Attached are screenshot for BSC & HAM CPS spectra. Nothing notable for high frequencies.
Fairly quiet so far. (although the useism has taken a step up in last 18hrs & we continue to have sizable EQs over last few hours).
LVEA Zone Temperature StripTool looks fairly normal/flat around 20.1degC for last 43hrs.
Went through most of the Ops Checksheet with nothing to note.
On the CDS Overview, the edcu is red (this is due to BRS work).
DOE Ground Water Services was on site yesterday morning to perform annual testing of our well. I had to manually run the well for ~ an hour which is longer than I expected and as a result, I overfilled the well tank. There is a high level alarm sounding in the maintenance building but it is nothing to be concerned with, the annoying alarm will silence when the level drops.
IO alignment is complete, until final check, and until after a fix for IM2 and IM4:
Other IO changes to the HAM2 payload:
I've updated the BSC-ISI model in the repo and it is ready for installation. - HAM-ISI model is still pending (should be very soon).
Following requests from Jim, Sheila, et al, this change will update the BSC-ISI watchdog to make the ISI more robust during earthquakes. We have changed the model to increase the hold-the-damping-on-even-if-stuff-is-saturating time from 3 to 60 seconds.
The updated model is in the userapps repo at SVN revision 16398
Installation instructions and more technical details are in technical note T1700481.
The approved ECR is E1700367
and the FRS is 9309
I've not pulled work permits.
-Brian
[Spelling Ed from DAY Ops Shift duties so he can join IO Team for HAM2 activities.]
Day's Activities:
This afternoon, Travis and I finished moving the SR2 into it's new position 5mm closer to the BS (+Y). We had to jockey dog clamps and pushers to do this, attempting to hold the pointing of the suspension as close to it's previous condition as possible. Of course, we will realign this suspension ~next week when we run the PSL beam through the corner from PR to SR and everything inbetween. (We need to resuspend the SR2 and check for health via TFs first, as well as fix the SR2 UL M2 AOSEM, and finish HAM3 and elliptical baffle builds.)
While on HAM4, I reseated the 4" HWS mirror/mount that I moved last week, and we removed all tools and misc hardware lingering from the HAM4 work thus far.
Knock on wood that today was our final day of tedious alignment of the ITMX - we managed to get the suspension free of all mechanical interference points, while simultaneously having it point to the center of the OPLEV (with only ~30 and 120 urad of P and Y bias slider help), with all 20 AOSEMs centered in their range.
We need to rerun the full suite of TFs, but wish us luck!
Betsy, Travis
Re-terminating of the Y2-8 ion pump connector work will continue tomorrow. Had issues with connecting one of the safety interlock connections.
Hugh, Jim, Krishna
This is regarding the idea of placing a seismometer on the BRS-Y platform in order to improve coherence and hence tilt-subtraction. The brief summary is that we found excess low-frequency noise in the seismometer when it was placed on the table on BRS-Y. Hugh had attempted to reduce the noise by adding thermal insulation to the table and the instrument. The last alog on this is 35426.
In the recent data I looked at, this excess noise looks non stationary - it seems to get worse with elevated wind. This suggests it is likely not thermal. The seismometer data shows that the table has much higher motion in the 10-100 Hz as compared to the ground. Since the seismometer isn't held securely on the table, it is possible this excess high-frequency motion is causing it to move ever so slightly, which could look like excess low-frequency noise (step changes). We are attempting to mount the seismometer securely to the table using M10X1.0mm threaded rods and holes to be machined into the table. Results and plots will be posted in a few days.
How is the table secured to the ground?
The table on which the BRS sits, is sitting on three legs, not sure if these are hard point or rubber mount... Good question, Krishna?
J. Kissel
Taking the data from this morning's aLOG as my target frequencies, I've tuned BRD S/Ns 015 and 016 to match the highest vertical (a.k.a bounce) and roll mode frequencies of the newly suspended H1 SUS ITMX. Results are attached and tabulated below:
'Target [Hz]' 'BRD S/N' 'Mean Fit Freq. [Hz]' '% Diff [%]'
'9.806' '015' '9.853' '0.49'
'' '016' '9.812' '0.066'
'' '' '' ''
'13.927' '015' '14.033' '0.76'
'' '016' '13.926' '0.009'
Setup for measurement is as described in the HAM ISI TMDs (see 38741) using corroborating evidence from both the laser vibrometer and a primitive shadow sensor setup. To calculate the above "Mean Fit Frequency," I used a least square's minimization with a Lorentzian seed function,
fit(f) = (1/Q)^2 ./((f - f0).^2 + (1/Q)^2) + C
to fit both sensors' data to obtain the measured frequency with the noise, and then took the mean of the two sensors' results.
Betsy and Travis are installing these BRDs as I type.
Krishna, Jim, Hugh,
Today, we began the task of replacing the gravitational actuator/damper with capacitive actuators for the BRS-X. This will make it nearly identical to BRS-Y in control and behaviour‌ and eliminate the few incidences of damper problems we have had with BRS-X.
We opened up the vacuum can and machined a hole in the aluminum thermal shield for the capacitor plate wire. The capacitor plate was then glued to the thermal shield using TorrSeal epoxy, which takes a day to set fully. Tomorrow we will complete the wiring. We will also attempt to reduce the offset between the center of mass and the pivot axis to make it less sensitive to horizontal acceleration.
Sometime over the last few days, we lost functionality of the F1 (Face 1) BOSEM on the main chain of ITMX, causing a bit of hunting and measuring. The chain would not damp and although there were peaks in proper places on all TFs, the noise floor for the PITCH measurement was well below where it should be. This pointed to F1 as having issues and hence the swap.
S/N 091 came out, 049 went in.
| WAS (091) | IS (049) | |
| OLV | 17,900 | 29,350 |
| OFFSET | -8,950 | -14,675 |
| GAIN | 1.676 | 1.022 |
Note, we are seeing that many OSEM OLVs have been decaying over time. So, while we aren't surprised that 091 was much lower than it should be, likely we need to take a closer look at all BOSEM trends.
WP7196, DCC-T1700481, ECR-E1700367, FRS-9309
Brian, Jim, Dave:
h1isi[etmx, etmy, itmx, itmy, bs] models were compiled and restarted. Their DAQ channel lists increased by six slow channels each (example list for ITMY shown below).
The isi2stagemaster.mdl was modified. It now uses C code in the file ISIWD_LONGDAMP.c in place of ISIWD_GPS.c
+: slow channel H1:ISI-ITMY_ST2_WD_MON_RESET_PENDING added to the DAQ
+: slow channel H1:ISI-ITMY_ST2_WD_MON_HOLD_REMAINING added to the DAQ
+: slow channel H1:ISI-ITMY_ST1_WD_ST1_FLAG_GEN_DELAY_TIME added to the DAQ
+: slow channel H1:ISI-ITMY_ST1_WD_ST1_FLAG_GEN_DELAY_TIME_MON added to the DAQ
+: slow channel H1:ISI-ITMY_ST1_WD_MON_RESET_PENDING added to the DAQ
+: slow channel H1:ISI-ITMY_ST1_WD_MON_HOLD_REMAINING added to the DAQ
We have added a couple of convenience functions to LDVW that might be of interest. BRUCO (brute force coherence) written and maintained by Gabriele Vajente, WDQ (omega scans on a standard list of channels) uses either the Matlab W pipeline maintained by Brennan Hughey or dmt_wscan written and maintained by John Zweizig. LDVW provides a simple web interface that launches a condor job to run these processes. Results are available on the leas-jobs machine at CIT. BRUCO interface is available at https://ldvw.ligo.caltech.edu/ldvw/Bruco WDQ is at https://ldvw.ligo.caltech.edu/ldvw/Wdq If any part of the interface is counterintuitive or you have any trouble running please contact me at joseph.areeda@ligo.org
J. Kissel I'm headed into the optics lab to tune the Bounce / Roll Dampers (BRDs) for ITMX, so I want to reconfirm the measurements of the highest vertical (a.k.a. "bounce") and roll modes of the newly suspended ITMX. The previous measurement of the highest vertical & roll modes (LHO aLOG 39106) was taken while the SUS was still rubbing all over and poorly aligned. Travis and Betsy worked hard yesterday to find that the majority of Top to Top mass transfer functions were clean (barring a believed malfunctioning F1 sensor); aLOG pending. The results are reported below (and ASDs of the PUM/L2 and UIM/L1 stage are attached). Meas Date 2017-10-20 2017-10-31 V4 [Hz] 9.8022 9.8056 R4 [Hz] 13.927 13.927 The frequency bin width of both measurements is 0.5 mHz, and the effective bandwidth is 0.732 mHz. The bounce frequency has increased ever-so-slightly (9.8056 - 9.8022)/9.8022 = 0.00035 (i.e. 0.03% increase), but the roll mode remains the same. Given our ability to tune the BRDs is at the 1% level, I expect / will be happy to get something between the 9.8056 .* (1 + [0.01 -0.01]) = [9.9037 9.7075] range.
Richard came into my office this morning to tell me that the laser was off. Pulled up the laser status screen to find that power watchdog was activated ~12 hours 20 minutes ago (~6 pm Monday). The laser was brought back.
As Peter says, the laser tripped due to a trip of the power watchdog. Looking at the relevant channels at the time of the trip, the 35W FE watchdog tripped first, with the HPO watchdog following ~6 seconds later due to the loss of the FE (see attachment). Beyond that, I can't find any reason for the FE power watchdog to have tripped. All channels for the FE and NPRO appear to be behaving normally up to the trip, then everything cuts off (as it should with a power WD trip). Unclear at this time why the FE watchdog tripped. Maybe another occurance occurrence of the mysterious NPRO shut down that we can't explain but seem to see a few times a year?
Edit for spelling at 10:30 PDT.
Filed FRS 9337 for this trip.
We usually blame NPRO shut-downs on the UPS the NPRO is plugged into. Maybe that's what happened here again?
It appears that some process variables are taking on values with characters that do not conform to JSON: Oct 18 11:10:12 conlog-master conlogd[7334]: terminate called after throwing an instance of 'sql::SQLException' Oct 18 11:10:12 conlog-master conlogd[7334]: what(): Invalid JSON text: "Invalid escape character in string." at position 44 in value for column 'events.data'. Oct 18 11:41:52 conlog-master conlogd: Server started. prefix: 'H1:CDS-CONLOG_' database: 'conlog' max inserts: 500 max time: 30 keep alive period: 30 time source: 0 create_messages: 1 Oct 18 11:59:32 conlog-master conlogd[8524]: terminate called after throwing an instance of 'sql::SQLException' Oct 18 11:59:32 conlog-master conlogd[8524]: what(): Invalid JSON text: "Invalid escape character in string." at position 44 in value for column 'events.data'. Oct 18 12:19:34 conlog-master conlogd: Server started. prefix: 'H1:CDS-CONLOG_' database: 'conlog' max inserts: 500 max time: 30 keep alive period: 30 time source: 0 create_messages: 1 Oct 18 12:34:12 conlog-master conlogd[9092]: terminate called after throwing an instance of 'sql::SQLException' Oct 18 12:34:12 conlog-master conlogd[9092]: what(): Invalid JSON text: "Invalid escape character in string." at position 44 in value for column 'events.data'. The same error repeats at each loading of the channel list.
Restarted the MySQL server with the general query log enabled to log the SQL query causing the crash. Restarted conlog. Crash did not occur on restart this time. Will leave log enabled for now.
The parts in question are D1101911 & D1101910.
HAM2: I've installed a QPD cable strain relief on IM4 trans. See first attachment.
HAM3: Done with MC2 trans and POPB, but not POPA.
For POPA, 1/4-20 screw wouldn't go in easily (I only tried for a minute or two) and I didn't bother to keep trying because I remember that this was a problem in the past for some of the QPDs (alog 19168 for TMSX).
One set of parts is left on the wire shelf on the work table in HAM2 cleanroom (top right of the wire shelf). If the 1/4-20 wouldn't cooperate, I would recommend to do the same solution as the above alog entry.
Done with ASC QPD strain rellief.
POPA strain relief was installed as planned using a long undersized screw (8-32), a nut, and some vent washers.
All segments of POPA, POPB, MC2 trans and IM4 trans responded to flashlight.
(Posted by Keita using Ed's account.)
