Sheila let me know that they had some trouble last night with reducing the RF45 modulation depth at 2W.  Things like PR gain became noisy, and POP90 increases, but things got better as they powered up. 

The reason, back in fall 2016, that we had been doing RF45 before increasing power is that we had (at that time) been going straight to 50W in the IncreasePower state, and we saturated the AS WFS (alog 29750, bullet point 3) once at 50W.  Also, the AS port trigger PD that is used for the fast shutter was close to saturation (alog 28909). 

So, now that we have split out going to 20W versus going to higher powers (right now 30W), we can move the 45MHz modulation depth reduction in between. 

I have done this, but since the state numbers were only 1 apart, to make the order correct on the states "ladder" screen, I have swapped numbers 504 and 505.  State number 504 was the RF45 reduction, and state 505 was IncreasePower.  Now, 504 is IncreasePower and 505 is reduce_rf45_modulation_depth.

After injecting some squeezing in the IFO and saw nothing we thought it's a good idea to go back to the HD and measure squeezing with PSL LO. Sadly due to another 30% pump transmission drop (alog46290) the max non linear gain we could get is 2. The phase noise measured from LO error signal (in loop) suggested 3 mrad. Fringe visibility was 98.5% (I used 98% for the model below). 

The measured loss in the path can be found in Haocun's github here. The calculated total efficiency is 0.8147. This measurement suggested we have extra loss up to ~20% that was unaccounted for.

The first measurement Haocun took after the vent can be found here (we locked LO with Mephisto LO at the time). We still had intensity noise issue at the time so the data points were all over the place because the nlg was unstable. I'm not sure if the extra loss has always been there or degraded over time after the vent.

Today while locking ALS we had fast glitches in both x and y arms which caused us to lose lock. I had a quick look at the same channels I looked at when we had extra intensity noise on the ALSX laser back in September.

The glitch in the x arm coincided with a jump in a few channels in the laser: the pump diode 2 power, the noise eater, and the IR anf green output powers. We lost lock a few seconds after the glitch. First attached plot shows the glitch in green arm transmission (bottom middle), and other laser-related channels. 

The y arm has glitches much more frequently than the xarm, their frequency and severity comes and goes. Looking at the same channels at the y-end during a period of frequent glitching (second attached plot) shows nothing obvious that is coincident with the glitches. There are some noteworthy things: the RMS noise on the green and IR powers out of the laser is significantly worse for ALSY, compared to X. And there are also frequency glitches in the pump diode 1 noise monitor.

Conclusion: the als-x laser maybe has some kind of intermittent mode-hopping. It's unclear from these plots what is going on with the y-arm.

Laura Nutall, Georgia, Craig, Sheila, Stefan, Thomas, Danny

Laura pointed out this alog from Valera 42651, based on this we decided to try switching the upper stages of DARM control from ETMY to ETMX.  The attached script should do the switch, some of these things need to be added to the guardian before we try to use the use_ETMX flag to go to ETMX.  The attached screenshot (and template) show the measurement from L2 excitation to DARM for both test masses once we set up the actuation path.  We haven't seen fringe wrapping since making the switch at 6:30 UTC.  Craig has been updating the calibration, which is why our range has been changing. 

Locking notes from earlier today:

• We had difficulty with engaging ALS DIFF when the sensor correction gain was wrong.  The clue that something was wrong was that although our ground BLRMS indicate that the ground motion was under 1um/sec rms, by counting fringes we could see that the differential motion of the Y arm was about 3 um/sec rms peak.  With the correct sensor correction Georgia counted  about 1.5 um/sec peak velocity by letting the green arms swing. Once Hugh and Jeff fixed the sensor correction, 46315, we were able to lock ALS DIFF without any problems, so it seems like this 2 Hz oscillation during the engagement is dependent on the input motion. 
• We had some problems with the ASC engagement again. When the dither loops came on, the other loops were unable to follow and we lost lock once.  (We don't have soft limiters for the dithers, and we reduced the CHARD gain so that the dithers aren't covered by the CHARD noise in DARM).  We increased the CHARD P gain by 16 dB during the engagement to match the gain of CHARD Y, and changed the +46dB filter to +30dB so that the final gain is the same (we also adjusted the gain in LOWNOISE_ASC to have the same overall gain for CHARD P).  We also increased the gain of SRC1Y during the dither loop engagement by 20dB, but weren't able to do the same for SRC1P.  If we continue to have difficulties with the dither engagement we can try increasing the gain of SRC1P (by less than 20dB) and both SRC2 loops, or we could add soft limiters to the dither loops. 
• We lost lock due to an ~3Hz instability of the soft P loops a couple of times at the end of the ENGAGE_SOFT_LOOPS state.  We decreased the gain of CSOFT P by a factor of 2 and decreased the gain of DSOFT P by 25%, by commenting out the final gain increases for these in the guardian. 
• Once we were at 30W Georgia moved the A2L gains back to what they were last week, and we didn't see any dPd theta instability, so it seems that these gains are fine. 
• We reduced some of the wait times that we had for waiting for POP18 to decay during the power up.  We also removed some of the logic that was stepping the PRCL gain as we increased power, because we noticed that the PRCL gain was too high and we had gain peaking.  We could further simplify this process. 

J. Kissel

As mentioned yesterday/Monday, I fixed a bug that Shivaraj and Saravanan found in the time-dependent correction factor calculations for the optical plant (see LHO aLOG 46260). Woo!
We managed to recover the IFO and hit nominal low noise this evening for about 30 minutes while I was present. Woo!
I was looking around the new system for success, and for about 10 seconds prior to the lock loss, it was all "working" roughly as expected. This time period was surrounding Jan 10 2019 at 01:15:54 UTC -- see attached screenshot of kappa_C, f_C, f_S, and Q_S -- the optical gain, the DARM coupled cavity pole frequency, the detuned signal recycling cavity's optical spring frequency, and its Q.

There still things left to be fixed before it "works" enough to start considering applying these to DELTAL_EXTERNAL:
(1) Obviously, the signs are wrong (negative) on the optical gain (kappa_C) and spring Q (Q_S). This is a problem remaining with the signs of the reproduction of the various reference model parameters (aka "EPICs Records"). These have been created and installed using new(-ish) python code of which all the sign errors have not yet fully shaken out when replicating the current, mixed DARM actuation scheme. Evan and team are still actively working on this (see G1900019).

(2) The gating system (which smoothes out the TDCF for application to DELTAL_EXTERNAL) and override system (which allows us to bypass non-functional parts) don't yet have all settings turned ON in such a way that the relative actuator strength changes -- a.k.a. kappa_UIM, kappa_PUM, and kappa_TST -- are spitting out live values. In this brief moment in time where the optical plant parameters were working, all of the actuator kappas were bypassed and spitting out +1.0.

(3) We're still using a pre-ER13 measurement to determine the DARM loop reference model, which in turn determines the EPICs records, which determine the calculated TDCFs. The actuator measurements were quick and dirty. The IFO was still at 20 watts. Thus -- when I see that the corrective gains, the kappas for both actuator and optical plant, have magnitudes at the 10-30% level, I'm not yet sad. Now that we're electing to stick at 30W input for a while, while we hunt for noise, we should re-take -- or in some cases take for the first time -- quality measurements that will get the system to levels needed for ER14 / O3. 

However, with all the above qualifications -- we made good progress today. We can comfortably (albeit imprecisely) say that the optical gain needs ~10% correction, the DARM coupled cavity pole is at ~420 Hz, the optical spring is at ~15-20 Hz, and the Q is very close to critical at ~1.

Attached below are images of the HR surfaces of the H1 ITMs with IR light resonating in the arms (19.7 W of 1064 nm laser light).

They were captured by Jeff Bartlett a few weeks ago using the "Pcal" ITM camera systems mounted on the spools at the ends of the beam tube manifolds in the LVEA, near the cryopumps.

The camera settings were:

ITM-X
#109 ISO 800, shutter 1/800
#110 ISO 800, shutter 1/400
 
ITM-Y
#016 ISO 400, shutter 1/500
#110 ISO 250, shutter 1/200

The assembly drawing for the camera system is LIGO-D1500073-v9.

Kentaro Mogushi's technical note regarding camera system upgrades to minimize aberrations introduced by the vacuum windows  is LIGO-T1700540-v3.

Note that these are the first IR images after implementing Kentaro's upgrades.  Image quality may improve slightly as the systems are adjusted.

J. Driggers, S. Dwyer, J. Kissel, P. Thomas

Prior to yesterday's rebooting of most front-end machines (LHO aLOG 46288), Patrick, Jenne, Sheila, and I went through the SDF system and reconciled differences. I attach the highlights of what we accepted (into the SAFE.snap files -- OBSERVE.snap files were NOT reconciled). The highlights come mostly from the ISC systems that are infrequently maintained. This may (or may not) be useful as (a) a recovery aide today, and (b) a record / indication of the progress / changes in the global control system over the past few weeks and the holidays -- especially as some of the team responsible for the changes have left the site for good.


h1asc model:
- Accepted changes to the TR X/Y and A/B QPD to DHARD and DC5 to elements of the ASC input matrix. This reflects the changes made by Hang before the break to implement the radiation pressure compensation (RPC) stystem (e.g. LHO aLOG 43849). These elements of matrix were last changed in mid-December (e.g. Dec 14, 23, 26 , and 28)
    - Several other Radiation Pressure Control TRAMPs were accepted as is.

- Accepted the reduction of the CHARD Y A gain from 1.4 to 1.35, and the turning OFF of the DHARD Y B  loops by setting the gain to 0.0. This is something we don't necessarily want to keep, but was part of the spot position exploration vs. TMS drift on Monday night.

- Accepted the CHARD Y to ITMX element of the ASC output matrix at 0.87 (where it had been saved [accidentally] in December as 0.93, which everyone agreed is very wrong).

- Accepted ASC_AS_A_RF45 I and Q individual segment filter gains as 1.0 instead of 1.413. These had accidentally been saved in the SAFE.snap because two weeks ago, we were reconciling SDFs in a state just before nominal low noise. That means we were *past* the state that reduces the RF45 modulation depth, and compensates that optical gain change by increasing the gain of the AS_A_RF45 sensor from 1.0 to 1.413, which we *do* want when we're in nominal low noise -- i.e. when we reconcile against the OBSERVE.snap -- but now when we're DOWN, or in low acquisition states -- which is what we want after computer reboots. Thus ACCEPT as 1.0.

- Accepted any and all ASC-ADS (alignment dither system) changes in filters, gains, limiters, tramps, oscillator frequencies, etc as is. These are all a part of the new dither alignment system mentioned above. Most were changed on Dec 21 with the system was installed (LHO aLOG 46136).

h1ascimc model:
- Accepted MC2 TRANS QPD SEG[1-4] FM9 filter = "+9dB" gain as OFF. This filter was originally installed in concert with a -9dB analog whitening gain in mid-December when Sheila found that we had saturated the QPD's ADC before a high-power lock loss on Dec 21 2018 LHO aLOG 46084. However, whether it be from settings loss or intentional, the analog whitening gain was restored to "nominal" +36 dB gain a day later, and then the compensating gain filter was turned OFF a few days after that. Somehow, in those three days, though, someone saved with compensating filter as ON in the SDF file. Since the analog whitening gain is no back to regularly being at +36 dB, there is no longer a need for this +9dB compensation filter to be on. It's been accepted as OFF.

h1lsc model:
- Accepted changes to the LSC output matrix elements for OSC 1 2 3 to PRM, SRM, and BS matrix as now regularly ON (with whatever values were present). This is the part of the new length dither line system to constantly monitor the length loop gains during power up. LHO aLOG 46082

- Reverted LSC PD DOF matrix (the DRMI 3f input sensing matrix) elements as ZERO. The safe.snap file correctly indicates that the DRMI 3f matrix elements should be zero when the IFO is down. The current EPICs values were non-zero because the last lock loss was during a DRMI 3f lock, and the elements hadn't been reset.

- Reverted the LSC TR X NORM OFFSET from -0.36 to 10. This offset (as far as we can tell) isn't used, but it's been 10.0 for a very long time prior to its change on Dec 21. Also the LSC TR Y NORM OFFSET is also 10. No findable aLOG, so we reverted.

h1lscaux model:
- blindly accepted all ~110 differences.

h1omc model:
- Accepted  the change in filtering of the OMC-LSC_PD_IN bank -- the first filter bank on the way to the OMC length control that band-passes the OMC DCPDs. The OMC length dither line band pass was reduced from "BP4100" in FM5 = [4000 4200] Hz to "nBP4100" in FM7 = [4050 4150] Hz on Dec 31 -- see LHO aLOG 46197. 

h1susbs model:
- Accepted FM10 = "BounceRoll" in the BS M2 LOCK L filter bank as OFF. This is the Beam Splitter's Bounc and Roll mode notch. It's unclear why it's off, but it's been this way for quite some time, so I accept it.

h1susitmpi model:
- Reverted all UPCONV output matrix elements to 0.0. This matrix last had values changed and populated while Sebastien and Terra were here getting Qs of test mass mechanical modes in early December 2018 (LHO aLOG 45812), and are no longer needed. 

- Accepted all oscillator frequencies.

Generic Things Across Models:
"QPD Offsets"
    - Any and all QPD and WFS OFFSETs have been accepted as is in h1asc, h1ascimc, h1omc, h1alsex, h1alsex, h1iscex, and h1iscey. As we're continuing to explore and find the optimal spot position for the interferometer -- especially after we've switched to an ETM alignment dither system (see e.g. LHO aLOG 46272) -- these will be occasionally updated.

"A2L gains"
    - these DRIVEALIGN P2L and Y2L gains across many globally controlled suspensions (like h1susetmx, h1susetmy, itmx, itmy, even h1susmc2) are coupled with the above mentioned exploration of spot positions. Jenne says, theoretically, with the new ASC dither system, we should no longer be regularly updating these, so we've accepted them as is.

"Not28Hz" filters
    - in h1asc model, accepted DC 1 and 2 "Not28Hz" filters (FM9 for P and FM5 for Y) as ON. The crowd didn't have a satisfying answer for this resonant feature that intermittantly gives us problems, moves around in frequency, and/or may be solved -- according to a Dec 19th LHO aLOG 46084 after some commissioning of ASC loops. 
    - in h1susbs and h1susetmx models, accepted 28.75 Hz filters as OFF.

We should probably go on a campaign of removing these filters now that the problem has been solved. 

During yesterday's maintenance window I worked on getting the DBB up and running.  I implemented the mode matching solution shown in the first attachment and proceeded to align the DBB.  That completed I then tweaked the mode matching lens positions to optimize the mode matching to the DBB PMC.  I ended up moving the 1st mode matching lens 2.5mm towards the 70W amplifier, and the 2nd mode matching lens 11.2mm away from the 70W amp.  At this point the maintenance window was running short and further position tweaks only made the mode matching worse so I stopped here.  A single FSR scan of the DBB PMC is attached to show where I left things (see 2nd attachment).  The mode matching isn't quite as good as the last solution (implemented before the changes during the Nov vent, which rendered that solution invalid), the 20+02 peaks are higher but more equal than previously; it's possible some more lens position tweaking could improve this (will be explored during a future maintenance window).

I then tried to lock the DBB PMC, but once again it was unable to lock.  I looked into the locking threshold for the DBB PMC and then noticed something that I hadn't when I last worked on the DBB: the DBB PMC transmitted photodiode (TPD) was not registering a voltage on MEDM.  This voltage is used as the locking threshold for the DBB PMC, and it consistently read zero.  I hooked up a multimeter individually to each port of the DBB PMC TPD (0dB, 40dB, and 80dB) and all 3 were indeed outputting a voltage.  Using the DBB's Manual mode, I manually moved the DBB PMC PZT and watched the voltage from the 0dB port change accordingly: as the DBB approached resonance the TPD voltage increased, as it moved away from resonance it decreased (all expected behavior).  Plugging the DBB cables back in and repeating the same experiment showed absolutely no change on the MEDM screen.  To get a more full picture, during the next maintenance window I will repeat this for the 40dB and 80dB ports of the DBB TPD (as I only used the 0dB port due to time constraints), although this likely explains why the DBB PMC won't lock; its resonance condition can never be met if the readout is not functioning correctly.  We will start hunting this down to get it fixed.

As a last item, I did a quick 1 minute duration Mode Scan (since it doesn't involve a locked DBB PMC) just to see what the higher-order mode count of the 70W beam is (as read by the DBB PMC).  I've attached the resulting measurement (3rd attachment).  It claims the 70W beam's higher-order mode content is 41.1% +- 9.666%.  This cannot be right as at last measurement we had a PSL PMC visibility of ~89%; I also checked the PMC after seeing this result and it was transmitting ~55W and reflecting ~11W (83% power throughput, which would be impossible if the higher-order mode content of the beam was really 41%).  So something is still off here, be it alignment, mode matching, or something else.  I will try to tweak up the mode matching and alignment to see if I can get the mode matching better (at least close to the previous solution) and see how this changes the Mode Scan measurement; if this doesn't work, a new mode matching solution will be explored.  Should that fail, then we may be in the "something else is wrong here" category.  It should be noted that LLO is experiencing a similar issue with their DBB; we're currently in communication with AEI to see if we can get this figured out.

FAMIS 8086

The following appear elevated:

ETMY_ST1_CPSINF_H2
ETMY_ST2_CPSINF_V2

BS_ST1_CPSINF_H1_I
BS_ST2_CPSINF_V1_I

ETMX_ST1_CPSINF_H2

ITMX_ST2_CPSINF_H3

ITMY_ST1_CPSINF_V2
ITMY_ST2_CPSINF_V3

15:40 UTC Vanessa to end X. Karen to LVEA.
16:29 UTC Karen back
17:12 UTC Gerardo and Filiberto to X-28 and Y-28 and mid station
17:14 UTC Nutsinee to ISCT6
17:24 UTC Ed to LVEA to reset HAM6 dust monitor
17:50 UTC Kyle back and forth between mid X and mid Y
18:06 UTC Chris opening rollup door, then to mid X
18:27 UTC Gerardo and Filiberto back
21:44 UTC Keita and Daniel to LVEA, work on ISS second loop
22:01 UTC Travis to end Y
22:03 UTC PSL tripped off, rack work
22:12 UTC Keita and Daniel back
22:23 UTC Travis to mid X
22:34 UTC Travis back
23:00 UTC Site weekly meeting

FAMIS 10491

Added 50 mL H2O to the H1 PSL crystal chiller.

The canister filters are white and free of debris.

There are no faults on the diode chiller.

• Trying out TJ's new initial alignment guardian.
  • Set some thresholds for green initial alignment (very slight mod on the ones TJ already had), so now green initial alignment can auto-offload.
  • Checks for returning true probably need to check that the managed guardian arrived at the correct state, as well as is done.  Looks like the manager guardian is happy and moves on whenever the sub-guardian is done with any state, including any that are on the way to the requested one.
  • We probably need to write a more sophisticated convergence checker, so that it looks at some running average, to see if the signals are moving about zero, rather than setting a threshold.  When the ground is noisy, the convergence checker with tight thresholds won't say ok, but if you loosen the thresholds then you won't actually be getting a fully aligned system.
    • In the ALIGN_IFO guardian, MICH_DARK_ALIGN state, the convergence checker has been commented out for this reason.  So, the new all-in-one alignment guardian won't really work for the MICH state until we get some kind of convergence checker that works.
  • As often happens, the SRC had trouble locking until I used SR2 (with SRM misaligned) to center on AS_C.  We will likely need to write a guardian state on the way to SRC align that does this for us.
• I had done most of my initial alignment in useism_mod_wind, which seemed better than windy.  But, the wind is continuing to go down, and SRC align keeps losing lock (SRY by itself is fine, but with the ASC is not fine), so I've just changed to useism for my ISI sensor configuration.
  • Nope, still bad.  Going back to regular "windy", since that's what is on Jim's recommended table, given that the microseism is at 0.8 um/s, not all the way up to 1.5 um/s.
• Looking at Yarm, ITMY moves much more than ETMY (alog 46313).
• PSL rack work on ISS diagnostics caused flow sensor to trip and close the external shutter.
  • Reminder to self on how to open the shutter:  Sitemap -> PSL -> Laser -> gray box near bottom of screen, just to left of middle, with title External Shutter Request -> RESET button next to the word Flow (top of the 2 in the gray box) -> Click RESET on the small screen that pops up.
• Added logic to CHECK_MICH_FRINGES in ISC_LOCK guardian to turn off the LSC-MICH1 gain if the BS is hitting its limiter.  This happens (infrequently, but can and does occasionally happen) when BS is too far off for MICH to catch lock, as Sheila noted in alog 46302.

Following up on alog 46240, we made the following mods to the ISS Second Loop, S/N S1700062, which is now installed again (PCB D1600298):

• Connect feedback of open OpAmp U2B to negative input and ground the positive input.
• Increase the gain of the sum by increasing R32 to 10K (from 1K).
• Swapped the corresponding OpAmp U3 from OP27 to OP37.
• Decrease the gain by 10 in the readback of ERR2 after the sum by changing R62 in the SUM signal block from 4.53K to 45.3K.
• Decreased the analog excitation input gain by 10 by increasing R33 to 100K (from 10K).
• Decreased the digitla excitation input gain by 10 by increasing R77 to 100K (from 10K).

To account for the increased analog gain we reduced the VGA gain to 0dB (from +20dB). We increased the REFERENCE_IN_MTRX_1_2 gain to 10 (from 1) and the REFERENCE_IN_MTRX_1_4 gain to 0.5 (from 0.05). The reference offset was also adjusted to 537 (from -35.14). At first glance, the output offset fluctuations are significantly reduced.

[PatrickT, Jenne]

We've struggled to keep Yarm ALS locked this morning since we have particularly bad ground motion today.  We started down a path of trying to figure out why.

The Yarm SUSPOINT length projection from the seismic tables doesn't seem very coherent with the power fluctuations in the Yarm.  This led us to look at the angular motion of the optics. 

In the attached ndscope figure, you can see that the oplevs for ITMY and ETMY, pitch and yaw are on the same y-axis range of 0.8 urad.  However, ITMY pit is moving more than this scale.  (Stefan has some plots up that show that a similar phenomenon is happening on the Xarm).

We took a measurement of the ITMY pit oplev loop, with a thought toward using the oplev to damp this motion.  However, before we go further down this path, we're looking with Sheila and Stefan if there are any inconsistencies in the seismic isolation or suspension top stage damping that may have potentially come in during some maintenance or computer reboot time.

Chandra R., Gerardo M., Kyle R.

Today we removed the existing (iLIGO, circa 1998) 2500 l/s ion pump and replaced it with another iLIGO one which had been rebuilt.  GV14 and GV15 were hard-closed throughout the day to facilitate this.

This exercise provided an interesting data point.  The pristinely baked, dry etc... new ion pump internals/interior, though having been vented with dry N2 this morning to prepare it for installation, was, in fact, exposed to the higher-than-normal moist air (raining today) for a few hours while the pump exchanging was taking place.  This exposing is required and purging while being exposed for this short time frame isn't typically needed.  However, today, upon pumping back down, starting the pump and combining it with the X-mid volume, I was surprised at how "wet" the resulting pump had become as a result of this brief room-air exposure.  The resulting pressure was abnormally high and I was hesitant to expose X1 and X2 to this water load.  The interesting part is how rapidly this "water source" was depleted.  What this shows is that the water accumulation on the vacuum baked surfaces while exposed to the wet room air hadn't had enough time to permeate deeply into the surface and was, thus, easily and mostly removed in a short time period upon re-exposure to vacuum - fascinating!

TVo and I went on a leak hunt and went through the table, lines, and then opened the chiller. Opening chiller revealed a puddle under the pump (see attached picture). It looks like we have a leaky pump shaft seal. Filed WP8036 and went to swap the chiller with the spare. The spare chiller would give us a "Low Fixed Pressure" warning and would then shut down. The low threshold was set to 3.0psi and when starting the chiller it would only read around 1.4psi. No flow was ever observed while we tried to turn this on. We eventually gave up and swapped the leaky one back in.

Back in April, Greg Grabeel fixed what is currently our spare from an issue that seems to be exactly the same (alog41314). He ran the chiller in a closed loop for a week back then, so I imagine that it should work now as well. Perhaps the issue we had today was just operator error. We will look into getting parts to fix the leaky seal, and how to get the spare running again.

Helium leak tested nipple-to-GV 16.5" joint as well as nipple welds with background of 2 x 10-9 torr*L/sec -> OK.  Currently there is a zero-length reducer with a 1 1/2" valve installed in place of the nominal 2500 L/sec ion pump which has been removed and is being refurbished.