Dave asked me to look at h1ecaty1 after reports of a glitch. Upon logging in I found what is in the attached screenshots. I restarted the IOC and also firefox. When the firefox update asked me if I would allow it to make changes to the computer I clicked no.
[Jenne, Hang, Gabriele]
more locking
tcs - HWSX &Y alignment CO2y up, CO2x needs aligned
psl - ISS power supply swapped worked, still looking at peaks maybe due to water line touching table, still need to do modematching
pem - accelerometers and microphones
sus - quad violin mode expansion, add isi ff to bs sus
asc - Hang model update
pr3/sr3 oplevs being resurrected
vac - ion pump at ex replacement, no engineered fall protection on clean room, needs laser safe
efm - electronics swap for more gain
wind fence instrumentation install at ex
supply fan maintenance starting tomorrow
digging outside at mid-x
communication survey sent out last week to assess issues with in-chamber data/communications still open
Power fluctuations most likely due to ISS work performed between 7/17 - 7/19. Chiller flows are on a marginal downward trend.
Laser Status:
SysStat is good
Front End Power is -0.006326W (should be around 30 W)
HPO Output Power is 0.2797W
Front End Watch is GREEN
HPO Watch is RED
PMC:
It has been locked 4 days, 0 hr 58 minutes (should be days/weeks)
Reflected power = 18.65Watts
Transmitted power = 46.11Watts
PowerSum = 64.75Watts.
FSS:
It has been locked for 0 days 9 hr and 11 min (should be days/weeks)
TPD[V] = 2.572V (min 0.9V)
ISS:
The diffracted power is around 0.41% (should be 3-5%)
Last saturation event was 4 days 0 hours and 46 minutes ago (should be days/weeks)
Possible Issues:
Front End Power is Low
PMC reflected power is high
ISS diffracted power is Low
LRA out of range, see SYSSTAT.adl
Summary:
For a cold interferometer with no self heating (ie IFO power) onto the surfaces, the TCS settings are as follows. What is being used for the calculations come from the Nebula page and the Finesse aLIGO parameter files and the contrast defect is calculated using the amplitude overlap described here. Depending on the various configurations of the interferometer, the amount of calculated contrast defect will be different:
| Configuration | Required CO2 Power(Watts) |
Required RH Power(Watts) |
Calculated Static Contrast Defect (PPM) | Description |
| Xarm and Yarm Locked | 0.0 | 0.269 on ITMY | 3.86 | Propagate the arm modes to the beam splitter and measure the mode overlap, this will not see the CO2 lensing and is dominated by the ITM+ETM radii of curvature. |
| PRX,PRY, IMC Locked | 0.094 on ITMX (Annular) | 0.0 | 186 | After turning on the ring heaters to adjust the ring heaters for the X-arm and Y-arm modes. Lock PRX and PRY individually and take the overlap between the two modes at the ITM HR surfaces. This is sensitive to both the thermal lensing of the ring heaters and the CO2 laser, however, we'll want to compensate the ITM RH configuration for sideband. |
| Simple Michelson, only IMC locked |
0.100 on ITMY (Annular) |
0.0 | 160 | Propagate the IMC mode towards the ITMs and take their overlap at the HR surface, because this is not resonant, the overlap will be proportional to just the difference in ITM radii of curvature. However, this overlap is sensitive to the substrate thermal lensing, but we won't be using this configuration anyway. |
Details:
Generally, the contrast defect is ratio of the power at the AS port to the REFL port when locked on a dark fringe, in other words, it's the amount of junk light that is present in the interferometer.
This junk light is a symptom of a few different causes, for example, an imbalance of reflectivities between the ITMX/ITMY will cause non-perfect destructive interference at the AS port and AS-AIR camera will see a round beam. Another cause of contrast defect could be misalignment between the ITMs+BS which will result in seeing a 01/10 higher order mode. But if both of those causes are fixed with stringent designs specs and alignment loops closed, then what is left over will be dominated by mode mismatch which can be fixed by a combination of RH and CO2 lasers. Because this model doesn't take into account the self heating, these numbers will have to be adjusted based off the amount of optical absorption on the HR surfaces as well as the substrate.
The power required for the lensing is calculated using the TCS actuator calibrations:
Where 
is the calibration factor in microdiopters/watt.
It seems, according to alog-43005, that the radii of curvature reported by the nebula page might not be correct since most of the phase change seen by the beam (both sb and carrier) in the simple michelson configuration is due to the ITM ROCs.
Craig, Sheila, Matt, Lisa
We went through an inital alignment and locked PRMI. The build ups are about a factor of 3 lower than what we had in May, so we tried walking the alignment around a lot.
For the initial alignment we had to do a few things:
Once we saw that our build up was not good, we decided to try to bring the beam back onto the POP QPDs (We think that this path was not modified during the vent, we used to be on the QPDs and now we are too high). We moved IM4 to center in pitch on POPA, and then moved the X arm to get the green and red co-aligned. This resulted in worse build ups in PRMI.
We can see something on the PR2 camera, which could be scatter but we can't be sure what it is. We attempted to minimize the camera sum after Matt adjusted the mask to zoom in on these spots, but this also resulted in lower build ups.
We started a VCO scan in single bounce with the OMC locked at 1216361323+60
This is for a whistle investigation. It will run for 12 hours. If you want to stop it from moving the IMC VCO in the morning, you can quit the matlab script running on opsws13
Edit: It looks like this ran from about 6:11 UTC July 22nd to about 15:50 UTC.
For future reference, we had a PRMI lock with the beam on the POPA QPD at 5:38:57 UTC on July 23 2018
Matt, Lisa, Craig Previously I had made some in-loop green frequency noise measurements of ALS_COMM and ALS_DIFF. Today we made some out-of-loop IR frequency noise measurements. First we locked green to the Xarm in LOCKED_SLOW_W_ETM_WFS, then IR to the Xarm via the XARM_IR_LOCKED ALIGN_IFO guardian state, so that IR had control of MC2. Then we measured the frequency noise of transmitted green light beating with frequency-doubled IR on ALS_COMM using the COMM PLL. The COMM PLL is calibrated into micrometers, so I calibrated it back into IR Hz and plotted it below. X-arm IR Frequency Noise RMS = 12.4 Hz = 180 pm We have some concerns about this measurement, in particular, we are missing some peaks present in Alexa's measurement of the same frequency noise (Figure 6.1 of her thesis). LHO alog 10511 gives their old method of estimating IR frequency noise. Our measurement scheme today is closer to LLO alog 11265, only we locked using POP_A_45 instead of AS45.
At 22:00 PDT the cell phone alarm system crashed when trying to send an hourly keep-alive email. Systemd restarted the program 30 seconds later, at which time it was able to send emails.
I've opened FRS ticket 11133 to cover this.
Matt, Lisa, Danny, Craig, Hang
=====================================
In the ALIGN_IFO guardian, we created two new states called PREP_MICH_DARK_ALIGN_WFS and MICH_DARK_ALIGN_WFS. These states actively control the MICH alignment when locked with the dark fringe, using WFS signal (AS_B_RF45_I for both pitch and yaw) instead of dithering. It was designed to be used for the contrast defect measurement in the simple mich configuration, as it should minimize the contamination due to misalignment.
We first used this state to measure the contrast defect, and got a large contrast of 3%. See the first image. A 02 mode was clearly visible. Then Danny noticed that the CO2 Y heating was left on (1 W central heating) during this measurement.
We then turned off the central heating and instead put 1 W (~3.8 W delivered) of annular heating on Y. After a while we again locked the MICH in dark with angular control on using the WFS. The dark fringe got improved significantly. We measured ~ 18 ct in AS_A_NSUM the dark state, while in the bright fringe we had ~ 10000 ct. This leads to ~0.2% contrast defect in the simple MICH config. See the second plot for the dark fringe residual beam (note that we modified the exposure time to max in the second image). Also note that we did not follow the entire heating process and we might overshoot in the annular heating. The contrast defect could be further improved.
======================================
We also did some alignment in the corner station.
For SR2, we looked at AS_A_NSUM in single bounce and scanned SR2 in both pitch and yaw. We recorded the edge values where we start to see drops in AS_A_NSUM, and then put SR2 alignment to the middle of those edge values.
For IM4/PR2, we used the INPUT_ALIGN guardian state to align them to XARM. It seemed that guardian state of misaligning ITMY needed to be updated as ITMY beam still showed up in the AS PD even after we requested it to be misaligned. In order for the input ASC loops to work, we needed to manually misalign ITMY by >30 urad in yaw.
The PRX/SRY guardian states did not work properly currently. We could lock PRX in length but the alignment part didn't work. For SRY even the beam looked roundish in the camera, we could not acquire it to be locked in length.
To clarify, Hang tells me that they were requesting 1W with the CO2 rotation stage, which was actually giving them 3.8W injected into the vacuum.
Also, this level of annular heating was temporary, to compensate for the fact that they had not waited for the central heating to have fully cooled off.
Craig, Georgia, Jenne, Matt, Lisa, Keita
We found this afternoon that our problem with ALS glitches is actually due to a frequency crossing of the DIFF VCO. We were able to reproduce the glitches with the Y arm blocked by moving around the control voltage out of the DIFF PLL. Removing the DIFF FDD has no impact.
We have locked ALS twice by first locking ALS_DIFF, then locking COMM. (It used to happen in the other order.)
Hang, Danny, TJ
Yesterday we found that there was a range of values for SR3 that we could see both ALSX and Y beams on the HWS table (42981). Today Hang helped us get SR3 in that window while compensating with SR2 to to keep AS_A and AS_B on pds with good sum. Danny and I went to the table to double check that we now had two clean beams on table, but were only able to get ALSX on the HWSX camera. For HWSY, the SLED is not aligned to the irises, even though it was working before we did this SR3 move...odd. We will have to do a bit more aligning to make sure that the ALSY beam coming onto the table is not clipping.
Current SR3 opticalign values: P=423.5, Y=-152.0
SR2: P=1030.7, Y=378.5
If you run out of options, changing the HAM4 ISI positional and rotational biases might be another knob to turn in order to get both ALS beams onto the HWS table as well as no clipping in the SRC. . Although it’s not clear from the in-chamber Hartmann paths which direction you'll want to move.
One of the things that was noticed on ISCT1 is that some of our beam combining beam splitters have beams that are hitting the back side of their mounts rather than the beam dumps behind the mount. In particular, I think Keita said that ALS_BS6 on D1201103 had a problem. I don't know if Sheila and Georgia noted that also ALS_BS3 on that diagram has a problem currently, or just potentially.
One way to solve this problem is utilize a mount that gives clear access to all 4 ports of the beam splitter. It sounds like BS6 currently has a New Focus 9809 mount which should suffice, but we still have some clipping problems there. BS3 just has a standard mount, so definitely needs something. I have had success at the 40m ALS beam path using Thorlabs H45s. I also note that we have 2 New Focus 9925 mounts in the optics lab. I'm not sure if these are made any more, since I can't find a website for them, but they would work too, and we already have them, so we can swap one into the BS3 mount. I attach a photo from my phone of this last mount.
(Mostly writing this down so I can remember the part numbers, since I struggled to remember what the 45deg mounts from Thorlabs were called.)
Sheila, Craig, Hang After Georgia and Sheila locked ALS COMM this morning, we proceeded to lock ALS DIFF. Our first attempts to lock failed as the ETMY ESD linearization was off. When we switch it on ALS DIFF locking worked. We were also able to get the ISC_LOCK guardian to reach LOCKING_ALS. The plot shows the spectrum of H1:ALS-C_DIFF_PLL_CTRL_OUT_DQ, which is already calibrated into micrometers just like ALS COMM. ALS DIFF RMS = 6.8 Hz = 48 pm
There is a more than 3 orders of magnitude difference below 0.1Hz. How?
Blue is an old reference with the ALS DIFF not locked, shouldn't have included it. A better comparison is in Alexa's thesis, figure 6.8, where she cites 40 pm RMS in loop noise, or 5.6 Hz RMS. Her plot is at the error point of the PLL, while ours is at the control point, so this isn't a direct comparison. I would make a direct comparison if I knew the TF from DIFF PLL ERR to CTRL, but the medm screen is broken and only the gods of interferometry know where the foton files live.
Congrats, team!
Lisa Barsotti showed me an old lock of ALS_DIFF made by Kiwamu and Sheila in the pre-O1 days. Their ALS_DIFF frequency noise was 60 pm RMS, or 8.5 Hz RMS. Our number is not much better, even after our optics replacement significantly increased the finesse of the arms. Looking at Kiwamu's first attachment in the linked alog compared with ours, most of the RMS came from high frequency. Our high frequency RMS seems to be around two times lower, so the new arms are doing something. Low frequency noise hits us harder, yielding a similar RMS.
