PcalX bump from 0-120 Hz

TJ, LouisD, TonyS, DriptaB, FranciscoL

There was an unexpected "bump" on PcalX at around GPS 1395546818. This was shortly after TJ attempted Detchar safety injections (76727).

Looked into H1:CAL-PCALX_TX_PD_OUT, RX_PD_OUT, OFS_PD_OUT, OFS_AOM_DRIVE_MON_OUT and OFS_ERR_OUT for any issues.

Observations:

• Bump travels across spectrum from 0 Hz to 120 Hz for an indefinite amount of time. The figures "PcalX_DTT_120.png", "...85.png", and "...40.png" illustrate the bump appearing at 120 Hz, 85 Hz, and 40-ish Hz, respectively.
• OFS_ERR_OUT sees the bump at what seemed to be a loop troubleshooting. The top plot on "PcalX_DTT_zoom_ERR.png" shows the bump appearing at ERR_OUT.
• Troubleshooting did not solve the issue. We did not see any bumps this morning so the issue seems to be gone.

Some ideas on where to look for the source:

• AOM issues
• "RF whistling"

There was a similar issue on (70817) where the "OFS servo was saturating" after Detchar injections. The OFS was not happy after troubleshooting this time.

Further investigation is on process.

Opslogin0 web client now working again

Connection to nomachine via web is once again working.

A+ HRTS assembly, balancing and characterization report (for O5)

Ryan C, Austin, Oli, Jeff K, Betsy, Dave, Eric, Fil, RAL/CIT team and Rahul

Happy to report that we have finished the assembly of our first A+ HAM Relay Triple Suspension (HRTS) (freestanding version) with Class A parts as per the assembly procedure described in D1900449_V7.  HRTS is a new small scale triple pendulum suspension required for Balanced Homodyne Detection system. This is first of the twelve suspensions we have to deliver for both the sites (six each for LHO and LLO, which includes one spare per site) for observing run O5.

HRTS comes in two configuration, freestanding (table mounted which is discussed in this alog) and suspended (to be built, will be attached to the new beamsplitter suspension BBSS). The details about the assembly and characterization along with pictures are discussed below,

Attachment01 shows the AR side of the SUS, attachment02 shows the HR side. For isometric view, please see attachment03.

This multistage suspension (three suspended stages) has blade springs at the upper stage and top mass for vertical isolation. There are two blade springs at the very top stage (D2100389) as shown in the picture here. The blade spring are mounted on spring loaded clamps which can be adjusted for Yaw dof and height (sus point). The two wires (diameter 0.006in) from the two top stage blade springs suspends the Top Mass (D2100362) which has four blade springs, as shown here.

Wire diameters are as follows:-

Top Stage to Top Mass = 0.006in, length 115mm

Top Mass to Penultimate Mass = 0.004in, length 115mm

Penultimate Mass to Optic = 0.0025in, length 160mm

Suspended masses (as per specifications):-

Top Mass = 750gm

Penultimate mass = 802gm

Dummy optic = 300gm

The Top Mass blade spring clamps are spring loaded (just like the top stage) which gives them the ability to adjust for blade tip height. We use a tool (T2400063) provided by the RAL UK team for measuring (while adjusting) the blade spring tip height (this tool can only be used for this stage) - as shown in this picture. A calibration block has also been provided for calibrating the tool before using it. The penultimate mass is suspended from the top mass blade springs using four wires of diameter 0.004in. The Optic (dummy optic for now) is suspended using two wires in a loop from the penultimate mass - see picture here. All the wires were pulled using their dedicated wire jigs, which defines the wire length and position within the wire clamps.

We had some issues with wire installation procedure (wires at the top mass stage kinks/breaking too often during handling). After discussions with colleagues here and at RAL we now have a new tool (currently under fabrication) which will aide in wire installation procedure. Also, with practice we are getting better at handling wires of this thickness (thinner than human hair).

Once the assembly was complete, we measured and adjusted the height of the top stage blade springs, which comes out to be 22.5mm from the top plate. This adjustment was made to lower the entire chain such that each stage aligns with their respective position, as marked on the frame. The bottom edge of the dummy optic is now approximately suspended at 40.5mm from bottom of the frame (s/n 002) - which as per the design specifications. The other degrees of freedom like pitch, yaw, roll also looks respectable without any major adjustment, although we can further improve the pitch on penultimate mass (using balance mass and pitch adjuster mechanism provided at two stages).

HRTS is controlled using six BOSEMs, their flag/magnet attachment are as per the controls arrangement document E2300341. After attaching the BOSEMs and with little adjustment all six flags looked nicely centered.

Before centering them we measured their Open Light Current (OLC) and calculated the offsets and gains which are as follows,

BOSEM D060106-E (s/n)	OLC	Offsets (-OLC/2)	Gain (30,000/OLC)
F1 (S1900810)	26807	-13403.5	1.119
F2 (S1900795)	31010	-15505	0.967
F3(S1900754)	30165	-15082.5	0.994
LF(S1900809)	29316	-14658	1.023
RT(S1900782)	28820	-14410	1.040
SD(S1900626)	26768	-13384	1.120

We have a dedicated test stand at the triples lab in the Stagings building, thanks to Fil, Eric and Dave. The hardware (power supply, satellite boxes, Triple Coil driver, IO, AA and AI chassis) can be seen in this picture. We also have a working MEDM screen for HRTS on X1 controls thanks to Jeff Kissel and Oli. Using this infrastructure we started testing out our suspension. The first challenge was too much vibration in the lab due to turbulent air, vibration due to building doors opening/closing etc. Our suspension is sitting on a heavy optical bench and we have also used a teflon sheet for absorbing some of the ground vibrations. However, since the turbulent air in the lab was too much for us to take any meaningful measurements, hence we covered the SUS in two layers of foil and then clean room cloth - as shown over here. We then exited the lab and it took 30mins or so for things to calm down (while HVACs in the lab are still running). Long story short, we took the first top to top transfer function measurements for HRTS and the plots are attached below. I am attaching the DTT plots as well as the ones processed in Matlab.

If you look at the DTT plots, the coherence is not too bad despite all the external vibrations leading to saturations (we have 18bit DAC). In Matlab we can compare our results against the model (Mathematica from M Barton, imported to Matlab). The Longitudinal (L),Transverse (T) and Yaw (Y) dof aligns nicely with the model, ie all peaks and magnitude looks good. Pitch (P) also has most of the peaks at right places (except a missing peak at 4.5Hz), and is off in magnitude which we are investigating. Vertical (V) is noisy (which is expected as the suspension can be easily excited in vertical motion) and has a cross-coupling from Roll (R) (at around 1.8Hz)? Roll (page8) looks the worst of all especially the shape and magnitude at frequency below 2Hz.

I am still fine tweaking the balancing of the suspension, further isolating it environmental noise and discussing with colleagues to take better measurements. In the meanwhile this is a decent start for us, eleven more to go.

Tagging EPO -- Rahul has new babies! This is the newest type of suspension -- and by far housing the smallest triple suspension. So cute!

Changing PRM Position to Improve Build-ups

Jennie W, Jenne

 

Yesterday, on advice from Sheila to we are trying to decouple the alignment changes to IM1, IM3, IM4, PRM and PR2 in this entry from any changes we may be making to the jitter coupling.

So the plan is to move just PRM to the place Gabriele and I had moved it to. This was about 284 counts on SUS-PRM_M1_DAMP_INMON on 22:44:25 UTC on 2024-03-21 when we had being making these changes.

Firstly I did this on the alignment sliders and then realised this is compensated for by camera servo YAW1 which changes the PRM to keep the spot on the BS constant so we need to actually change the camera servo offset to move the PRM.

 

Jenne has turned on ADS lines but we will remain on camera servos for this test and tweak the ASC-CAM_YAW1_OFFSET which will move PRM.

 

We checked the PR2 A2L gain first by injecting lines on PR2 in OSC8 in both and yaw, as Annamaria suggested we would need to check the spot wasn't moving too much on PR2. See first image.

Previous example of tuning A2L gains for PR2 (and I think the last time it was done) are here.

We took a spectrum and changed the A2L gains in SUS-PR2_M3_DRIVEALIGN_Y2L_GAIN to minimise our injected lines in LSC-PRCL_IN1 as these lines did not change much in DARM as we changed the gain.

These were mimimised at P2L gain of -0.31 and Y2L at -7.3 (nominal gain are -0.61 for P2L and -7.4 for Y2L so only the pitch gain changed substantially). Template for this is saved in /ligo/home/jennifer.wright/Documents/Noise_DARM/20240326_PR2_dither.xml.

 

Then I checked the jitter coupling by doing injections using the noise budget templates in:

/ligo/gitcommon/NoiseBudget/aligoNB/aligoNB/H1/couplings/IMC_PZT_{P,Y}_inj.xml after taking a background measurement in each case with the injection off.

We inject into either IMC-PZT-PIT_EXC for pitch and IMC-PZT-YAW_EXC for yaw measurements.

 

The P measurements are saved in /ligo/home/jennifer.wright/Documents/Noise_DARM/20240326_IMC_PZT_P_inj_start.xml

Took Y measurements (no injection and injection) without squeezing as somwething happened to squeezer and it is no longer injected. Saved in /ligo/home/jennifer.wright/Documents/Noise_DARM/20240326_IMC_PZT_Y_inj_start.xml

Moving PRM down in yaw (via camera offset) made build-ups worse but more power on POP B and B. First cursor on plot shows this point.

 

Measurement Times:

PITCH DOF, no injection: 2024/03/26 23:08:40 UTC

PITCH DOF, injection: 2024/03/26 23:12:22 UTC

 

YAW no injection: 2024/03/26 23:22:18 UTC

YAW injection: 2024/03/26 23:33:45 UTC

 

Plots
Top left - dark red is DARM with injection on, yellow is background with inj off.

Top right - blue is cleaned DARM with injection off, red is with injection on.

Bottom left - green and brown are two jitter witness channels with no injection, red and blue are these same two channels with the same injection on.

Bottom right - green and brown are two jitter witness channels with no injection, red and blue are these same two channels with the same injection on.

 

Made jitter better slightly on IMC-WFS_B_I_YAW.

Could not easily improve A2L coupling on PR2 here. Tried running the dither template for PR2 again and changhing the gains but nominal gains was close to optimal where P2L has -0.61 and Y2L has -7.4.

But took jitter references again just for yaw degree of freedom. For plots:

 

Measurement Times:

YAW DOF No injection: 26/03/2024 23:59:10 UTC

YAW DOF Injection: 27/03/2024 00:02:15 UTC

 

Plots
Top left - dark red is DARM with injection on, yellow is background with inj off.

Top right - blue is cleaned DARM with injection off, red is with injection on.

Bottom left - green and brown are two jitter witness channels with no injection, red and blue are these same two channels with the same injection on.

Bottom right - green and brown are two jitter witness channels with no injection, red and blue are these same two channels with the same injection on.

 

After I had reverted the changes on PRM, Jenne did some IM1 and IM3 yaw steps detailed here (of about 20 counts on sliders instead of the 200 counts Gabriele and I were doing before).

From this image where first cursor is PRM change and second cursor is IM1 and 3 change; we can see that when IM1 and 3 were being changed, the PRM was pulled to the same place as I had moved it to but this time it made the build-ups better.

Jenne and I conclude that this alignment change is thus not just the PRC being re-aligned but some input clipping or pointing.

Also in Gabriele and I's moves we only moved YAW after aligning pitch and so this may also make a difference.

These YAW moves in IM1 and 3 also unlocked the IFO so we need to look closer into why this happens (Jenne suggests some fast loop is running away), but I think this YAW alignment is something we should move towards maybe with other changes to make sure all the loops stay stable.

---

Just to clarify we reverted all the A2L gain changes and alignment changes after we lost lock.

EX bias change test

Artem, Sheila

We changed the bias on ETMX while adjusting the drivealign gain to compensate, similar to what was done in 73913.  We might be seeing a small difference in the DARM spectrum between 30-40 Hz. 

We scaled the drivealign gain with the bias, but also adjusted the gain at each step to keep the DARM OLG the same. 

voltage	references	quiet time start (UTC March 27th)	quite time end	DA gain	DA gain scaling (in addition to bias voltage scaling)	OLG change %
140	0-6			184.65	NA	NA
190V	8-13	15:41:40	15:58:51	126.895	0.93	0.5
240V	14-20	16:02:54	16:19:05	96.167	0.89	1
289.6V	21 -27	16:27 *	16:43:30	77.22	0.865	1.2
415	28-34	16:49	16:59:30	51.628	0.829	1.8

People walking in LVEA and plugging in CM board seemed to cause some glitches. no evidence of people walking on seismometers from 16:37 on.

The first attachment shows the DARM OLG after gain adjustments.  The second attachment shows long spectra at the various biases we checked.  We also looked at coherence with the susrack magnetometer (Y)  and ESD power monitor 18 V, and saw no coherences there.

The third attachment shows a comparison of the spectrum with 190V bias and 415V, making it easier to see that there might be a difference in noise.  A next step would be to do some repeated steps between these bias settings to see if the difference is repeatable, and doing a broadband PCAL to DARM injection to check that the calibration is as consistent as we think it is between these settings.

Attached whitened DARM and my ESD noise model spectrograms for different bias levels. Not sure it shows something, in particular for 240V bias there were people walking so there are some glitches (I thought last 6 minutes were quiet - from seismic - but apparently they were not..). 415V bias spectrogram also has some glitches..

Wed CP1 Fill

Wed Mar 27 10:11:47 2024 INFO: Fill completed in 11min 43secs

Gerardo confirmed a good fill curbside.

SQZ_ANGLE_ADJUST servo turned off March 25th

Naoki turned off the SQZ_ANGLE_ADJUST servo on March 25th, new nominal is DOWN. We had turned it on in 76556.

It appears that the bucket squeezing (yellow BRLMS that effects range most) is more stable without the servo, with the cost that but 700  to 1750Hz squeezing is less stable (blue and green), plot attached. In the two locks last night all BLRMs seem to take longer to thermalize, probably caused by the IFO being unlocked for longer and cooling down between locks.

Ops Day Shift Start

TITLE: 03/27 Day Shift: 15:00-23:00 UTC (08:00-16:00 PST), all times posted in UTC
STATE of H1: Observing at 149Mpc
OUTGOING OPERATOR: None
CURRENT ENVIRONMENT:
    SEI_ENV state: CALM
    Wind: 12mph Gusts, 9mph 5min avg
    Primary useism: 0.03 μm/s
    Secondary useism: 0.29 μm/s
QUICK SUMMARY: Locked and Observing for 4 hours. Seems the relock was automatic.

Lock losses while moving input beam

All three times we tried to move the input beam (76534, 76607 and yesterday), we caused a lock loss when moving in the yaw direction.

Approximate lock loss times: 1394946678 1395096385 1395535842

All lock losses appear to show the same behavior:

• IMC transmission drops in a "small" step before anything else happens. IMC transmission then stays constant for 120 ms, then drops quickly
• This corresponds to a "glitch" in REFL_SERVO_ERR, the start of a "ramp" in REFL_SERVO_OUT, and a discrete step in LSC error signals (PRCL, MICH, SRCL)
• Shortly after that, all signals go crazy, but there is no evident oscillation

Those lock losses are indeed very fast, and this seems to point to a CARM problem.

Ops Eve Shift End

TITLE: 03/27 Eve Shift: 23:00-07:00 UTC (16:00-00:00 PST), all times posted in UTC
STATE of H1: Observing
INCOMING OPERATOR: None
SHIFT SUMMARY: Just started observing. Had a lockloss a couple of hours into my shift and then a couple earthquakes that kept us from locking for a while. After that, I was having issues with DRMI and PRMI and couldn't get them to catch by adjusting by hand, so I reverted the OPTICALIGN offset changes that Jennie had made earlier with the beam walking. That actually ended up making the alignment worse, and I had a lot of issues trying to revert those optics to different times, and then reverting all optics (except sqz) to multiple different times, before being able to get past DRMI/PRMI after an initial alignment.

This means that the first offset script(76732) didn't get to run at 10pm either. :(

Attached is the one SDF I had to accept to get into Observing.
LOG:

2300UTC Detector locked for 2 hours, commissioning going on

** This was due to Nutsinee and Naoki''s diode centering at HAM8 **
2316 ZM3 started was constantly saturated
2319 Saturating stopped

0050 Lockloss
0121 Started MANUAL_INITAL_ALIGNMENT
0153 Earthquake mode activated due to incoming earthquake from Fiji
0156 INITIAL_ALIGNMENT complete, relocking
0206 Lockloss from ACQUIRE_DRMI_1F
0210 Lockloss from FIND_IR (due to earthquake), sitting in DOWN while earthquake passess

0319 Starting to relock
0323 Earthquake mode activated due to a nearby earthquake off the coast of Oregon
0325 Lockloss at FIND_IR due to nearby earthquake, holding in DOWN
0341 Starting to relock again
0357 Earthquake mode activated due to Alaska earthquake  
0416 Cycling  through CHECK_MICH_FRINGES, so I went to DOWN and ran MANUAL_INITAL_ALIGNMENT only for PRC, MICH, and SRC
0431 INITIAL_ALIGNMENT done, starting relocking
0451 Lockloss after cycling through PRMI/DRMI

0524 Reverted all optic offsets (except SQZ) to 03/26/24 21:16UTC (beginning of the lock before maintanence)

0544 Reverted all optic offsets (except SQZ) to 03/26/24 06:50UTC (beginning of last lock)
0527 Started INITIAL_ALIGNMENT
0621 Started relocking
0706 NOMINAL_LOW_NOISE
0711 Observing

Start Time	System	Name	Location	Lazer_Haz	Task	Time End
21:47	ISC	Jenny	CR	n	PRM moves	23:47
22:52	PEM	Robert	LVEA	n	Shaker setup and injections	01:20
22:59	SQZ	Nutsinee, Naoki	FCES	yes	Center diode	23:39
23:21	SHG	Julian	OpticsLab	y(local)	Working on SHG	00:59

Ops Eve Midshift-ish Status

We lost lock 1.5 hours ago and have not been able to get back up due to an earthquake from Fiji rolling through. Currently sitting in DOWN until the ground motion slows

camera servo offset stepper tests set to run overnight

I set the camera_servo_offset_stepper.py script to run for CAM 1 servo between 10pm and 2am PDT and the CAM2  servo between 2am and 6am as the IFO unlocked at 6pm today when the first script was meant to run. Ths can run without knocking us out of OBS and has been cleared with Jenne.

ADS lines on during thermalization (during commissioning)

I have by-hand re-engaged the ADS lines (and turned on the ITMY line that we don't usually turn on), so we can watch what they do during thermalization, while the spots are controlled with the camera servos. 

The lines have a bit been on-and-off, since we were going to NLN_CAL_MEAS, but in general the lines have been on at the nominal low noise height of 30 counts in the oscillators. 

The first plot is a zoom of what happened while we were naturally thermalizing, including some interesting-ness in the yaw3 value.  It's off and on again due to my not having it turned back on very quickly, but the values around -22 mins are what they had converged to when we switched to camera servos (and the lines get turned off).  Then, I turn the lines back on, and the yaw3 value is quite different from what it had been. I'm not sure if that's a huge deal or not, but it's something I hadn't really been actively aware of until today.

The second plot is over a much longer time, while Jennie and I were also moving some pointing around.  The cyan circles are of the time from the first plot.  The green-purple-green-blue horizontal bar in the middle is roughly representative of the times we were doing different things. Green times are normal, nominal thermalization, with all optics in their usual positions.  During the purple time Jennie was moving the PRM via the camera servo setpoint. During the second green time she had reverted things, so it's back to nominal thermalization.  The blue time is when I was moving IM1 and IM3 via their sliders.  

One conclusion I have is that the time whne we were moving IM1 and IM3 via their sliders (steps of -20 counts each in yaw, waiting, then another set of -20 slider counts each in yaw), the ADS YAW3 line goes farther from zero, indicating that we'd have higher ASC coupling, so it makes sense that the range might have looked worse when Gabriele and Jennie were doing this test a while ago, even though the buildups got better.  I don't think I had gone even quite as far in yaw as Jennie and Gabriele had gone the other day, but we had a fast lockloss, even though nothing was actively being moved (I was holding still for a few tens of seconds).

Jennie has more details on the tests we did this afternoon. 

 

Attemped to run Detchar safety injections

Following instructions in this google doc and here's a past alog70817 to reference running as well.

I started the injections the first time at 1395523038, but ran into an HTTP error when trying to upload the event to GraceDb. I also had forgotten to turn off the PCALX lines and Robert had a shaker going on during part of that time as well. Scrap this one.

Second round with PCALX lines off, no shaker injections, and still ran into the HTTP error at the end of the injections. Start time of these - 1395523842.

Looking at the PCAL AOM, it seems that it has had a lower voltage for ~1 hour before these were ran. PCAL team is currently looking into this and if it would have affected these injections.

High CLF in the IFO

Naoki, Daniel, Nutsinee

Today we increased RF6  from -22dBm to -13 dBm and 8 dBm. We saw excess noise at 8 dBm above 300Hz but no excess noise at -13dBm. REF 12 is the squeezing at -22dB before we started the test. Using the time from alog76553. REF9 and REF10 both show squeezing at -13dBm RF6 at different squeeze angle where one has a better sensitivity at low frequency bucket. REF13 shows squeezing at 8dBm RF6. The excess noise above 300Hz cannot be improved with squeeze angle. Investigation is required.

We turned off ADF sqz angle servo during the test. We readjusted the ADF squeeze angle demod phase and accepted the new value in the SDF.

We are parking RF6 at -12dBm. Since Daniel didn't like the unlucky number 13.

 

Loop	Was (-22dBm RF6)	Now (-12 dBm RF6)
CLF gain	10	0
LO gain	-7	-12
FC LSC gain	-2.6	-0.86
FC ASC gain	0.1	0.03