsws1, opsws4, opsws5, opsws7 - Install python-matplotlib version 1.1.1rc from official Ubuntu 12.04 distributions. For opsws2, which is still running Ubuntu 11.04, build and install version python-matplotlib 1.2.1 from source. The official up-to-date version of matplotlib for Ubuntu 10.04 and Ubuntu 11.04 workstations is 0.99.1, which according to Hugo is not new enough. This also required installing build prerequisites, which ended up being a large installation of software. I will wait for Hugo to see if this installation meets his needs before deciding if it should be done anywhere else. A gentle reminder for those of you building custom software, please bear in mind that we are running many Ubuntu 10.04 LTS workstations, so you should develop for that if at all possible.
Work by ISCT1/IOT2L (LVEA) – Corey Work in H2 Building - Dave Trip to End X – Corey Work on H2 PSL enclosure – Pablo/Michael R. Start Spectra Test on Beam Splitter – Arnaud P. Work on BS Optical Lever – Thomas Vo Work on IOT2L (LVEA) – Cheryl Going to end Y to work on ALS table – Corey Back to work on H2 PSL enclosure – Pablo/Michael R.
#1) I found the hardware in the attached picture shoved under the foil that was protecting my class-A hardware - unknown cleanliness of this package and contents now puts into question the cleanliness of my parts. #2) my parts are now on both tables in the cleanroom by HAM2. If you need to get into this cleanroom, call me.
[Sheila, Kiwamu]
We found that there was a small mistake in the slow control channel linking. This was fixed in the software and committed to the SVN.
The mistake:
The VCOs had two RF monitor channels cross-wired. "ReferenceMon" was taking the ADC signal which was for "DividerMon" and vice versa.
Correction:
In order to correct them we simply swapped the links in the twinCat system manager file, H1ECATC1.tsm. This modification was applied on three corner VCOs, namely IMC, ALS-DIFF and ALS-COMM. This change was then commited to the SVN. Also Sheila modified the link document [1] accordingly. She will be applying the same mods for the VCO at the end stations.
[1] DCC LIGO-E1201049-v4
Mark B and Thomas V While showing Thomas how the numbers in the HLTS OSEM2EUL matrix for the optic (M3) were derived, I noticed a nasty bug: the code in /ligo/svncommon/SusSVN/sus/trunk/HLTS/Common/MatlabTools/make_sushlts_projections.m had M2 in two places where it should have had M3. I fixed it as follows (and committed the fix): OSEM2EUL.M3 = [ 0.25 0.25 0.25 0.25 ; ... % L [ 0.50 -0.50 0.50 -0.50]./M3.PitchArm; ... % P - corrected by Mark B, 5/17/13, had been M2.PitchArm [-0.50 -0.50 0.50 0.50]./M3.YawArm]; % Y - corrected by Mark B, 5/17/13, had been M2.YawArm The corrected output is 0.25 0.25 0.25 0.25 2.40963855421687 -2.40963855421687 2.40963855421687 -2.40963855421687 -4.15973377703827 -4.15973377703827 4.15973377703827 4.15973377703827 We typed these numbers into the SUS_CUST_HXTS_M3_OSEM2EUL.adl screen for PR3 and SR3 and updated the safe.snap files. LLO should do likewise.
The equivalent HSTS file had the same bug and has been fixed also: /ligo/svncommon/SusSVN/sus/trunk/HSTS/Common/MatlabTools/make_sushsts_projections.m . However the M2 and M3 levels have the same lever arms, so it doesn't make any difference to the values in MEDM.
Andres took Transfer Function and Power Spectra data on I1-SR2. The results look positive and the plots are posted below. All scripts and data files have been committed to the SVN vault. We are waiting for the testing group to review the results before proceeding.
After looking through the undamped and damped transfer functions provided for SR2, these TFs raise absolutely no concerns, since all measurements are in good agreement with the model and are consistent with other HSTS suspensions. Power spectra plots now include the expected sensor noise (black dashed line), which shows the measurement agrees with the prediction in the majority of cases. Where there's a discrepancy between the expected sensor noise and the measurement (e.g. M2_WIT_P_DQ and M2_WIT_L_DQ) SR2 measurements are consistent with previous HSTSs, implying the issue is most likely with the calculated sensor noise. Thus, I will investigate this sensor noise calculation further... In the interim, please consider SR2 approved for India. Nice work!
Attached are plots of dust counts > .3 microns and > .5 microns in particles per cubic foot requested from 5 PM May 15 to 5 PM May 16. Also attached are plots of the modes to show when they were running/acquiring data.
Attached are plots of dust counts > .3 microns and > .5 microns in particles per cubic foot requested from 5 PM May 14 to 5 PM May 15. Also attached are plots of the modes to show when they were running/acquiring data.
I moved all of the MC TRANS optics about 6 inches toward the MC REFL path, to compensate for the change in the steering mirror pointing. There are some optics and electronics yet to be installed, but the analog camera and PD100 are installed and cabled up. Kiwamu helped with moving the periscope to it's new location, and we both noticed that the MC TRANS beam is going to be near the edge of the viewport as it exits the vacuum system. How close to the edge is unclear until we have a beam.
The PLL at the Y end has been locked for about 30 minutes. I've changed the slow servo back to the old code, which has a simple 1/f response if Pf is set to zero. The settings are: In 1 Gain = -15 dB fast gain 4dB Slow servo "ugf" -0.055 Pf=0 With these settings the UGF is 14.3kHz, the phase at the ugf is -100 deg. From the transfer function it looks as though higher gains would be possible, but if you add more than 3 dB of gain an oscillation starts. As far as I know there are no diagnostics in the slow control system to detect the oscillation.
Craning Forklift over Xarm in the morning
Ski visitors: Snow Valley & other contractor
PCal Alignment & laser hazard in H2 LAE (Pablo/Rodruck)
H1 PSL Shutter closed at 10am (Kiwamu)
Praxair on site at 11:05
ISC (Sheila) & HEPI (Hugh) work at EY
HAM4 Doors Off
BS Oplev work (Thomas)
IO MC Trans periscope move (Cheryl)
Now that BSC3 is on the test, we are working on gettting it checked out and connected for testing prior to integration. Today, Fil helped me re-route the in-air cables (because this ISI is sitting 180 from the previous 3 ISI's, so the cables no longer reached the right corners), then Greg helped me ring them out with the emulator, to verify the routing. After that, I powered off the electronics, wired everything up I could (some of the CPS in-air cabling is improperly gendered, requiring adapters I didn't have on hand) and powered everything up. Tomorrow, balancing and lockers, and if we are lucky, TF's over the weekend.