Sheila, Daniel, TVo

Executive Summary

It doesn't look like the TMDS caused any extra absorption on ETMX.

Following my aLOG-38476, where I alluded to the possibility of increased loss due to absorption as a reason we had trouble locking last night, we wanted to try to measure a loss in each arm and compare them to each other.  This is done by looking at the reflected power and seeing the difference between locked and unlocked.  In the attached time series, the first drop is the X-ARM locking, the second drop is the Y-ARM locking.

For this measurement, we configured the IFO as such:

- The arm cavity of interest was first locked on ALS to get aligned well (it would be mis-aligned during the actual measurement)

- The other arm cavity was misaligned

- Then we locked a single arm with Guardian (ALIGN_IFO)-

- We also turned on the DC Centering on the AS WFS in order to maximize their power.

- Then to get a decent dip in the reflected power off the arm cavity, we increased the input power from the PSL from 2 Watts to 25 Watts.

Below is the summary of results:

XARM

YARM

Some of the sensors in the AS port didn't give us good results when we locked and unlocked but it's not fully understood why.  The total loss is a combination of mode-matching, alignment etc and these were not taken into account.

J. Kissel, S. Dwyer

We're getting verbal alarms about timing errors constantly because the end station GPS receivers are in the wrong configuration (see LHO aLOG 38439).

To save our sanity until the GPS receiver configurations are fixed, we've commented out the TIMING error function from the verbal alarms script by excluding it from the , all_tests list defined on line 1253 of 
     /opt/rtcds/userapps/release/sys/h1/scripts/VerbalAlarms/VerbalAlarms.py

This should be put back in once the GPS receiver configuration is fixed.

It seems that our trouble locking last night was due to a change in the initial alignment references.  We are now locked at DC readout with a recycling gain of just over 30.  

In order to lock I changed the TR_CARM offset (sqrt(TRX+TRY), each normalized to single arm power) at which we transition to the REFL PD from -41 to -35.  This allowed us to lock with a recycling gain of just over 20.  After that I moved the ITMs by hand with all the ASC except the soft loops on, and we could recover a good recycling gain with the old TMS offsets.  That means that the problem was just our initial alignment references. 

At DC readout, I used the scripts described in 37033 to reset the green alignment references, although I had trouble getting the Yarm to lock in green. Screenshot shows the changes in settings. 

TITLE: 09/01 Day Shift: 15:00-23:00 UTC (08:00-16:00 PST), all times posted in UTC
STATE of H1: Commissioning
SHIFT SUMMARY:  efforts to recover NLN continue
LOG: activities today

• Sep 01 2017 15:24:15 UTC Richard starting Initial Alignment
• Sep 01 2017 15:34:36 UTC Liz in LVEA vertex
• Sep 01 2017 15:34:04 UTC Karen to optics lab
• Sep 01 2017 16:04:20 UTC Karen out of optics lab, heading to VPW
• Sep 01 2017 16:07:47 UTC ThomasV locking X arm for absorption measurements
• Sep 01 2017 16:11:43 UTC Liz out of the LVEA
• Sep 01 2017 16:12:22 UTC Ed to cleaning room
• Sep 01 2017 16:20:45 UTC Kyle and Corey at EX in LVEA to move equipment
• Sep 01 2017 16:30:51 UTC I reset the Noise Eater
• Sep 01 2017 16:31:06 UTC Instrument Air alarm at EX
• Sep 01 2017 16:47:08 UTC Ed to LVEA to plug in phones
• Sep 01 2017 17:18:35 UTC Kyle and Corey at EY
• Sep 01 2017 17:34:14 UTC Kyle and Corey done at EY
• Sep 01 2017 17:34:35 UTC BRS and SC back on
• Sep 01 2017 18:14:41 UTC Kyle to EX VEA, leaving BRS on
• Sep 01 2017 18:16:34 UTC Ed out of LVEA, restoring phones
• Sep 01 2017 18:42:29 UTC Kyle back from EX
• Sep 01 2017 22:25:01 UTC Ed to Clean Storage
• Sep 01 2017 22:30:04 UTC Ed back from Clean Storage
• currently: Sheila and Thomas have H1

Summary: It floats!

Previous update here: 38330

On Thursday afternoon, Travis helped me pull the wires and get them clamped. The one issue we had was two screws failing and almost breaking off. Neither of us thought that we were over torquing them, but it is possible. Luckily we stopped as soon as we felt them giving so they were still partially attached and we could just back them out slowly. I forgot to get a picture of these but they are still sitting on the table and I'll grab one next week.

Today I went in to release and balance the suspension and successfully was able to do so. It took a fair amount of poking, sliding, and exchanging masses, but eventually I used 19.65kg excluding the screws.

Attached are a few pictures of the wires and the floating suspension

And what else floats?

Apples! Bread! Very small rocks!

A duck!

J. Kissel

Even though we no longer suspect there's anything wrong with the ETMX test mass after discharging on Tuesday and Wednesday (aLOGs with convincing refuting evidence to come from Sheila and Thomas; see claims of suspicion in LHO aLOG 38476), I've gathered some digital red camera images from the camera archive, 
     /ligo/data/camera/archive/2017/09/01/
         H1 ETMX (h1cam25)_2017-09-01-22-04-39.tiff (or without the linux handling of badly used characters 'H1 ETMX (h1cam25)_2017-09-01-22-04-39.tiff')
     /ligo/data/camera/archive/2017/08/29/
         H1 ETMX (h1cam25)_2017-08-29-02-02-08.tiff ('H1 ETMX (h1cam25)_2017-08-29-02-02-08.tiff')

imported them into matlab (I'm still using the workstation's default version 8.0.0.783 [R2012b]), 
    [dcreadout_0829,~] = imread([imageDir{1} 'H1 ETMX (h1cam25)_2017-08-29-02-02-08.tiff']);
    [dcreadout_0901,~] = imread([imageDir{2} 'H1 ETMX (h1cam25)_2017-09-01-22-04-39.tiff']);

and subtracted the images, 
    difference_ETMX = (dcreadout_0829 - dcreadout_0901);

and plotted them, e.g. 
    figure(1);
    imagesc(dcreadout_0829);
    colormap(gray);

just to see if there were any major changes in point scatters because of the TMDS use.

There is no major change in point scatters as a result of the use of the TMDS system on ETMX.

Pump down curve attached.  

Corey helped me move the TMDS stuff from the X-end to the Y-end today.  I'll file a WP today for September 4th (Monday-Thursday) TMDS of ETMy.  

BRSX looks stable. BRSY shows a re-centering by Jim W.

[ colour code reminder: GREEN = front end restart, BLUE = daq restart, PURPLE = unexpected or unanticipated restart, RED = crash ]

model restarts logged for Thu 31/Aug/2017 - Wed 30/Aug/2017 No restarts reported

model restarts logged for Tue 29/Aug/2017
2017_08_29 10:01 h1iopsusey
2017_08_29 10:01 h1susetmy
2017_08_29 10:01 h1susetmypi
2017_08_29 10:01 h1sustmsy
2017_08_29 10:13 h1hpietmy
2017_08_29 10:13 h1iopiscey
2017_08_29 10:13 h1iopseiey
2017_08_29 10:13 h1pemey
2017_08_29 10:14 h1alsey
2017_08_29 10:14 h1caley
2017_08_29 10:14 h1iscey
2017_08_29 10:14 h1isietmy
2017_08_29 13:18 h1hpiham1
2017_08_29 13:18 h1hpiham6
2017_08_29 13:18 h1iopseih16
2017_08_29 13:18 h1isiham6
2017_08_29 16:38 h1broadcast0
2017_08_29 16:38 h1dc0
2017_08_29 16:38 h1fw0
2017_08_29 16:38 h1fw1
2017_08_29 16:38 h1fw2
2017_08_29 16:38 h1nds0
2017_08_29 16:38 h1nds1
2017_08_29 16:38 h1tw1

Maintenance day. BIOS work on h1susey, unexpected restarts of h1seisy and h1iscey. Unexpected restart of h1seih16. DAQ restart for Beckhoff ecatc1 upgrades.

model restarts logged for Mon 28/Aug/2017
2017_08_28 11:53 h1susitmy
2017_08_28 12:00 h1broadcast0
2017_08_28 12:00 h1dc0
2017_08_28 12:02 h1fw0
2017_08_28 12:02 h1fw1
2017_08_28 12:02 h1fw2
2017_08_28 12:02 h1nds0
2017_08_28 12:02 h1nds1
2017_08_28 12:02 h1tw1

SUS model fixing with associated DAQ restart

model restarts logged for Sun 27/Aug/2017
2017_08_27 16:16 h1lsc
2017_08_27 16:18 h1omc
2017_08_27 16:18 h1susitmx
2017_08_27 16:20 h1susitmy
2017_08_27 16:23 h1broadcast0
2017_08_27 16:23 h1dc0
2017_08_27 16:23 h1fw0
2017_08_27 16:23 h1fw1
2017_08_27 16:23 h1fw2
2017_08_27 16:23 h1nds0
2017_08_27 16:23 h1nds1
2017_08_27 16:23 h1tw1

New ITM-SUS, LSC and OMC code. Associated DAQ restart.

model restarts logged for Sat 26/Aug/2017 No restarts reported

ETMY and ITMY both seem to be either near or out of spec.

Added 100ml to crystal chiller.

• H1 unlocked
• ran Initial Alignment
• ThomasV has X arm locked in red for absorption measurements

We preform more FINESSE simulations on the AS72 scheme, focusing on how the signal changes when we switch to the new SRM (T_srm=0.37 -> 0.32), and possible solutions if we are sitting on the ill-conditioned regions. 

Conclusions:

1. If the AS72 sensing matrix measured now (with differential wavefront distortion; T_srm=0.37) behaves well, the sensing should be fine after we replace the SRM.

2. If the sensing matrix for AS72 is close to being degenerate, after replacing the SRM the sensing is likely to be even worse.

In case that the sensing matrix is bad (e.g. we are sitting at AS port gouy phase 45-75 deg in the plots), possible solutions:

    i. Increase the SRC one-way gouy phase to ~20 deg (e.g. with SR3 ring heater; current nominal SRC gouy phase is ~18 deg). This should be the BEST solution. For more details, please see 37222.

    ii. Use ASA_72Q and ASB_36Q for SRM/BS sensing. This scheme work only for a very narrow (AS port gouy phase + diff lens) space. 

    iii. Use ASA_72Q and (ASA_72I - ff x ASA_DCQPD) where the ff is set to decouple the spot centering. This should cover the AS port gouy phase space corresponding to 45-70 deg in the plots. 

Sheila, Jeff, Corey, TVo

After opening the gate valve at ETMX, we started locking to measure the charge in-lock using Sheila's script in aLOG-38387

However, we had trouble getting past the CARM_10PM state and Sheila noticed that the alignment might be bad because of the ratio of REFL power locked to unlocked was too low and the PRC Gain was lower than normal as well.  Even after running initial alignment twice, the same problem still persisted so we're thinking there may be some extra absorption in the arm cavities, we will follow up tomorrow.

Jonathan, Dave:

we were reminded this week of a second trend NDS1 issue if the user asks for data which spans a DAQ restart. NDS1 has an optimization turned on which reads the channel list from the first second trend frame file in the time sequence and applies it to all subsequent files. It looks for a missing file as an indicator the DAQ was restarted and the channel list needs to be re-read. This worked when second trend frame files were one minute in length, but does not work when they were extended to ten minutes in length.

The temporary solution is that I will rename the partial second trend frame file by pre-pending a dot (ditto with its associated md5 file) after I have ascertained that LDAS has archived the file.

The attached dataviewer one hour second trend plots of the EX building outside temperature spanning yesterday's DAQ restart show a before and after this fix was applied. Before the fix the temperature went from 99F to +/-400F, after the fix the correct temperature is shown with a 10 minute data gap.

The long term fix is to use the run-number in the frame header as a channel configuration index.

This completes WP #7119 and WP# 7134.  

PT525 finally "woke up" when, by opening GV20, the X-end pressure changed by a factor of x10.  We have noted in the past how the new-to-us Bayard-Alpert/parani wide-range gauges differ from our traditional Cold Cathode gauges, both in magnitude and general behavior.  This difference is greater as the real pressure increases.  The Parani portion of the new gauges agree well with our traditional convection enhanced paranis but in the HV pressure ranges they are just "different".  We know that the CC gauges pump with about 1 or 2 l/s pump speed while the nude BAs don't pump.  For the conductance due to our arrangement of fittings, this can explain a difference between the two gauge technologies of up to 50% but not factors of 3, 5 and 10! 

The newest odd behavior shows up as a complete "lock up" of the gauge.  It gives valid but non-changing output in the example shown recently at the X-end (see attached graphs)  
 

 
Note:  In WP #7134, I mistakenly indicated that I was leaving the 1.5" valve which isolates GV20's gate annulus from the rest of the annulus piping "open" after opening GV20 and that this was a policy change.  In fact, I am leaving this 1.5" valve "closed", as is standard policy when the 44" gate is open.  I believe that the new, "Chandra era", change in practice applies only to when the 44" gate is closed for prolonged periods.  For this case, we will now be leaving the 1.5" open so as to eliminate the need to "dump" the un-pumped gate annulus volume into a temporarily connected aux. pump cart.  

J. Kissel

I've measured the effective bias voltage of the H1SUSETMX and H1SUSETMY ESDs in the traditional fashion (as measured by the zero in actuation strength [rad/V_S] from each quadrant reported by the optical lever). 
Excellent news: All Quadrants of ETMX ESD report 0 +/- 3 [V] effective bias voltage. AWESOME.
where, via the same, single-number metric (the mean and standard deviate of the weighted mean in all four quadrants, over two angular degrees of freedom), ETMX used to have 17 +/- 19 [V] effective bias voltage. Of course, you should look at the plots, since this single-number metric only works when all the weighted means of quadrants are very similar.

Note, however, this is a bulk measurement of all the ways that the ESD can be effected by charge. As Leo has discussed (see T1500467), and Sheila has recently resurrected (see LHO aLOG 38387), the force on the test mass is determined by 

    F_ESD = A (V_B - V_S)^2 + B (V_B - V_S) + C (0.5*(V_B + V_S) - V_R)^2 + D (0.5*(V_B + V_S) - V_R) + E 


where V_B, V_S, and V_R are the applied bias voltage, the applied quadrant control "signal" voltage, and the voltage of the surround cage components, respectively. A, B, C, D, and E are coefficients that depend on the geometry of the ESD pattern, the distance between the test mass & reaction mass, as well as the surrounding cage components (ring heater, EQ stops, fiber guards), and the polarizability of the test & reaction mass. This effective bias voltage measurement is a measure of a collection of all but one of these coefficients,
    V_eff = 0.25 (D - 2 C V_R - 2 B) /  (C - 4 A)

Once we recover the interferometer, Sheila and I will remeasure each coefficient independently, a. la LHO aLOG 38387, and BSC ISI excitations as in LHO aLOG 38132. Let's hope this has some impact on the sensitivity.

ETMY's effective bias voltage remains high (using the above metric, 51 +/- 34 [V]), but we're attacking the charge on that chamber next week. May we have the same amounts of success there!

Three cheers for Rai and the systems design team!