E. Bonilla, A. Fernandez-Galiana, J. Kissel

A few days ago, I took the first transfer functions of the newly free, newly installed H1 SUS OPO, and instantiation of an optical parametric oscillator suspension (OPOS) -- see initial results in LHO aLOG 40749. The data can be found here:
    /ligo/svncommon/SusSVN/sus/trunk/VOPO/H1/OPO/SAGM1/Data/
        2018-02-27_2209_H1SUSOPO_M1_WhiteNoise_L_0p02to50Hz.xml
        2018-02-27_2209_H1SUSOPO_M1_WhiteNoise_P_0p02to50Hz.xml
        2018-02-27_2209_H1SUSOPO_M1_WhiteNoise_R_0p02to50Hz.xml
        2018-02-27_2209_H1SUSOPO_M1_WhiteNoise_T_0p02to50Hz.xml
        2018-02-27_2209_H1SUSOPO_M1_WhiteNoise_V_0p02to50Hz.xml
        2018-02-27_2209_H1SUSOPO_M1_WhiteNoise_Y_0p02to50Hz.xml

In order to bring this new suspension type into the fold, I've started to create the standard testing and commissioning infrastructure for analyzing the data in full, namely
    - With extensive help from Edgard, I've updated the single stage suspension model generating function,
        /ligo/svncommon/SusSVN/sus/trunk/Common/MatlabTools/SingleModel_Production/generate_Single_Model_Production.m
      to accept the new dynamical model,
        /ligo/svncommon/SusSVN/sus/trunk/Common/MatlabTools/SingleModel_Production/ssmake_voposus.m
      and the new parameter file
        /ligo/svncommon/SusSVN/sus/trunk/Common/MatlabTools/SingleModel_Production/oposopt_h1susopo.m
      such that I can produce a dynamical model of the OPOS on the fly.
    - I've copied over Alvaro's updated OSEM2EUL and EUL2OSEM matrix generation script into a more standard name and format,
        /ligo/svncommon/SusSVN/sus/trunk/VOPO/Common/MatlabTools/make_susopos_projections.m
      such that it's functional, and can spit out those matrices for use on the fly.
    - I've created a new DTT measurement analysis script that calibrates the raw data, and compares that data against the model,
        /ligo/svncommon/SusSVN/sus/trunk/VOPO/Common/MatlabTools/plotOPOS_dtttfs_M1.m
      so that we can debug an individual set of TFs for rubbing and expected/unexpected cross-coupling, and save the data to a standard, matlab friendly format so that we can compare different measurement sets over time and across sites.
The last thing on the to-do list is to create a "plotallopos_dtttfs.m" that will then contain a list of all measurements of any OPOS that will do that comparison over time and across sites. Also, as I'm sure you've noticed, there're still lingerings of "vopo" in all the scripts and folders that I haven't touched yet in the interest of getting the data processed. I'll save both for another day.

After all that infrastructure work, and comparison against the model, we clearly have our work cut-out for us, and the terrible T and Y TFs explain why I wasn't able to get rudimentary damping loops to work.
TJ has said that he's moved around a cable by H3 that he thinks was rubbing today, so I'll make attempts to remeasure the thing ASAP. Didn't get to it tonight 'cause of the work needed to create the above infrastructure. Sorry!

J. Kissel, E. Bonilla

I've formally processed Travis' interim transfer functions that he took a few days ago after he and Betsy were approaching happiness about the suspension being free (see LHO aLOG 40585). Both chain's results are interesting:

Main Chain:
   - I see some high Q cross coupling on the higher frequency resonances that I don't like between 2 and 3 Hz. M0 P to P shows it the worst, and it seems to have R0 T, M0 Y, R0 V, and M0 L modes that aren't expected, and indicate interaction between the chains.  We should look take a look for subtle rubbing at or between the lower stages of the main chain.

Reaction Chain:
   - I've worked with Edgard to get the new model for the AERM reaction chain suspension up and running, and incorporated into the traditional testing & commissioning infrastructure. The required changing several of the files inside the 
       /ligo/svncommon/SusSVN/sus/trunk/QUAD/Common/MatlabTools/QuadModel_Production/
     directory, but all is now functional. 
     I'm excited to say that the model matched the data extremely well right out of the box. Also, the obvious, expected changes between an ERM and an AERM are evident in the middle-frequency modes, most obvious in the Longitudinal and Yaw TFs, which typically only depend on highly controlled parameters -- the mass and lengths of wires. We only had to 
         - adjust the mass of the PenRe (Reaction Mass' PUM) to what Betsy reports in LHO aLOG 40291, (from 65.1 to 64.356 / kg)
         - update the PenRe's Pitch and Yaw (and associated off-diagonal) moments of inertia to match -v5 of T1500563, 
         - fiddle with the Roll moments of inertia for the PenRe and Annular Reaction Mass itself (I2x from 0.998 to 0.6 / kg.m^-2, I3x from 0.3064 to 0.4 / kg.m^-2)
     in order to get the data to match as well as shown. The metrics for success (and what originally didn't match well) were the T / R modes (now modeled) at 0.91 and 2.68 / Hz. The latest model parameter set is committed to 
         /ligo/svncommon/SusSVN/sus/trunk/QUAD/Common/MatlabTools/QuadModel_Production/quadopt_aerm.m
    The biggest descrepancy from the model is the Pitch transfer function, but we've seen this with every reaction chain type -- the ESD, PUM adn UIM OSEM cabling have always significantly altered / stiffened up the Pitch TF. No surprise or alarm there.
    Nice -- glad to see a new suspension type measured and successfully compared to its first-principles model!

    But, we've still got some work do to before we're happy with the dynamics and are confident the isolation stages are free.

Started CP4 bake today, with some hiccups. The heater was being controlled by a TC that was in a blind spot and not reading a well represented temperature, causing the enclosure to heat up too fast. The heater is now controlled via TC#1 which is clamped to a valve body on bottom of CP4, and can be read out from camera image found here:  https://lhocds.ligo-wa.caltech.edu/cr_screens/png/video0-1.png. The other three TCs are:  TC#2 reads air near GV 11 (east end); TC#3 reads air at exit of heater duct (west end); TC#4 reads air at top near GN2 regen pipe. 

We discovered a bigger problem - the air in east end of enclosure is not being circulated because the supply and return are on top of each other at west end, so the east side (GV11) is heating at a higher rate. Tomorrow we will work with contractors to install an extension duct at return to balance air mix. We decided to stop the heater for tonight to protect GV11. (edit:  the heater had already tripped because at least one of the four TCs had exceeded the setpoint+10C - good to know the safety protocol works!)

A quick rundown of today's activities in HAM6:

• Installed mirror into ZM1 suspension
• Balanced the tip-tilt as best I could without a laser. When we have a beam I will most likely need to adjust the pitch slightly and then recenter the OSEMs.
• Rerouted the beam diverter cable and placed the BD on the South side og the table. The iris used to keep the alignment for ZM1 is currently in the location that this BD needs to be, so we will have to put it in its final position when we can get the iris out of the way.
• Cleaned up some of the other cables below the table since I was already down there for the BD cabling.
• Routed and cabled ZM1.
• Sheila and I then connected the correct cables to each respective OSEM, took and open light measurement, and then proceeded to center them with their new offsets. They will need to be recentered after the pitch is adjusted with a beam.
• We packed up the extra screws, tools, and such and left only what we think we will need. I also removed the cable tags that Hugh had placed to identify each cable.
• Jeff K had mentioned that the OPOS was rubbing, so I poked around and found the H3 OSEM cable to have twisted a bit in one of its cable brackets which was causing it to lightly touch the suspended platform. Hopefully that fixed it.

Pictures to come tomorrow.

J. Oberling, E. Merilh, M. Heintze

Today we looked into the scatter issue noted yesterday.  We removed the new FE pick-off and looked at the scatter again, see attached pictures.  As can be seen, still a lot of scatter.  The going theory is that we are not optimally mode matched from the NPRO to the MOPA, and the scatter is being caused by amplified spontaneous emission (ASE) due to the unused gain from the mode mis-match.  Since the FE is running good, we have good visibility to the PMC, and the mode scan from the most recent DBB measurements indicated low higher-order mode content in the beam, our solution will be to create a spatial filter to block this scatter; this will likely be installed in front of WP02.

We then reinstalled the new pick-off and optimized the alignment, being sure to tilt it off enough to avoid saturating the PD.  It is currently not calibrated, we will have to do that at some point in the future (all PDs that measure power will be re-calibrated at the end of the install).  There was a request to take a transfer function of the FSS out to 10MHz to investigate a bump at ~5MHz noticed at LLO.  Thinking this would be an quick thing to do (famous last words...), we decided to get it in before lunch.  Unfortunately, the FSS would not lock.  After playing with several settings for tens of minutes, Matt was finally able to get the FSS locked.  We then set about tweaking the alignment, which was actually a good bit off.  At this point we broke for lunch.

After lunch, having abandoned the "quick" FSS TF to make progress on the 70W install, we began setting up for measuring the beam caustics of the FE for mode matching.  WP02 was re-installed in the system (we had to remove it to fit the Wincam for beam profile measurements during the installation of the FE pick-off, AMP_WP01 and AMP_PBS01).  We then removed all of the unnecessary optics in the FE DBB path (M09, M10, CCD01, CCD02, L05, L06, the bullseye sensor and its mirror) and set up a ruler for accurate moving of the Wincam during the caustic measurement.  Matt gave me an output coupler to use to dump the majority of the FE power, and we mounted a pump light filter (to keep any pump light from contaminating the measurement).  We also confirmed that lens L15 has a 200mm focal length, and the rough location of this focus for ease in setting up for the measurement.  This complete, we called it a day.  Tomorrow morning we will measure the beam caustics of the FE and begin modeling for a mode matching solution.

TITLE: 03/01 Day Shift: 16:00-00:00 UTC (08:00-16:00 PST), all times posted in UTC
STATE of H1: Planned Engineering
INCOMING OPERATOR: None
SHIFT SUMMARY:

CP4 Bakeout begins, HAM6 Squeezer work continues, EY Pcal work completed, EY TMS checked.
LVEA is Laser HAZARD now & EY is back to SAFE.
LOG: (All times UTC)

• 16:03 Resetting EY Dust Monitor (Jeff B.)
• 16:08 MY--CP4 Bakeout will be beginning at 16:30, may get alarms (Chandra)
• 16:16 Moving snorkel from MX to MY, for MY electrical/camera work by Ken (M&M)
• 16:50 EY--PCal alignment work for the day begins (Rick, Travis, Niko, Marek, Dennis)
  • 19:44 Breaking for lunch
• 17:05 EX--Moving legs of Test Stand out of the way at EX (M&M)
• 17:21 HAM6--ZM1 & beam diverter work (TJ)
• 17:42 PSL work (Jason, Ed)
  • 17:58 Matt joining
  • 22:25 Continuing work

Patrick coverage from 18:00-19:00

• 18:05 Rick called to say EY is laser hazard
• 18:07 Dave tried to add reservation and crashed reservation code. I restarted display on nuc2. Display on video4 needs restarting.
• 18:19 Betsy to end X to move furniture
• 18:33 Well testing company through gate
• 18:36 Tour leaving CR

• 19:44 EY Laser SAFE
• 19:45 Laser safety hardware work in LVEA (will kill lasers...i.e. at EY)
  • 19:47 Turned OFF TCS x & y  lasers (Corey for Richard)
• 21:38 Cabling for Chiller summing chassis in CER (Dave & Fil)
• 21:38 Squeezer Table cabling (Fil)
• 21:50 Continuing HAM6 work (TJ, joined by Sheila, Nutsinee, Terry)
• 21:58 LVEA being transferred to Laser Hazard (Sheila)
• 21:53-23:09 EY TMS work (Georgia, Keita)
  • Also transitioning to Laser Hazard/Safe
• Cheryl noticed Y3_PT210B_PRESS_TORR slowly increasing on Wall TV (video0); usually these are flat.  Gerardo consulted, he said it was due to the temperature increase.
• 22:09 Guardian work for HAM2, BS, ITMy (Jamie)
• 22:11 Increase transimpedance for OPO Trans DCPD (Marc)
• 23:36 EY TMS transfer functions (Betsy)

[Keita, Georgia]

With the new EY alignment, Keita and I adjusted the TMS pitch and yaw to steer the green ALS beam (reflected off ETMY) back onto the in-air table.

Keita adjusted one of the weights on the TMS table to improve the alignment in pitch. It is now 4" from where it was in O2, in the -y direction.

The TMS slider values are now P = 14.6 urad, Y = -217 urad, with the reflection roughly aligned. The full slider range in pitch is +/- 608 urad, so we think even with this rather large offset there is enough range to steer the beam 1m over 4km.

Finally, some of the OSEM sensors on the upper stage of the TMS suspension are not in the center of their range, something we might want to go back in and fix.

J. Kissel

While confirming the at-vacuum driven health of H1SUSITMY, I've found that -- although we've cured the RO RT OSEM shorting problem (LHO aLOG 40787) --  the main chain (M0) still *sometimes* needs as 50000 ct vertical offset in order to prevent noisy, potentially rubbing-like V and R transfer function results. I say *sometimes* because I measured the M0 V and R TFs on 2018-02-26, saw noisy results, and the first vertical mode was ~0.04 Hz higher than expected. Suspecting that we still needed the 50000 ct vertical offset that was used during most of O2 (I couldn't find an aLOG to reference), I turned ON the offset, measured V and R (and the rest of the DOFs) today, 2018-03-01 and saw that the TFs cleaned up nicely with the 1st vertical mode in the right place. But -- just to be sure -- I took the V transfer function again *without* the offset, and saw that the transfer functions still look acceptable.

The first attached image shows the results and what I mean regarding "noisy". Hopefully the legend straight-forwardly labels the above described measurements.

With these initially confusing results, I started trending temperature and vertical positions of corner-station BSC suspensions.  The history: after the short vent to repair the ITMY R0 RT OSEM, the vertical position dropped to ~ -100 [um], then at 2018-02-26 08:58 UTC (about 01:00a PST local!!), shot up to ~ -80 [um], stayed there until 2018-02-27 20:55 UTC, then it began to exponentially continue to rise to pre-vent quick-vent value of ~ -50 [um]. The 2018-02-26 measurement I took happened to be in the ~1.5 day 80 [um] time, where I suspect the suspension was/is just subtly rubbing like it was in O2 that drove whomever to add the 50000 ct V offset originally. 
I also show a 50 day trend of all BSC SUS. All suspensions show the same vertical position evolution as ITMY, indicative of a corner-station-wide, in chamber temperature trend, not something quirky about ITMY.
The in-loop, averaged corner station temperature sensor for Zone 1A near the BSCs reports a flat temperature for many moons, not unexpectedly.
The vertical position of ITMY during the extra clean 2017-12-20 measurement was +30 [um]. That measurement was taken in air, just before the doors went on over the holidays (see LHO aLOG 39829) -- showing that if the suspension is riding high due to cool temperatures, it's possible to get a very clean transfer function, even at air, so something is clearly interfering / rubbing / messed up when the SUS runs hot.

Frustrating. Especially, and probably because, it's intermittent -- we forgot about the need for this offset, and didn't address it either times we opened up BSC1 to play with ITMY. And we *always* forget about temperature during the chaos of a vent. Stinks. 

So, for now, I leave the 50000 ct offset in place (this is about 1/4th the range of the DAC), but we should continue to watch this suspension with respect to vertical position and temperature, hoping that it regains more of its vertical position and continues to improve whatever is going on.

The .pdf attachments show the full suite of M0 and R0 transfer functions taken on 2018-02-26 when the main chain was low / rubbing, and on 2018-03-01 when the 50000 ct vertical offset was applied (and the temperature had risen a bit) showing clean transfer functions. I did not take a full suite today (2018-03-01) without the offset.

On a happier note that the Reaction (R0) chain looks clean after the repair of the R0 RT OSEM. And for the record, the attached results are with and without the 50000 ct main chain V offset, and they show no difference -- implying that whatever is going on with the Main Chain is an independent problem.

Overall:  EY, HAM6, & PSL work looking good

HAM6: 

• TT testing, cabling work, and then would like to go to laser hazard tonight/morning (will stay hazard for a while). 
• NOTE:  If you can avoid walking around HAM6 this would be good, it's a congested area & there will be an IR laser in operation there.). 
• Work on making a work platform on south side of chamber. 
• Possibly ready for beam mid-next week (so need covers off Septum window). 
• "VOPO might be rubbing," Jeff K. 
• Will need viewport simulator.

EY: 

• Today:  TMS Check (Keita & Georgia)
• Tomorrow:  B&K (Jeff K, Jim), Ground loop checks (Fil)

EX (Next Week)

• ALS/PCal Lock & Tag Out
• Door off Monday
• Need to be thoughtful of visits to EY & limit traffic there.

Guardian work ongoing, so will be going up & down. (Jamie)

HVAC Work in OSB Labs slated for next week.  Need to cover optics tables for this work (Corey).

J. Kissel
FRS Ticket 9683

Checking in on the top-mass health of H1 SUS ITMY after the corner station volume has passed below 1e-5 Torr, I've gathered an high frequency, amplitude spectral density of the Main (M0) and Reaction (R0) chain top mass OSEMs. All sensors show a flat noise floor consistent with the expected self noise of the OSEM (~5e-11 / m.Hz^{-1/2} above 10 Hz)  and no longer show any significant comb-like features. 

As such, I recommend we close out the above cited FRS Ticket.

Noticed that the Observatory Mode for H1 was set to "Locking" this morning (, so I set it back to the nominal "Planned Engineering".  It was like this since Tuesday morning, so I'm assuming something was booted during Maintenance and took us back to this "default" setting.

So have about 45hrs of "Locking" which should be "Planned Engineering".

A new frequency overview screen is now available for the squeezer, see screenshot 1. The left side lists the relevant VC(X)Os, whereas the right side presents measured and calculated frequencies describing:

• The two AOM frequencies, the LO frequency, and the frequencies of the PSL/IMC VCO and the squeezer VCO. The AOM 1 frequency and the LO frequency are inputs that are set to 200.000 MHz and 3.125 MHz, respectively. The AOM 2 frequency is the readback of the VCXO frequency.
• The difference between the two AOMs, and its difference to the LO frequency. The later will be zero, once the CLF loop is locked.
• The difference between the beat frequency and double the squeezer VCO frequency. This will be close to zero , when the TTFSS is locked.
• The difference between the PSL/IMC VCO and the squeezer VCO. This describes the difference in laser frequencies between the main laser in the interferometer and the squeezer laser. This difference should be kept small, before the LO is engaged.

The squeezer VCO has available two error signals that can be used to tune its frequency using the tune offset (see screenshot 2):

1. Internal: The error signal is the difference between the measured VCO frequency and a given set frequency. The later is nominally 79.2 MHz, and identical to the PSL/IMC VCO.
2. External: This is the difference between the PSL/IMC VCO and the squeezer VCO. It can be used to keep the squeezer VCO at the same frequency as the PSL/IMC VCO.

The intended strategy during squeezer locking is:

1. Before the IMC is locked, the squeezer VCO should be kept near its nominal set frequency.
2. Once the IMC is locked, the squeezer VCO should follow the PSL/IMC frequency.
3. Once the the squeezer laser is locked to the OPO, the VCO tune input is no longer effective, and one has to use the OPO PZT position to keep the frequency of the squeezer laser close to the main laser.
4. After the LO loop is locked, the two lasers will be phase locked. This implies that both VCOs will be running at the same frequency.

The VCO capability to control its frequency has been copied over to the VCXO. The VCXO frequency can now be kept near a nominal set frequency, which is nominally 203.125 MHz. I doubt this feature is necessary, but it is there if needed.

TITLE: 03/01 Day Shift: 16:00-00:00 UTC (08:00-16:00 PST), all times posted in UTC
STATE of H1: Lock Acquisition
OUTGOING OPERATOR: None
CURRENT ENVIRONMENT:
    Wind: 12mph Gusts, 10mph 5min avg
    Primary useism: 0.06 μm/s
    Secondary useism: 0.41 μm/s
QUICK SUMMARY:

• CP4 bakeout starting this morning.
• VOPO/Squeezer work at HAM6 work continues
• Hardware Watchdog is RED at EY (known issue)

SQZ6 enclosure was moved south of HAM6. Cabling for table in the SQZ bay was moved to new table location. SQZ team will let us know if we missed a cable. Power and E-Stop cables still need to be terminated.

h1iscey front end glitched at 14:15 PST. We are holding off on its restart until we contact EY group.

After a lot of experimentation, I have found a way to improve the attenuation of frequencies below 9 Hz in the calibration by 1-2 orders of magnitude, without significantly increasing the computational cost or latency of the pipeline. Here is a list of what I've changed and what I've kept the same:

• In the actuation path, before the actuation filter, I have added a separate high-pass filter using Matlab's firls() function (documentation). firls() designs a linear-phase FIR filter that minimizes the weighted integrated squared error between the desired frequency response in specified intervals and the approximated response of the filter. The filter I've found most effective is designed to have a frequency response of 1 from 9 Hz to the Nyquist rate, and 0 from DC to 8 Hz.  I found improvement when I weighted the stop band more heavily by a factor of 100.  This filter was 2 seconds in length for the tests that I ran.
• I shortened the actuation filter itself from 6 seconds to 4 seconds. (The main reason we needed such a long filter before was to prevent spectral leakage, which is now taken care of by the previous high-pass filter.) It still contains half of a Hann window to the fourth power to roll off low frequencies, in order to further attenuate those frequencies. I found that this roll-off actually worked better than any version of least squares filter I tried in its place.
• In the inverse sensing filter, I replaced the half Hann window to the fourth power with Matlab's least squares filter using firls(). For some reason, this worked better in the inverse sensing filter than the Hann window stuff, even though it did not work better for the actuation. I designed this filter to have a frequency response of 0 from DC to 5 Hz (weighted by a factor of 100) and a frequency response of 1 from 9 Hz to the Nyquist rate. This is what seemed to produce the best results. I did not add a separate high-pass filter in the inverse sensing path, as it seemed unnecessary, and filtering at 16384 Hz gets expensive.

Of all the things I tried, this is what worked the best. Reasons I did not make this even better include:

• I haven't figured out a better way yet
• It is possible to do better with longer or more filters, but since we are doing this in the time domain with FIR filters, adding more filtering gets computationally expensive and adds latency. It is necessary to balance computational cost with quality, as we do not want the pipeline falling behind in low-latency or taking forever to produce offline data on the clusters.

Several plots are attached to show the new features. The first 5 plots are the frequency responses and comparisons to the ideal models for each of the filters used. The last 3 plots are comparisons of C01 data with data produced using the new filters. The attenuation is better by about 1-2 orders of magnitude, and there is just a very small amount of ripple added below 20 Hz.