Lisa, Evan
Attached is a spectrum of the suppressed DARM error signal from last night. The RMS is about 3×10−13 m down to 0.1 Hz, and (as we've already seen) almost all of it comes from motion between 1 and 5 Hz. By looking at the recent DARM OLTF for H1 (LHO#17153), we see that we easily have more room for a boost below 10 Hz or so (and Dan is already working on this).
Also attached are spectra of recent "best" locks from H1 and L1, along with the GWINC prediction for 10 W of input power.
P.S.: The rms traces of the calibrated DARM signals in DTT appear to be completely bogus. The attached plot was done independently in python.
P.P.S.: Attempting to print these spectra in pdf, eps, or ps causes DTT to crash.
Some plots with the best L1 and H1 sensitivity curves so far.
The attached image describes the effect of the boost filter we'd like to add to DARM. The blue trace is the current set of DARM integrators, boosts, RGs, and so on (I haven't included the suscomp filter). The red trace has the new boost in FM7. The blue trace is the same as the comparison that Peter made between H1 and L1 back in February. Compare the red trace in the attached plot (the change we'd like to make) to the L1 trace in his comparison; this boost will bring us more in line with the Livingston loop shape. (The changes made two weeks ago to the DARM loop were made to the suscomp filter.)
The boost adds 20dB of gain at 3Hz, takes away ~1.5dB at 10Hz due to a pair of complex zeroes, and loses about 4deg of phase at 50Hz. The last DARM OLTF we measured says we have 45deg of phase margin at the 50Hz UGF, and probably about 30dB of gain at 3Hz, so this filter should significantly improve the suppression in the 2-5Hz band without costing too much phase.
We might want to add an RG filter at 0.45Hz to suppress the pendulum resonance, too. Or increase the existing RG in FM3.
Dan, Sheila
We had a quick look at last night's lock, it seems that the degradation in the range over time was not due to an alingment drift, since we can see that the build ups stayed high durring the lock. It seems like the wind picking up again this morning was pretty well correlated with the range decreasing. We don't know how to get DMT sensmon data (or any DMT data) from the control room. Dan made this plot by logging into the cluster to find the DMT sensmon frame files.
Dan, Sheila
The attached screenshot shows a one second trend of the beckhoff channel that is the fast shutter trigger readback, along with two RCG channels, the first is AS_C sum, this PD should be the source for the fast shutter trigger, and ASAIR_LF, which should see similar power fluctuations.
The beckhoff channel seems to be 0.2 seconds ahead of the RCG channels. This timing difference sometimes makes it appear as though the fastshutter shut first, causing the lockloss, as in the second attached screenshot.
We would like to prove definteively that the shutter is only shutting when it should, but we don't have a readback that we trust.
Dave and I looked into this somewhat and we don't think that the timestamps of the EtherCAT systems are used at all by the frame builder. As a matter of fact the internal clock of these machines was about 3 seconds off. Since it is hard to believe that the EtherCAT systems violate causality, the most likely candidates are frame builder, data concentrator or maybe the EPICS gateway. This would also mean that all data transported through EPICS to the frame builder could be affected. More investigations are pending.
Following up on my log 17295, the 10 hz CPS low pass elliptical filter is now running on all BSC ISI's. I've spent today working on ETMX trying to push the St2 filter down more and adding a similar filter to St1 but it's not ready for show time, so only the "old" filter mod is running.
First attached plot shows the X and Y St2 sensors, X is red ( gnd STS, St2 CPS& GS13), Y is blue, Y is running the "old" elliptical filter, X is running the more aggressive setup. You can see that from ~2 to 15 hz the more aggressive (red) setup does better. Unfortunately, the St1 elliptical filter interferes with the sensor correction (changing the phase of the CPS plant), so the ground STS signal doesn't cancel the CPS signal as well, increasing motion by ~2x at ~.5 hz. Or maybe the ground is a little worse, idk. I plan on tuning this a little more, maybe modifying the sensor correction as well (when I understand it better), and will try again when the opportunity presents itself.
The next two plots show changes I've made to the blends. Second plot is St1, red is stock, green is the filter I tried today(with problems at .5 hz), blue is what I will try next.
Third plot is St2 blends, red(not installed) is new, blue is the only thing running right now.
model restarts logged for Mon 23/Mar/2015
no restarts reported
Time Code Translator work at end stations, WP5116
Dave, Jim:
We racked up the missing TCT in EY, then connected it to one of the unused fiber pairs running between the VEA rack and the computer rack. We verified it was using ports 19 and 20 in the MSR (same as EX). Once connected the TCT synced to the time and the error LED on the MSR sender was extinguished.
We then ran 30 foot 50-Ohm coax, BNC terminated, between the TCT in the computer rack, through the wall, to the first port on the CER comparator. This was done at both EX and EY. The 1PPS port on the back of the TCT is a TNC connector, so we used a TNC-BNC adapter on that port for EX. We need to do the same for EY.
At EX, the comparator is now displaying a time of -22.2uS on the TCT port, roughly the time of flight between the buildings. We will revisit the TCT internal settings, they are most probably what was left from iLIGO.
Rebuild of ASC model
Daniel, Dave:
Daniel though h1asc would benefit from a clean build, install and restart. This was done, I'll see if this helped in the IPC errors.
PCAL filter modules, port from PEM file
Sudarshan, Dave:
Sudarshan discovered that I had forgotten to migrate the PCAL filters from when the CAL system shared a core with the PEM. He copied them from H1PEME[X,Y].txt to H1CALE[X,Y].txt, making a CAL_PCAL to PCAL name change in the process. The filters were then loaded into the PCAL models at both end stations.
The TNC to BNC adapter has been installed to connect the EY TCT to the timing comparator. Remaining work is calibration of the CsIII frequency standard, the time distribution system in the corner stations, and the TCT at the end stations.
(Kyle, Gerardo)
Turned off and decoupled two pump carts, from HAM1 and HAM2 annulus system.
The annulus ion pump for HAM1 still not connected to CDS, HAM2 is.
Scott L. Ed P. Chris S. Joe D. Completed tube cleaning to single door between X-1-6 and X-1-7, 40 meters today. Test results for cleaned section and next dirty section are posted here. Installed 70 support tube caps on previously cleaned sections and more on order. Continuous monitoring of beam tube pressures by control room operator during cleaning operations.
Since we are frequently running out of range on the end station VCOs, when there is high wind, we switched the drive of the fiber AOM back to the IMC VCO. This was the original scheme. It has the disadvantage that the end station green locking now depends on the IMC to be locked. On the other hand, this is probably not a big deal, when working with the fully locked interferometer.
Kiwamu, Sheila,
We found that we had to lock ALS DIFF and COMM sequentially with this change, instead of in parallel as we have been doing. This morning we found that we could not transition the tidal to common with this change, so we reverted it so we are again using the fixed frequency RF source for the fiber AOM. We labeled both of the cables as well, so that next time the wind picks up we can put this back. We will probably not need high bandwidth common tidal with this modification, so we can try turning this bandwidth down and leaving off the IMC F offset, then later engaging it with a slow ramp.
07:00 Cris and Karen into LVEA
08:08 IFO Fully locked at LOWNOISE_ESD_ETMY - ~18Mpc.
08:15 Gerardo into LVEA
08:25 Gerardo out of the LVEA
08:49 Joim installing new isolation filters on BS
08:49 Heintze out to LVEA -TCS
08:57 Gerardo out to LVEA - HAM1
09:08 Corey out to squeezer bay
09:14 Ellie out to LVEA - HAM1 area
09:18 Betsy out to LVEA
09:21 Jim W and TJ to EX
09:30 Jim B to EY to install Time Code Translator.
09:21 Jim W and TJ back from EX
09:53 Gerardo out of LVEA - Fil and Andres are still in pulling RF cables around HAMS 3,4
10:00 Jodi and Gary out of the LVEA
10:23 D Barker and J Batch back out to EY
10:25 HFD on site
10:58 Doug Cook out to LVEA - running a box of parts out to a cabinet.
11:08 P King into H2 enclosure
11:15 D Barker restarting ASC
11:42 P King out of H2 enclosure
12:07 Matt outnof LVEA
12:08 Dave amd Jim back to EX
12:10 Fil and co. out of LVEA
12:38 Dave and Jim back from EX - DAQ restart imminent
12:42 DAQ restart
12:56 Karen @ EY
13:43 Karen Leaving EY
14:50 Fill out to LVEA to make some measurements for cables
15:00 Kyle and Gerardo to MidX
15:07 Fil out of LVEA
This is the analysis of a quiet stretch of data in SR3 oplev from yesterday.
I checked to make sure of the following:
a) Alignment sliders have not been moved during this period
b) There is no actuation to SR3 at this time.
c) The signal levels on all SR3 quadrants are okay, that is the quadrants are not saturated.
I am posting here screenshots of plots of:
1) the time series for 16 hrs at 1 sec sampling rate
2) the time series for a 2 hr stretch with 256 Hz sampling from within the 16hr stretch
3) a RIN spectrum of the SUM signal for the 16 hrs sample going down to 10^-3 Hz to 0.1 Hz
4) a RIN spectrum of the SUM signal for the 2 hrs sample going from 0.1 Hz to 100 Hz
5) A spectrogram of the 2 hrs signal showing the broadband noise injections due to the glitches in the laser.
Note:
All data for this analysis is available in /ligo/data/oplevs at LHO. File name LHO_SR3_before_LD_Fix.mat
OPLEV Overview Screens UPGRADED
Following Stuart Aston's suggestions I have added:
1) a bar indicator to show the SUM value. This enables us to tell at a glance if the value is close to saturation
2) a set of menu buttons to call data analysis tools like Diaggui and Dataviewer, which open corresponding templates so that we can look at time series and spectra of relevant channels.
3) a button to every oplev screen which accesses the SUS or ISI corresponding to the Oplev, so that, if the oplev is misaligned, the corresponding suspension or table can be reached directly from this screen.
SITEMAP updated:
I have removed the menu items in ISI tab for reaching individual HAM ISI Oplevs. As we now have the overview screens and also command line scripts to access individual oplev screens I have reduced the complexity of SITEMAP by removed these menu items.
All screens have been committed to the svn and can be customised at LLO
Screenshots attached:
Removed REFL_B from end station EtherCAT systems. This completes ECR E1500169 and issue tracker 1025 at LHO.
This time I caught it on the test stand with better error reporting. Test stand code error: Mar 24 10:14:29 conlog-test-master conlogd: Channel Access Exception: Channel Name: Unavailable, Native Type: Unavailable, Native Count: 0, Read Access: Unavailable, Write Access: Unavailable, IOC: Unavailable, Message: Virtual circuit disconnect, Context: h1asc0.cds.ligo-wa.caltech.edu:60755, Requested Type: TYPENOTCONN, Requested Count: 0, Source File: ../cac.cpp, Line Number: 1214 Mar 24 10:55:15 conlog-test-master conlogd: ../database.cpp: 661: insert_dbr_string_value_execute: Error executing m_p_insert_dbr_string_value_prep_stmt. Parameters: H1:SYS-MOTION_Y_SHUTTER_A_NAME, 1427219715488583185, 0, �, 0, 0. Mar 24 10:55:15 conlog-test-master conlogd: ../conlog.cpp: 314: process_cac_messages: MySQL Exception: Error: Incorrect string value: 'xC0' for column 'value' at row 1: Error code: 1366: SQLState: HY000: Exiting. Mar 24 10:55:16 conlog-test-master conlogd: ../conlog.cpp: 346: process_cas: Exception: boost: mutex lock failed in pthread_mutex_lock: Invalid argument Exiting. Seems to have non ascii characters in the value of the H1:SYS-MOTION_Y_SHUTTER_A_NAME channel. Production code error: Mar 24 10:14:29 h1conlog1-master conlog: Channel Access Exception: Channel Name: Unavailable, Native Type: Unavailable, Native Count: 0, Read Access: Unavailable, Write Access: Unavailable, IOC: Unavailable, Message: Virtual circuit disconnect, Context: h1asc0.cds.ligo-wa.caltech.edu:60755, Requested Type: TYPENOTCONN, Requested Count: 0, Source File: ../cac.cpp, Line Number: 1214 Mar 24 10:17:01 h1conlog1-master CRON[21576]: (root) CMD ( cd / && run-parts --report /etc/cron.hourly) Mar 24 10:39:01 h1conlog1-master CRON[21583]: (root) CMD ( [ -x /usr/lib/php5/maxlifetime ] && [ -x /usr/lib/php5/sessionclean ] && [ -d /var/lib/php5 ] && /usr/lib/php5/sessionclean /var/lib/php5 $(/usr/lib/php5/maxlifetime)) Mar 24 10:55:25 h1conlog1-master conlog: ../conlog.cpp: 301: process_cac_messages: MySQL Exception: Error: Incorrect string value: 'xEC>xBF' for column 'value' at row 1: Error code: 1366: SQLState: HY000: Exiting. Mar 24 10:55:26 h1conlog1-master conlog: ../conlog.cpp: 331: process_cas: Exception: boost: mutex lock failed in pthread_mutex_lock: Invalid argument Exiting. They may have failed on different channels, but this narrows it down.
Rescanned the channel list, no changes. Restarted Conlog. 28 unmonitored channels: H1:ALS-X_REFL_B_DC_GAINSETTING H1:ALS-X_REFL_B_DC_HIGH H1:ALS-X_REFL_B_DC_LIMITS H1:ALS-X_REFL_B_DC_LOW H1:ALS-X_REFL_B_DC_NOMINAL H1:ALS-X_REFL_B_DC_OFFSET H1:ALS-X_REFL_B_DC_RESPONSIVITY H1:ALS-X_REFL_B_DC_SPLITTERR H1:ALS-X_REFL_B_DC_TRANSIMPEDANCE H1:ALS-Y_REFL_B_DC_GAINSETTING H1:ALS-Y_REFL_B_DC_HIGH H1:ALS-Y_REFL_B_DC_LIMITS H1:ALS-Y_REFL_B_DC_LOW H1:ALS-Y_REFL_B_DC_NOMINAL H1:ALS-Y_REFL_B_DC_OFFSET H1:ALS-Y_REFL_B_DC_RESPONSIVITY H1:ALS-Y_REFL_B_DC_SPLITTERR H1:ALS-Y_REFL_B_DC_TRANSIMPEDANCE H1:GRD-TCS_ITMY_LOGLEVEL H1:GRD-TCS_ITMY_MODE H1:GRD-TCS_ITMY_NOMINAL_S H1:GRD-TCS_ITMY_REQUEST H1:GRD-TCS_ITMY_REQUEST_S H1:GRD-TCS_ITMY_STATE_S H1:GRD-TCS_ITMY_STATUS H1:GRD-TCS_ITMY_TARGET_S H1:SUS-SR2_LKIN_P_OSC_SW2R H1:SUS-SR2_LKIN_Y_OSC_SW2R
08:08 IFO fully locked @ ~ 18Mpc (according to SensMon).
Broke lock @ 08:41 :( Not bad for a maintenance day!
It seems as though after dropping lock the IFO came back twice (on its own or with some help from Ed?), with one locking attempt that failed in locking DRMI_1F.
The attached plots show the guardian state from 13:20 UTC to 16:20, the longest steps in the relocking process are still finding IR and locking DRMI, the long DRMI lock was about 8 minutes, the total time to recover from the lock loss was about 40 minutes (inlcuding the failed attempt), and the second time it took 15 minutes.
I think we can say that we now have some "one click" locks, unless Ed had to intervene in some way to lock this.
Evan, Sheila, Dan, Kiwamu, Lisa SUMMARY The wind dropped below 20 mph in the evening, and we started locking. After some roll, bounce and violin damping, we made it to DC readout @ 8W. We successfully transitioned DARM to ETMY with low noise ESD and got some MICH feed-forward cancellation going. We improved the noise between 30-200 Hz, and reached 28 Mpc (according to SensMon). The noise was more stationary than previously observed . However, the OMC is still shaking , and it is not good. Dan has a resonant gain for DARM ready to be tested to reduce the noise around 3 Hz, but we didn't manage to get that done tonight. We leave the interferometer locked at 28 Mpc starting at March 24, 2015 9:17 UTC . DARM control on ETMY with low noise ESD We did the transition from ETMX to ETMY L2 for DARM, but the interferometer was very glitchy, and we decided to abandon L2 control and try the new Den's strategy to do the feed-forward to ITM(Y) L2. Evan got about a factor of 10 subtraction (FM9 has BS plant inversion in LSC-MICHFF filter bank, gain +0.048). ASC cut-off filters Since none of the ASC loops had cut-offs, we decided to add some . They are now engaged by the Guardian. A more fine tuning will happen at some point. Started switching coil driver to low noise We started switching the coil driver to low noise . Evan will post a table later with the successful switches we did after acquiring the DRMI lock. We need to do more work on this, some failed.
Spectra attached.
Here is a plot of the lingering coherences with LSC DOFs.
Very nice progress!
Excellent news. Great progress.
This is fantastic progress, congratulations!
For anyone that would like to look at the data quality or run any DetChar tools on this data, the following GPS times can be used:
1111223896 - 1111239923
Duration: 16027 seconds
This entry should make Lisa a contender for her "best log entry" prize. Nice going, everyone!
These are the coil driver states that we've been using so far:
OLD | PRM | PR2 | SRM | SR2 | PR3 | SR3 | BS |
M1 | 1 | 1 | 1 | 1 | 1 | 1 | |
M2 | 2 | 2 | 2 | 1 | 1 | 1 | |
M3 | 2 | 2 | 2 | 1 | 1 | 2 | — |
OLD | IX | IY | EX | EY |
R0 | 1 | 1 | 1 | 1 |
M0 | 1 | 1 | 1 | 1 |
L1 | 3 | 3 | 2 | 2 |
L2 | 2 | 1 | 1 | 1 |
Now we've managed to switch to the following configuration for acquisition (following the tables in LLO#16565):
ACQ | PRM | PR2 | SRM | SR2 | PR3 | SR3 | BS |
M1 | 2 | 2 | 1 | 2 | 2 | 2 | 1 |
M2 | 2 | 3 | 2 | 3 | 3 | 3 | 2 |
M3 | 2 | 2 | 2 | 2 | 2 | 2 | — |
ACQ | IX | IY | EX | EY |
R0 | 1 | 2 | 1 | 1 |
M0 | 1 | 2 | 1 | 1 |
L1 | 3 | 4 | 2 | 2 |
L2 | 2 | 1 | 1 |
1 |
Green indicates a low-noise state, and red any other state. So far we have encountered the following things:
So what was supposed to be a simple favour for Aidan to turn the CO2y laser on turned out to be not so simple. Note: when I started this work I was unaware of the issues that Greg encountered last week (he has just put in an alog now for me alog 17302).
So at the moment the CO2y laser is still off.
Below is what observed/thought was the problem before hearing from Greg about the issues he saw previously (and is now alogging) , and also talking to Aidan about his recollections. Also will include what I think is going on for part of it and also recommendations.
Observation 1:
The rotation stage for CO2y is not working (I cant get it to do any rotation of any kind). I played with this before knowing that Greg had seen this last week also. I did not try playing with the CO2x. Based on past experience at LLO and LHO its going to be something with Beckhoff having to be reset (we really need to work out why these things just stop working). So until it is up and running I dont want to turn the laser on. And because I dont know how LHO like to proceed with doing procedures like this I took this no further. So what I need is for someone (maybe at start of next weeks maintenance period) to reset the HWP for me (probably just needs to be unplugged then replugged).
Observation 2:
I went out to the mechanical room to check the status of the power supplies etc. I immediately saw two things that peaked my interest.
1. Only one of the Kepco power supplies is on and wired up?? and the one that was on was drawing no current (see pic Kepco power supply). This I have seen before and indicates that one of the AOM drivers has faulted. But because things look to be hooked up weird (will discuss more later) looking on the tables I see that CO2y AOM driver is fine (indicated by the two green LEDs..see pic CO2yAOM driver), but the CO2x AOM driver has faulted (top LED red...see pic CO2xAOM driver). If you want to know what these fault lights mean please go to E1500133 and look at table 1 (someone should probably print this table out and stick it on or near the AOM driver for others for future reference). Now this has been seen before at LLO and the solution is to power cycle the power supply. However doing this will make the AOM start to work, and seeing as its not known when it faulted and what static diffraction setting the driver has been set to, once the reset occurs I could alter how much power is going to the interferometer (as CO2x laser looks like its being used). So again this is something should do during maintenance period (should note what power going to optic now and so when reset see how/if it changes and alter power accordingly).
2. The CO2Y laser power supply seems to be drawing a funky amount of current. 2.5Amps actually (see pic TCSy power supply), even though the laser is off. Ive even turned the key on the controller off, so it shouldn't even be in standby mode where only draws 0.7Amps. Talking to Aidan he seems to think that maybe the AOM driver/AOM is drawing current from this supply (him telling me that made me recall we saw numbers like this back in the day when we were installing and the laser was off but we had the AOM driver on). He seemed to recall the AOM driver and CO2 laser are driven off the same supply still here at LHO (which was the original design), even though the plan was to change to having the CO2 laser and AOM driver driven off separate supplies (as has been implemented at LLO...the Kepco power supplies power the AOM drivers the other supplies power the CO2 laser only). But this doesnt seem to totally match with what Im seeing in the mechanical room.
So looking at the wiring, here is what I think is the current setup at LHO. We are in some funky mid change between having one power supply do both laser and AOM driver and having separate supplies for all. In more detail......The two shelves showing the power supplies for TCS are shown in TCSracklowershelf and TCSrackuppershelf. How it should be as per the latest plan that I know of (and is at LLO) is that the on the lower shelf should hold the power supplies that power one table (ie say Y ) and for the upper shelf the power supplies for the other table (X) with seperate supplies for the lasers and the AOM drivers. What seems to be the case is that:
*on the lower shelf we have the Instek supply (I think thats the brand name....its the the one with the digital readout reading its drawing 2.5Amps that says TCSY on top of it) powering BOTH the CO2y laser and the CO2y AOM driver (I think this because the laser is OFF, as is the controller, but the AOM driver on the Y table has not faulted and is "running".The x table AOM driver has faulted and is not "running" and 2.5amps is an approximate number that the AOM driver could be drawing).
* The Kepco power supply on the lower shelf is wired up but not drawing any current at the moemnt and thus and looks to be powering the AOM driver for the X table (why... because the x table AOM driver has faulted its not drawing any current). It should be wired to be controlling the Y table if supply in the position at..or at the very least labeled clearly as to what it is powering. Mind you I could be wrong and its powering something else
* The instek supply (the one showing the digital readout of 22.42 Amps) on the upper shelf is powering ONLY the CO2 x laser (its drawing this current as the laser is on).
*The Kepco power supply on this upper shelf is not even wired up and is thus off.
What I suggest is at some stage when have some time (maybe during maintenance period next week), is to wire the system as per the new design (and as how is at LLO). One Kepco power supply is used for the AOM driver on one table and the other Kepco power supply be wired to the AOM driver on the other table. And then just have the other power supplies individually controlling the CO2 lasers. Should then also have the power supplies for one table on one shelf and the power supplies for the other table on the other shelf.
Also when turn laser back on need to see if still only drawing half the current like Greg saw and if so investigate (probably cable connection is the problem like in the past). SO quite a few things would lile to do at next weeks maintenance period.