Alexa, Gabriele, Sheila, Evan

Summary

We have updated the DHARD P filters so that we have a 3 Hz BW when we are on resonance. When we first engage the WFS at 50pm CARM offset, the loop is stable at 100mHz BW. This loop would also be stable at 3 Hz BW; however, we don't want to have to adjust the loop gain as we reduce the CARM offset, so we have left the loop at low BW at 50pm CARM offset. The DHARD Y filters are installed for a high BW configuration; however, they have not been tested yet.

Details for Pitch

The old DHARD P configuration was as follows: FM3 (0.1:0), FM4(:0.01), FM7(SB3.8), FM8(invDhP), FM9 (SB9.8), FM10 (ELP10),Gain = 8. Note: this is the inverse plant that Evan measured in LHO#16520.

The new DHARD P configuration is as follows: FM2(zpk([-1.5+i*4;-1.5-i*4;-3.5+i*1.5;-3.5-i*1.5],[-41+i*70;-41-i*70;-46+i*100;-46-i*100],160000)*gain(2)), FM4(:0.01), FM7(SB3.8), FM9 (SB9.8), Gain = 21. Evidently, we have removed Evan's plant inversion and replaced it with a compensation filter which we designed from Gabriele's model. The attached image shows the difference between the old and new configuration modulo the gains.  

At 50pm CARM offset, we were able to close this loop first with a gain of 21 (low BW), and then with a gain of 165 (3 Hz BW). Since the optical gain changes as we reduce the CARM offset, we decided to leave the loop at 50pm CARM offset with a low BW. Once we were on resonance, this loop was stable and gave us a 3 Hz BW as desired. This is in the guardian now.

Details for Yaw

The old/nominal DHARD Y configuration is as follows: FM3 (0.1:0), FM4(:0.01), FM7(SB3,14), FM8(invDhP), FM9 (SB9.8), FM10 (ELP10),Gain = 8. Note: this is the inverse plant that Evan measured in LHO#16566.

In preperation for a higher BW loop we have installed FM2 (same compenstation as for P), and FM6 (zpk([-0.376+i*15.215;-0.376-i*15.215;0.22+i*19.22;0.22-i*19.22], [-0.405+i*18.071;-0.405-i*18.071;-0.207+i*11.314;-0.207-i*11.314],1)gain(0.488863)). FM6 is designed to compenstate the resonant peaks between 1.5-3Hz in Gabriele's model LHO#17001 (this was confirmed by measurement). The second attached image shows the comparision between the nominal/old configuration and these new filters modulo the gains.

This loop has not been test yet. The guardian currently sets this to the old configuration.

[Alexa, Evan, Sheila, Gabriele]

The attached MATLAB file is a model of how the ITM/ETM plant changes due to radiation pressure. I considered the actuation from the PUM and radiation pressure reaction on the test mass. I used the QUAD model to obtain the mechanical transfer functions I need. I checked that the modeled PUM to test mass transfer functions matches very well a measurement I took this morning. The optical stiffness matrix is computed following Sidles Sigg PLA 354, 167.

The laser input power and recycling gain are parameters of the model.

The first attached plot shows how the ETM actuation transfer function changes when we have 2.8 W in inupt of the IFO. The main low frequency peaks in both pitch and yaw are splitted, but the important thing to note is that the higher frequency features do not change significantly. This is expecially important for yaw: the 2-3 Hz structures don't move. My understanding is that those structures are parallel resonances coming from the upper stages of the QUAD, which are not affected by radiation pressure. It seems therefore that a reasonable strategy to implement larger bandwidth DHARD/CHARD loops is to compensate for the high frequency features, since they shouldn't change significantly with radiation pressure. Then we can develop a controller that works for both pitch and yaw, without a complete plant inversion.

The second attached plot shows the effect of radiation pressure in the hard/soft basis. Peaks move as expected.

Alexa, Evan, Gabrielle, Sheila

Thanks to Keita's work on ETMX Damping (alog 16895) and L2P (16968), we are now locking ALS DIFF with no oplev damping on either ETM.  Attached are spectra from all four stages of both quads, in both cases the ETMY oplev damping was off, for ETMX the solid lines are the old configuration with oplev damping on, the dashed lines are the new configuration described in Keita's alog with the OpLev damping off.  

The wind gusts are at around 30mph, we could see from the ALS control signals that the arms are moving more than usual, so I changed the end stations to the high blends and we are using BRS sensor correction at end X.  (configuration described in alog 16583)

Gabriele, Alexa

Gabriele had accidently unlocked the MC this morning, and had trouble relocking it. I know people had trouble with the MC yesterday morning, but there was no alog about this ... :(

The input power this morning was set to 5 W. When MC would try to acquire lock, the MC2 M3 LOCK L would reach the limit. I adjusted the input power to 2.8W and we locked immediatly. The IMC "DOWN" guardian adjusts the MC common mode board input gain depending on the input power (Power <4, 4 = 10); it's possible that the init and locked gain set for 5 W is not correct. This is something we new was not ideal, and we will correct. If you are having trouble locking the MC and the alignment looks good, I would suggest requesting the input power to 2.8 W and see if that works.  

model restarts logged for Thu 26/Feb/2015
2015_02_26 02:46 h1fw0
2015_02_26 10:55 h1pemey
2015_02_26 12:16 h1pemey
2015_02_26 15:29 h1fw1
2015_02_26 15:38 h1broadcast0
2015_02_26 15:38 h1dc0
2015_02_26 15:38 h1fw0
2015_02_26 15:38 h1fw1
2015_02_26 15:38 h1nds0
2015_02_26 15:38 h1nds1
2015_02_26 15:39 h1nds1
2015_02_26 20:40 h1fw0

Three unexpected restarts. Thursday maintenance: h1pemey modified as part of RFM investigation (was reverted). DAQ restart to acquire new HAM1 vacuum gauge.

model restarts logged for Fri 27/Feb/2015
2015_02_27 12:01 h1fw0

One unexpected restart.

Elli, Dave, Evan

Summary

After today's work to get SRMI locked (LHO#16993), we were able to take some preliminary measurements of the SRC using the aux laser. More work will be necessary before quoting any numbers about the length or Gouy phase.

Details

A brief description of the setup: on IOT2R, there is an auxiliary NPRO (Lightwave) which shoots into the back of IM4 and thereby probes the corner optics. Some of the light is reflected back onto IOT2R (along with some PSL light). An NF1611 is used to read out the beat of the PSL and the aux laser. On ISCT6, there is a second NF1611 which probes OMC REFL (which also contains a beat of the PSL and the aux laser, after these beams have been through the corner optics).

For this measurement, we detuned the aux laser by about 200 MHz relative to the PSL. We then locked the IOT2R beat frequency to the LO of a network analyzer (actuating on the aux NPRO PZT). While Elli kept SRMI locked, we swept the analyzer's LO by about 10 MHz. We recorded the transfer function (ISCT6 beat) / (analyzer LO).

Then Dave clipped beam in front of the ISCT6 PD, and we took another TF. This should in theory allow modes of odd order to contribute more strongly to the ISCT6 beat note.

Then Dave unclipped the beam again, and we took another TF with a −200 MHz detuning.

We're still trying to make sense of the data, but here are some initial impressions:

• The sweeps for the negative detuning have much broader peaks than the sweeps for the positive detuning.
• The positive-detuning sweeps have two peaks per FSR, regardless of whether the PD is clipped or not. There also is not much difference between the clipped and unclipped data (by eye, anyway).

Also, some useful numbers:

• The as-built SRC length is (as far as I know) nominally 56.008 m, as given in D0902838, so we expect an FSR of 2.68 MHz.
• The SRM transmissivity is supposed to be 37%, so the finesse of the SRC at zero detuning is 13.5. [Although I am somewhat confused here: Kiwamu claims that the installed optic is SRM-w14, but galaxy lists SRM-05 (with 20% transmissivity) as "H1 install".]
• The above two numbers give the SRC HWHM as 99 kHz.
• With an as-built SRC Gouy phase of 19°, we expect a transverse mode spacing of 141 kHz.

The data are attached. 02 is at positive detuning, no clipping; 03 is at positive detuning, clipping; and 04 is at negative detuning, no clipping.

Kiwamu, Dave, Evan, Betsy, Elli

Today we locked SRMI in preparation for measuring the SRCL guoy phase.  The lock is not particularly robust, but we are hoping it will be good enough to takea preliminary measurement...

Initial settings:

MICH is locked to LSC-ASAIR_A_LF with input matrix value 0.25, a H1:LSC-MICH_OFFSET offset of -200, a gain of H1:LSC-MICH_GAIN 1600 (or locking with  gain 800 then ramping up to 1600 also works), and an output matrix value of 1 going to BS. FM7 (zpk([1],[150;75+i*129.904;75-i*129.904,1,"n")) and FM9  (ELP40) are on in the MICH loop.

SRCL is locked to LSC-REFLAIR_RF9_I with an input matrix value -3.75, H1:LSC-SRCL_GAIN gain of -50000, a SRCL offset of 0 and an output matrix value of 1 going to SRM.  FM9 (zpk([10],353.553+i*353.553;353.553-i*353.553,1,"n")gain(10)) and FM10 (cheby1('LowPass",2,1,300)) are on in the SRCL loop.

Trigger settings are:

MICH: triggered by ASAIR_A_DC gain 1, upper thereshhold 1000, lower threshold 400.  Filter trigger thresholds On:1000 Off:400, 1seconds, FM2 triggered. (FM2 is zp1:0)

SRCL: triggered by ASAIR_A_DC  gain 1, upperthreshold -1000 lower threshold -1000.  Filter trigger threshold On:400 Off:400, 0.2seconds, FM2 and FM3 triggered. (FM2 is zp1:0 FM3 is zpk([1],[0.1],10,'n')resgain(0.684,3,10).)

Aquiring lock:

We are kind of triggreing the whole thing manually by setting up the above initial settings, and then turning off H1:LSC-CONTROL_ENABLE so no control signals are passed from the LSC output matrix to the suspensions.  If we don't do thisthe optics get kicked around and we don't lock.  We wait untill the AS_AIR flashes look slow (~1Hz) and then turn on H1:LSC-CONTROL_ENABLE.  About 20% of the time we then aquire lock (or we try again).  It is taking <1min to lock.

After aquiring lock:

Engage FM4 (zp 4:0).  Lock stretches are lasting for a few minutes  in this configuration, longest locks~10 mins.

Also: 

-The MICH bounce mode keeps ringing up at 17Hz so we have been damping it every now and then using by turning H1:SUS-BS_M2_VRDAMP_P filter.

-Flashes of SRMI were reaching the H1:ASC-OMC_A/B qpds which was moving OM3.  We disabled the feedback to OM3 by setting the H1:OMC-ASC_POS gains to zero. These gains have been returned to their previous values.

My goal was to find out whether or not the RCG upgrade on January 13th has fixed the 18-bit DAC Major Carry Transitions. However, no evidence of DAC glitches found BEFORE OR AFTER the upgrade.

There were loud glitches in SRCL on January 10th but they do not associate with zero-crossing of the SUS MASTER OUT channels. Even if there were any DAC glitches, unfortunately, they would have had hidden below the noise floor (the glitch at LLO here  is seen at 40 counts max after highpass). On January 16th the high SR3 counts (counts = |65535| and above) do not seems to generate any glitches that are significantly louder than the noise floor. I have also picked out some of the time that the glitches in MICH are well above the noise floor and see if they are coincide with the zero-crossing. All the results are attached below. No DAC glitches found. However, that doesn't mean they do not exist....

I used the 4th order 20 Hz highpass on the control signal because it worked well with Livingston glitch data (plot attached).

The glitch time I used for January 10th are [0.3915, 0.5791, 1.127, 4.474, 7.934, 8.265, 10.78, 11.33, 13.45, 18.88], and [1.952, 2.782, 3.401, 4.932, 6.713, 7.916, 9.398, 11.1, 14.68, 15.96, 16.38, 18.23] for January 16th. I used Laura's python script (from alog16354) to identify spots on the SUS data that correspond to the glitches.

Suddenly I found that the EX green beam is not clipped much any more by looking at the QPD spectrum, and it turned out that the green beam was totally misaligned. This happened as the EX green QPD centering servo was turned off yesterday, and eventually TMS drifted away.

Anyway, I turned the QPD centering back on, and the clipping was immediately back.

I adjusted the QPD PIT offsets and found that setting QPDA PIT offset to -0.9-ish while keeping QPDB PIT offset to zero-ish fixes the clipping and keeps both of the QPD NSUMs to be about 1.02 to 1.03 (good).

First attachment shows clipping spectra, the second one is after the big offset was set.

Elli took pictures of the green beam in both of these states, and the beam is at a nice position when being clipped, and the beam moves down below by maybe a couple cm or more when it's not clipping.

Since we cannot move TMS up or down, there's not much hope to unclip the beam without drastically off-center the green beam.

LVEA: Laser Hazard
Observation Bit: Commissioning   

07:15 Karen & Cris – Cleaning in the LVEA
07:50 Sudarshan – Going into the LVEA
08:07 Hugh & Jim – Going to HAM1 to unlock HEPI
08:45 Christina, Karen, Cris – 1st cleaning of cleanroom in LVEA
08:51 Stuart & Jason – B&K hammer testing ITM-X OpLev pier
09:06 Filiberto – In LVEA west bay working on Test Stand racks
09:49 Stuart & Jason – Out of the LVEA
10:05 Students to see Richard on site
10:17 Apollo on site to work on the VPW AC system
13:02 Kyle – At HAM1 working on pumping setup
14:00 Kyle – Out of the LVEA – GV5 and GV7 are open - HAM1 is on ion pump only
15:10 Apollo on site working on the VPW AC system

[Jason O, Stuart A]

Power spectra measurements taken earlier this week (see LHO aLOG entry 16917) suggested we should target the ITMY Oplev for our initial B&K hammer measurements.

Problems with the B&K data acquisition card yesterday prevented us from taking measurements. However, thanks to Calum, who had seen simialr occurrences at Caltech, we were able to get back up and running. Just to note that, the B&K Capabilities Wiki collates relevant documentation, and a Troubleshooting guide has now been added.

We proceeded to take B&K measurements for both the ITMY RX Leaner and TX Large Dwarf OpLev piers. A tri-axial accelerometer was clamped to a cross member near the top of each pier (see images of the set-up below). Measurements were taken using the "Simple Hammer Display 3.pls" template available in T1000697. A hammer trigger threshold of approximately 8N was found to work best. ASCII data has been exported from the B&K Pulse analysis software directly into text files (*.txt) so that it can be plotted using the "BandK_plot.m" script.

Plots are available below for the two pier types tested. The frequency of the first resonances observed (X response to X excitation) can be compared to previous measurements made on production units at Caltech, see T1100152, as well as those measured at LLO. The resonance measured at Caltech is denoted by the dashed black line in the plots and was taken when the various structures were bolted and grouted.   

Resonance Summary:

OpLev   Type   Pier Name     DCC#      Caltech   LHO 
ITMY    RX     Leaner        D1001297  34.5 Hz   21 Hz 33 Hz
ITMY    TX     Large Dwarf   D1000452  85.5 Hz   34 Hz

First structural resonances for both ITMY Leaner & Large Dwarf piers appear consistent with similar non-grouted piers measured at LLO. However, the Leaner pier exhibits an extra resonance feature around 34 Hz, which is most likely due to the bolting configuration (n.b. as well as no grout, there are also no vibration absorbers fitted). Interestingly, there is no evidence of the 3.3 Hz or 6.6 Hz features, present in the OpLev Spectra, seen in either pier.

All data and plotting scripts have been committed to the SUS svn:

/ligo/svncommon/SusSVN/sus/trunk/Common/MatlabTools/BandK_plot.m

/ligo/svncommon/SusSVN/sus/trunk/QUAD/H1/ITMY/BandK/
SimpleHammerDisplay3-ITMY-ITMY_OpLev_RX_LPier_T1a-SUSunlocked-ISIunlocked-VAnotfitted-Ximpact.txt
SimpleHammerDisplay3-ITMY-ITMY_OpLev_RX_LPier_T1a-SUSunlocked-ISIunlocked-VAnotfitted-Yimpact.txt
SimpleHammerDisplay3-ITMY-ITMY_OpLev_TX_LDPier_T1a-SUSunlocked-ISIunlocked-VAnotfitted-Ximpact.txt
SimpleHammerDisplay3-ITMY-ITMY_OpLev_TX_LDPier_T1a-SUSunlocked-ISIunlocked-VAnotfitted-Yimpact.txt

n.b. the raw B&K Pulse *.pls file for each measurement was also checked into the svn along with the exported *.txt file.

The plots below are the 24 hour OpLev trends

Altough the high frequency is kind of screwed up by the wandering line, we can get some interesting information about the lower frequencies.

The full report can be found at the following address:

https://ldas-jobs.ligo.caltech.edu/~gabriele.vajente/bruco_1108981336/

The most interesting coherence is with SUS-BS_M1_ISIWIT_PIT_DQ, which seems enough to explain most of the noise up to 100 Hz. This is consistent with what Sheila told me, i.e. that we're not fully using BS ISI.

For those interested in the BruCo details, I managed to reduce a lot the time needed to analyze the data, basically with the following modifications: split coherence computation into the single FFT computations, to reduce redundancy; parallelize the computation and expcially the disk access using all available processors. This brought down the typical execution time to analyze 10 minutes of data from 8 hours to about 20-30 minutes. The new code is attached.