andres.ramirez@LIGO.ORG - posted 08:24, Monday 13 October 2014 (14418)
ISS Difracted Power High this morning
The ISS Diffracted Power was over 12% this morning. Is been adjusted to 9.5 %
The ISS Diffracted Power was over 12% this morning. Is been adjusted to 9.5 %
model restarts logged for Fri 10/Oct/2014
2014_10_10 00:22 h1fw1
2014_10_10 07:43 h1fw1
model restarts logged for Sat 11/Oct/2014
2014_10_11 05:33 h1fw1
2014_10_11 21:41 h1fw0
all restarts unexpected.
I'm leaving the interferometer with LSC feedback disabled, the the green lasers trying to lock to the arms. We are currently having both high winds and high microseism, so anyone who is interested in looking at the motion of our optics or performance of ISIs when the ground motion is high could use this time, starting at 23:39 UTC October 11.
I've left the intent bit as undisturbed, could the first person to come in and commision or work in the LVEA toggle it to commisioning?
Alexa, Kiwamu,
We tried the DRMI tonight with the arm cavities held at a off resonance by ALSs. No exciting result yet.
(Too high speed in DRMI)
We could lock the DRMI only twice and they did not last for more than a couple of minutes. According to the band-limited RMS plots in the control room, we are having a monontonically increasing micro-seismic tonight, starting 10 hours ago. (Presumably) due to the high seismic, the DRMI fringe tonight is much faster than usual.
(ESD-UIM crossover issue)
Also we experienced an issue in which the ESD-UIM had a bad cross-over. When we closed ALS DIFF, it stayed locked fine for the first 10 minutes or so. But we noticed that the DARM feedabck was slowly developing a 1-ish Hz oscillation which gradually started using up the ESD range. This eventaully caused saturation in either or both of the ESDs and we lost the lock. In order to avoid this issuem. we tried decreasing the ESD LOCK_L gain from 1 to 0.8 on both ETMX and ETMY. This reduced the amplitude of the oscillation and also the in-loop spectrum got back to how it had been. We are concluding that the cross-over between UIM and ESD changed from yesterday.
After adjusting the ESD gain, ALS DIFF reliably stayed locked for a long time, typically order of 20 minutes. Each lock ends usually due to misalignment in ETMX or ETMY.
(ETMX ESD check)
This is just an observation of ETMX ESD.
In the process of checking the ESDs and etc, we checked the reponse of the ESD in the pitch direction by shaking it at 4.1 Hz and looking at it with oplev. We noticed that shaking it in pitch mostly gave us yaw. Shaking yaw gave a good yaw motion though. We checked ETMY and it was not the case.
(Sheila, Alexa)
Today we worked on the slow/tidal feedback to the ETMs. In doing so, we saw that the L2P for the M0 state was not very good at low frequency. Using the same method as was done for the BS (alog 14022), we found that the gain of ETMX L2P (with FM1, FM2 ON), should be changed from -1.0 to -4.0 at 0.01Hz. So far it appears that this is OK for high frequency. Meanwhile, we found that the gain of ETMY L2P (with FM1,FM2 ON) should be changed from -1 to 22 at 0.01Hz. We are not sure if this is okay at higher frequencies.
The slow/tidal feedback for both arms now works, and no longer causes a drift in the alignment. We made the following the changes to the feedback servo (this is all taken care of in the guardian already):
It appears that the DIFF PLL VCO is now in range, but maybe that's a coincidence with ground motion today....
One can probably bleed off the dc term (f<10mHz) by feeding back to HEPI. The path is all set up and the LSC signals are received by a HEPI ISC input filter. All that should be required is an integrator in this path.
We were having some problems with ALS DIFF earlier tonight, so we reverted the L2P gains back to -1. It is unclear at the moment if this was related at all.
~Two more 5.2mag aftershocks this morning on the Southern East Pacific Rise near Easter Island~
08:00 Morning checklist: Reset ALH; vacuum system looks ok - HAM6 cold cathode pressure reporting 10-7; CDS/DAQ - everything looks normal.
08:30 Morning meeting
08:42 Krishna to X-End to do some BRS testing/tweaking
09:20 Krishna back from X-End
09:50 Keita to End Y to align green WFS. He will be using ETMY/ITMY/BS/PR3
10:14 Dale into the control room with a tour group
12:04 Jason to End X to align OpLev
12:30 Adjusted ISS Diffracted power to ~9.5% by reducing the refsignal voltage from -2.01 to -2.05
12:37 Jason back from End X
14:45 Sudarshan into LVEA, crawling under HAM2
12:48 Keita back into control room from End Y
12:50 Sudarshan back from crawling
J. Kissel, K. Venkatewara
Jeff K. designed a model to use the GND_Supersensor (tilt-corrected STS) to do sensor correction on Stage 1.
To get the supersensor, we take the GND_STS and add the Tilt signal from BRS converted to velocity units by multiplying g/w. To further correct for the small amount of acceleration present in BRS, we further add a factor of F * V_STS * g/w^2 where F = M * d / I (M=Mass of beam-balance, d = Center of mass offset, I = Moment of Inertia)
Jim and I will try to implement and test this next week when not interefering with commissioning activity.
J. Kissel A reminder to add the damping control signal monitor (currently fed into ADC0 CH 30, see D1400077, as of LHO aLOG 14351) to the h1isietmx model's channel list. In fact, it's already included and sent into a filter bank called "ETMX_SPARE" in the GND_BRS block, so let's just change it from a filter bank to an epics output (like the drift monitor) and call it "ETMX_DAMPCTRLMON." Once added to the front end, we should add it to the MEDM screen.
J. Kissel, K. Venkateswara
We realized we had the wrong factor of 1/2 as described in alog entry 13448. 1/2 [V_sing / V_diff] was not needed/
The filters have been changed to account for the new calibration factor which is:
3.5e-6 [rad/pixel] * 1 [pixels/V_sing] * 1/2 [V_sing / V_diff] * 40/2^16 [V_diff/ct] * 1e9 [nrad/rad] = 2.1362 [nrad/ct]
J. Kissel, K. Venkateswara, Krishna copied, pasted, and posted too fast and still included my erroneous single-ended to differential conversion in the piece-by-piece equation, even though the answer quoted is correct. The calculation to achieve the correct calibration is 3.5e-6 [rad/pixel] * 1 [pixels/V_sing] * 40/2^16 [V_diff/ct] * 1e9 [nrad/rad] = 2.1362 [nrad/ct] Also, I've committed the updated filter file to the userapps repo. Also, also, we need to (a) Add the reference mirror filter bank and damping motor control signal* to the MEDM screen (b) Capture a new safe.snap with the reference mirror calibration filter turned ON as it is now. *The damping motor control signal isn't yet present in the user model, it's only available as a raw IOP test point. We should add this to the h1isietmx user model when we're making the front-end code modifications to create the tilt-free ground super sensor. I'll post this as a comment to the other suggested model changes so we don't forget.
At ~12:34PT I adjusted the diffracted power in the ISS to ~9.5% as instructed by Rick Savage. The refsignal is currently set to -2.05V.
J. Kissel I've arrived at numbers for refining the calibration of the ETM OPTICALIGN alignment offset sliders, using the same method as was used for the ITMs and BS (see LHO aLOGs 14265 and 14321, respectively). The new slider gains should be EX P = 15.344 [ct/urad] Y = 44.044 [ct/urad] EY P = 19.341 [ct/urad] Y = 51.013 [ct/urad] which are comparable to the ITMs' calibration, as expected. I have not installed these new gains, nor have I corrected the saved alignment offsets since it will be too disruptive to commissioning. These can be installed at anyone's convenience, but I'll install them in week if know one's gotten around to it. The to-do list: - Record current OPTICALIGN OUT values (in [ct]), to confirm you get the same answer after updates - Install above calibration gains - Restore to ALIGNED via guardian, enter in predicted offset values (in [urad]) below (adjust, if necessary, to get the previously recorded output value in [ct]) - Save ALIGNED value using pull-down menu in IFO_ALIGN MEDM screen - Restore to MISALIGNED via guardian, enter in predicted offset values (in [urad]) below (adjust, if necessary, to get the previously recorded output value in [ct]) - Save MISALIGNED value using pull-down menu in IFO_ALIGN MEDM screen - svn commit the updated aligned and misaligned .snap files in the ${userapps}/release/sus/h1/burtfiles/ directory - IF AMBITIOUS (and commissioning team gives you some time) -- perform refinement of optical lever calibration with these new values, as was done with the ITMs and BS (see LHO aLOGs 14312 and 14387). Details ----------- The current alignment offset slider calibrations for both EX and EY are 23.219 [ct/"urad"] 51.689 [ct/"urad"] and currect stored alignment offsets are P Y ["urad"] EX Aligned 274.1 63.9 Misaligned 240.0 102.0 EY Aligned 83.4 -64.8 Misaligned 62.0 -7.72 The ETM alignment slider values to hit the center of the ITM baffle PDs are P Y ["urad"] EX PD1 261.4 79.75 PD4 285.2 49.1 delta 23.8 30.65 claimed displacement = 2*delta 47.6 61.3 EY PD1 70.5 -83.5 PD4 100.5 -48.0 delta 30.0 35.5 claimed displacement = 2*delta 60.0 71.0 where the factor of two to calculate the displacement is from the single-bounce, optical lever effect. We know from D1200313 and D1200296 the ITMX and ITMY (respectively) PDs 1 and 4 are 11.329 [inches] = 0.287757 [m] apart vertically, and 11.313 [inches] = 0.28735 [m] horizontally on both ITM baffles, implying an angular displacement -- over the L = 3994.5 [m] arm lengths -- of P 0.287757 [m] / 3994.5 [m] = 72.03 [urad] Y 0.28735 [m] / 3994.5 [m] = 71.94 [urad] which means the correction factor for the OPTICALIGN gain is EX P 47.6 ["urad"] / 72.03 [urad] = 0.661 ["urad"/urad] EX Y 61.3 ["urad"] / 71.94 [urad] = 0.852 EY P 60.0 ["urad"] / 72.03 [urad] = 0.833 EY Y 71.0 ["urad"] / 71.94 [urad] = 0.987 or EX P = 1.5132 [urad/"urad"] EX Y = 1.1736 EY P = 1.2005 EY Y = 1.0132 which are corrections comparable to the ITMs (good sanity check, since they have the same actuation chain, and used the same offset calibrations prior to the refinement). So the new alignment slider calibation gains should be Old Cal Correction New Cal EX P 23.219 [ct/"urad"] * 0.661 ["urad"/urad] = 15.344 [ct/urad] Y 51.689 [ct/"urad"] * 0.852 ["urad"/urad] = 44.044 [ct/urad] EY P 23.219 [ct/"urad"] * 0.833 ["urad"/urad] = 19.341 [ct/urad] Y 51.689 [ct/"urad"] * 0.987 ["urad"/urad] = 51.013 [ct/urad] and the predicted alignment offsets that need be installed are old offsets ["urad"] * correction factor [urad / "urad"] = new offsets [urad] P Y [urad] EX Aligned 414.77 74.993 Misaligned 363.17 119.71 EY Aligned 100.12 -65.655 Misaligned -77.792 -7.822
Keita, Sheila, Kiwamu,
Keita pointed out yesterday that there was a strange lock loss event of the DRMI in which the guardian seemed to have switched the state to a unlocked state even though the PRC and SRC build-up were still sufficiently high. This is now understood and fixed.
We found out that this was due to our bad coding in the guardian. The "WFS_CENTERING" state had a unwated command (I don't exactly know what part was it) which brings the state to "DOWN" immediately. So every time when we requested the "WFS_CENTERING", it then shut down the lock. Sheila edited the code and now it does not show a funny lock loss any more.
The attached is time series from a second example of the funny lock loss. You can see that PRCL cut its input at t = 1.2 sec even though POP_RF18 and AS_RF90 were high. This was actually initiated by our request for the guardian to go to "WFS_CENTERING". Then it immediately transitioned to a wrong state DOWN, which shuts all the LSC control loops. Of course, this resulted in lock loss.
K. Venkateswara
I did another test to make sure the BRS damping works as expected. The attached pdf shows the results of the test. I disturbed the BRS by walking in its vicinity at t=200 seconds and the amplitude gets rung up. The turn-table starts damping it and it switches off after ~1000s.
DriftMonitor:
The channel H1:ISI-GND_BRS_DRIFTMON monitors the position of the beam-balance mirror pattern on the CCD camera. It was scaled to go roughly from +/-16k counts. But it looks like on the positive side it happens a bit earlier than that, roughly 14.5k counts. The second attachment (DriftMonOlder.png) shows the DriftMon output from August-September. It looks like it had already hit the edge on Sept. 4th. The code still works if the pattern goes over the edge but it can appear noisy if there is a partial pattern, so ideally we want to keep it in the working range.
The current DriftMon position is shown in the third file (DriftMonCurrent.png) after I completed the damper installation. It seems to be trending logarithmically towards -10k which is a bit lower than ideal but still in range. Adjustment of the position are not trivial and there are large drifts after each adjustment which are not always easy to predict, so adjustments should be avoided unless necessary. I will monitor it over time and if needed make further adjustments.
The setup script for the userapps directory has been modified to add paths for the new calibration subsystem, cal. The environment variables USERAPPS_LIB_PATH, USERAPPS_MEDM_PATH, and USERAPPS_SCRIPTS_PATH have been modified, and environment variables CAL_SRC and CAL_IFO_SRC have been added. This change is effective for new logins or new shells.
model restarts logged for Thu 09/Oct/2014
2014_10_09 08:51 h1fw1
2014_10_09 10:23 h1fw1
2014_10_09 13:12 h1fw1
2014_10_09 21:41 h1fw0
all unexpected restarts
re my 14373 entry, I figured I could correct these to the T1000388 even though there is a "Check the signs" Note in the doc as Fabrice says we'll just change the signs in the sym filters.
The matrix had correct elements in the primary transformations; only some of the cross coupling elements were variant. I went ahead and ran the populate matrices medm script. It crashed as it tried to load CPSALIGN values which are not in the data file. Makes me wonder how the GS13 & CART2ACT values got loaded before... So, I moved the CPSALIGN load section to the end and ran the script again. I then confirmed all the matrices complied with the T1000388--all good. Then confirmed the ISI still isolated--yes.
Then with the ISI just damped and HEPI not isolating( no Pos loops,) I drove HEPI in Z with a random noise signal band passed between .1 and 1 hz. Monitoring the HEPI L4Cs and the ISI Stage0 L4Cs, see attached, it appears that the vertical sensors are out of phase. So it seems we need that sign change in the SYM filters.
Made and committed safe.snap file.
Commissioners wanted the machine back before I could do other dof measurements, maybe I can do them with passive TFs. Either way, later.
Here is some horizontal data. I drove HEPI RZ but since the ISI L4Cs go to the Sensor Correction bank, it only has X Y & Z. So, here I look at the L4CINFs, the tangentially oriented sensors should be seeing this RZ. Compared to the vertical data above, I drove 4x as hard to get the coherence higher(still kinda poor.) It may not say the same thing, but, based on the phase, the signs are the same. Of course now this is before the input filters (I should have grabbed the _OUTs but too late for this run) much less the L4C2CART matrix so I'm really mixing apples and applesauce.
Jason, Sheila
There have been some glitches on the ETMX oplev, which don't seem to correspond to motion of the optic, the attached screen shot shows an example. I think that I have seen glitches like this that were much larger in amplitude, but am not sure.
One question for detchar is if they have a tool that can search for glitches like these, and give us some information about how often they are happening, how large they are, and maybe monitor to see if they are happening on other op levs.
Another issue is that sometimes the lasers fail, which is often foretold by a several seconds oscillation in the oplev sum. Can detchar set up some kind of monitoring for that?
Of course the real question is how can we fix it......
Hi Jason and Sheila, here at LLO Olmo Cerri, a summer student from UMISS, looked at the possible causes of glitches in OPLEVs. He worked with Suresh. Suresh earlier found that the the temperature variation could cause changes in the cavities of OPLEV laser and thereby changing the laser intensity which would look like glitches. Olmo looked at week long strecthes of data from a few OPELVs and charactersized how many glicthes we see (he wrote a matlab script to find the such gltiches). They also implemetned a simple temperature stabilization method (LLO a-log) to reduce the gltiches. It seemed to work well. Here is a presentation by Olmo on this work which shows data before and after the temperature stabilization setup. He is now looking into implemneting an online version of the code.
I set Olmo's code to run on a cron job with the results here: https://ldas-jobs.ligo-wa.caltech.edu/~codyca/MHoF_results/ We still need to put together a better way to present the results. The code is set to remove glitches caused by human interaction by looking at the OPTICALIGN offsets. Also, the results for LLO can be found here: https://ldas-jobs.ligo-la.caltech.edu/~codyca/MHoF_results/
Tuesday, October the 14th, was also a bad wind day, starting about noon. See Sheila's log 14475.