All BSC spectra looks to be nominal except for a slightly elevated offset in the BS_ST1_CPSINF_H1_I channel. All HAM spectra appears to be nominally sound.
It was not necessary to add water at this time.
TITLE: 09/12 Day Shift: 15:00-23:00 UTC (08:00-16:00 PST), all times posted in UTC
STATE of H1: Commissioning
OUTGOING OPERATOR: None
CURRENT ENVIRONMENT:
Wind: 5mph Gusts, 3mph 5min avg
Primary useism: 0.02 μm/s
Secondary useism: 0.10 μm/s
QUICK SUMMARY: Safety committee meeting today
Hang Sheila Georgia
Attached are 3 ASC measurements:
1) CHARD Y with the gain set to 0.3, which is the same gain as in the reference. You can see that the phase delay at 1.8 Hz is gone.
2) CHARD Y with the gain set to 1, we increased the gain to 2 before powering up but didn't remeasure.
3) CHARD P with the 46dB of additional gain
[Nutsinee, Haocun]
After putting ND filters before the OPO REFL due to the saturation issue, the OPO can be locked without problems, and the I and Q phases look good. (Nutsinee can give more details on this.)
Later on we took more squeezing measurements under this good locking condition, and I plot the data with some fitting curves as attached.
Homodyne Visibility: 97% and 96.87% (We had to do this, and before any adjusting it was ~86.7% with ~2mW unbalancing.)
So the estimated optical efficiency is 65% and 50mrad phase noise.
According to our budget here, our optical efficiency should be around 85%, which means we need to find out where are these ~20% loss from.
Next step:
- Swap PD or tweak electronics to have proper set-up for the saturation problem, instead of using ND filters.
- We'd better to adjust the crystal position again with OPO locking properly.
- Measure the optical loss on SQZT6, and dig out other potential losses.
Just re-checked the quantum efficiency of both diodes, which are pretty good. Very close to 1. The back reflection are ~10uW out of 1mW input.
However, found out that we were using 1mW on each diode instead of 0.5mW for those measurements, which could give some effects on the squeezing measurements.
Since I had some time while the NEG pumps were conditioning, I decided to explore the mobile power bank near HAM5 and disconnect all extension cords, it had a 6 of them but only 2 were on use, by me, the other 4 not used. After removing all extension cords I powered off the bank via the breaker on VEAC-02.
Likewise I coiled up several extension cords that were cluttering the biergarten and area near BSC8.
[Sheila, Hang, Gabriele]
In brief, the L1 LOCK gain in ITMX was 10 instead of 1 (as it used to be during O2). Therefore the reallocation of pitch and yaw to M0 was ten times stronger, and it modified the suspension angular response as seen by the ASC at the lower stages. This created a crossover at about 1 Hz that produced the right-half-plane zero we measured in CHARD and DHARD. This wrong gain was there only in ITMX. We reverted it to 1 as it was during O2. The gain was changed from 1 to 10 on Saturday June 2nd 2018 at about 21:55 UTC. Nothing relevant written anywhere in the elog.
The plot below shows the yaw to yaw response of ITMX, measured with the optical lever, in a few different configurations.
The expectation is that this will fix the DHARD/CHARD plants (43844)
We implemented the low pass filters in all test mass M0 stages, since it gives us some more phase and gain margin at about 1.8 Hz, where there is still a zero.
The story of how we found this problem is long and convoluted. A brief summary below:
As part of this story, Jenne Hang and I checked the calibration of the test mass oplevs using the baffle PDs.
ITMY pit oplev is about 4% larger than what I got from the baffle PD, yaw is 3.5% lower, ETMY pit oplev is 13% high and yaw is 2% high. Jenne did a similar check for the Y arm optics, and I think that she also found that the optical lever calibrations are OK. (for locations of baffle PDs see D1200657)
The check that there wasn't a cross coupling between pitch and yaw on the ITM optical levers, Jenne and Hang steered the green beams to PD2 +PD4 on the ETM baffles, and in both cases the optical lever response in yaw was about 2% of the pitch response.
It would be great if someone could follow up on the ITM pitch to yaw cross coupling during some down time (without the confusion about top mass offloading).
For the yaw to pitch coupling, we already have good measurements of all test mass actuation responses. We only have to repeat the ITMX measurements since they were taken with the wrong M0 offloading. This will take about (10+10)*100 s = 2000 s of interferometer offline.
I will post more details with the measurements tomorrow, and hopefully there will be time to repeat the ITMX measurement.
Commenting on Sheila's log about oplevs checks:
To check the calibration:
To check that true pitch motion is reported by the oplev as pitch-only, and not a combination of pitch and yaw:
All 3 pumps were conditioned today, but we are not satisfied with the results, two of the pumps reached their target temperature (250 oC) but did not hold for an hour (if that is the program intent, to keep or "soak" for an hour at target temperature), and the 3rd pump did not reached its target temperature, it only made it up to 198 oC. We are consulting the manufacturing company about such issue, if indeed there is an issue.
All 3 pumps are currently isolated from the main vacuum volume, and we will revisit their status tomorrow.
{Kyle, Chandra}
This morning we soft closed GV 1,2,5,7 to isolate the vertex, x-beam manifold, and y-beam manifold and connected different calibrated leak checkers to the back end of the main turbo pumps in order to leak check each section separately. We were looking for a big leak based on historical corner pressures (mid -9 Torr range) compared to now (mid -8 Torr range). In short, we found no leaks from spraying helium in a "fire hose" fashion. Note we have leak checked all new joints previously.
First we leak checked the YBM, suspecting IP5 isolation gate valve based IP5's relatively high current. We sprayed all conflats, viewports, purge ports, feedthroughs, large GVs, and some welds and saw no rise in the background of 2.4e-9 mbar-l/s of He.
Then we moved on to the vertex: started at BSC 1,2,3 and worked our way down to HAM 2 and then down to HAM 5. The He background was quite high: 1.8e-8 Torr-l/s and over the course of 1.5 hrs of spraying large amounts of He gas, the signal gradually rose to 3e-7 Torr-l/s. We did observe a notable rise at the RGA, but after extensive spraying and cycling its isolation valve, conclude there is no major leak at the RGA (above -7 Torr-l/s background, that is). Note that the RGA is close to the leak checker and He could have migrated to its exhaust.
We didn't have time to leak check the XBM, but based on pressure trends (attached), the problem lies within the vertex. We think we're dealing with the aftermath of being up-to-air for three months with nine chamber doors off, in addition to adding spongy materials such as viton, kapton, peek in the main vacuum volume.
AMU 28 is the dominant peak in today's RGA scan.
Corner is being pumped by six large ion pumps (five rebuilt and two with chevron baffles).
The three NEGs were regenerated today and were left valved out for now. Turbos were valved out and spun down at lunch.
The AMU 28 peak hasn't changed since the last RGA scan on 8/14 (with four IPs and three NEGs valved in).
I rediscovered today that, unlike the more common 44" electric gate valves having a nominal 4 minute 18 - 20 second motor run time for effective soft closure, that 48" valves GV1 and GV2 were 4 minute 40 sec and 4 min 39 sec respective motor run times. These times being determined by "ear" and experience not by measurement.
Levitation of turbo rotors no longer needed as they had been spun down a few hours ago
Plot shows the PSL warming up by 2.5deg. Both IO GigE cameras see the change, but the change on IO GigE 2 can't be explained by just the change caused by the EOM crystal.
- temperature change, PSL North = 2.5degF
- delta IO GigE camera 1 X (channel 1) = 0.73 pixels, 4.1um
- delta IO GigE camera 2 X (channel 5) = 11.2 pixels, 62.7um
- IO GigE camera 1 X distance from PMC waist = 0.689m, angle change = 5.9urad
- IO GigE camera 2 X distance from the output of the EOM = 2.7m, angle change = 23urad
- EOM input = 0.746m from PMC waist, angle change = 5.9urad, beam position change = 4.4um
- Expected change at IO GigE 2 = 20.5um (total beam postion change) = 3.7 pixels
- Unexplained beam position change at IO GigE 2:
- pixels: observed = 11.2, calculated_expected = 3.7, unexplained = 7.5
- um: observed = 62.7, calculated_expected = 20.5, unexplained = 42
EOM crystal properties:
- index of refraction = 1.771
- length = 40mm
- angle of wedge 5.7deg (2.85deg each side)
TITLE: 09/11 Day Shift: 15:00-23:00 UTC (08:00-16:00 PDT), all times posted in UTC
STATE of H1: Preventive Maintenance
INCOMING OPERATOR: None
SHIFT SUMMARY:
Maintenance from 8am - 1:30pm PDT this morning.
LOG:
Thanks for a thorough summary Corey.
While trying to relock today, we had much worse recycling gain after doing initial alignment than what we had on Monday. Hang found that at around 18:06 UTC (11 am local time) the stored values for the ITMY camera were changed. He restored them and Geogria accepted them in SDF, but we don't see any logs about model restarts. It sounds like there was an unintentional restart of that model (43932)
SDF is really not in good shape right now, and it will be challenging for commissioners to keep up with it while we are in this state. It would be great if operators/ other people can help us catch these things by being on the lookout for model restarts and being on top of either checking checking conlog when there are unintentional restarts or checking SDF before planned restarts.
At around Sep 09 2018 07:10:53 UTC (GPS 1220512271), the IMC and FSS lost lock. I was unable to recover the FSS. The FSS is railed at about 11.9 volts. I can move this value around by turning off the autolocker and moving around the common and fast gain sliders, but as soon as I reengage the autolocker it rails again. The laser crystal temp at the time of the unlocking was 0.31 K. I tried manually setting the temperature there, but the PZT just is not responding, and every time I ask it for control it rails to 11 V. I pressed pretty much every button and shook every slider on the PSL_FSS medm screens. I don't know what else to do from the control room. My uninformed opinion is that somehow the PZT lost it's lower DC power rail, since it always rails positively. However, I can get the Fast monitor to read negatively if I disengage the autolocker, turn on the "mode loop" (no idea what that is), and turn up the fast gain (Picture 2 below), so I'm not sure what's actually going on.
There is no DC response from the FSS refl PD. The value on the REFL PD doesn't change when the PMC is locked or unlocked, however I can see that the diode did show a higher level of reflected light during a time when the FSS was unlocked last Tuesday.
In the second attached screenshot you can see that as I have locked and unlocked the RFPD over the last hour or so there has been no response on the FSS PD.
Another issue that is confusing is that the PMC sometimes reports that it is locked when it is not. The PMC medm screen has a misleading statement that the PMC is resonant if H1:PSL-PMC_MIN_TRIGGER < H1:PSL-PMC_TRIGGER_OUTPUT < H1:PSL-PMC_REF
Looking in the PMC model this is unrelated to the logic that goes into the flag H1:PSL-PMC_RESONANT, which is missing from the medm screen.
Corresponds to FRS Ticket 11427.
It's IMC VCO failure.
See the first attachment, at some point IMC VCO failed and eventually came back.
No VCO output = no diffracted light into RefCav.
That's the reason why there was no response from FSS DC diodes even though Sheila locked/unlocked PMC.
Reference input of IMC VCO is taken off of an RF amp and 8-way splitter (D1000124) that is also used for COMM and DIFF which were OK (indeed everything except IMC VCO was OK, see 2nd attachment), so this cannot be the amplifier problem, it should be VCO itself or the cable connecting the splitter and VCO.
It happened again at about 11:50 AM (or 18:50 UTC).
See the third plot. This time I went to the floor and wanted to power cycle VCO, but since it doesn't have a convenient power switch I disconnected the reference input cable (that's where VCO OUTPUTMON dropped from -10 to -34dB). After playing around with it, swapping cables with one of ALS VCOs and such, I couldn't find anything wrong, reconnected everything back, and it seems to be working OK for now.
I cannot tell if this is an electronics or a shoddy connection or shoddy cable/attenuator.
IMC VCO showed this behavior 5 times (counting today/yesterday as one) in the past 4 months where the output intermittently dropped to -10dBm.
I also plotted COMM VCO output, you might think something happened to COMM too for #3 and #4, but it's actually not the case (2nd and 3rd attached).
I've found the FSS auto locker parameters totally out of whack.
Oscillation threshold of 3 doesn't make sense, even when it's oscillating I don't think H1:PSL-FSS_PC_PP goes above 3. I manually set it back to 0.6.
State transition delays of [delay1, 2, 3]=[10, 10, 10] sec don't make sense either. The servo is first turned on with low gain after the refcav is brought close to resonance for delay1 sec. Then wait for delay2 and the common gain is ramped. Then wait for delay3 and the temperature loop is turned on. I manually changed them back to [0.1, 1, 0.1].
I pulled the PSL-VCO (S1200558) and performed some testing on it today. Power checks out, tuning checks out, monitors work, I was unable to compete the phase noise test before noon. I checked the attenuators and the REF cable, these passed as well. I changed the REF cable position on the RF Amplifier Splitter from position 5 to position 7.
Based on Keita's trending, REFERENCEMON drops out, this would cause OUTPUTMON to drop out as well and it does. DIVIDERMON does not drop out all the way, curiously. The power OK signal was good throughout these incidents, the intermittent fault is not likely power related.
Looking at the past 4 months behavior posted by Keita above, the glitching seems to be periodic, and when I zoomed in most of this glitching is happening on Tuesdays before noon.
5 = Tuesday June 12, 11 AM
4 = Tuesday June 24, 11 AM
3 = Tuesday July 31, 8 AM
2 = Thursday August 16, 9 AM
So it misbehaved when somebody worked near the rack. Maybe poor connection at the connector (too little or too much torque?) combined with poor strain relief.
[Hang, Craig, Gabriele]
We moved PR3, PRM and the soft degrees of freedom to improve the beam centering on all four test masses and increase the recycling gain.
We started with a recycling gain of about 38.2 and finished with a recycling gain of 47.4 [+24%]
A lock loss terminated our work, so there is still some re-alignment to do to get the test masses centered. And there might be more recycling gain.
We added dithering lines in yaw to all four test masses, and we could improve the centering significantly. At the end all Y2L coefficients are zero, and the dithering signals are quite close to zero. The two ITMs are reasonably centered, as is ETMX. Instead ETMY seems to have a larger mis-centering.
If we need to put back the Y2L coefficients, we probably only need to set ETMY to about 4.
We finally lost lock after more than 6h continuous lock.
Wow, very impressive. Hiro's simulations (G1700140) gave a PRG of only 39, though that was including an ad-hoc loss per arm of 50 ppm.
This is awesome!
Our PR gain is POP_A_LF / IM4_TRANS_SUM with calibrations to take into account reflectivity branching ratios, but we're quite close to the edge of IM4 trans. I think our normalized yaw number is more than 0.8. (We haven't pico'ed yet, to keep an alignment reference, but I think it might be about time to pico and say that this is better than O2.) I'm suspicious that our PR numbers that we've been reporting might be systematically high, if the number we're dividing by is smaller than the total actual amount of power in transmission of IM4.
That said, this has been true since our vent recovery (and also to a lesser extent during O2), so the +24% is certainly real, but the absolute values we've been reporting for PR gain might be high. We should pico and check the calibrations to be sure.
Looking at TR_X_NORM and TR_Y_NORM (which are normalized so that the single arm build ups are 1, at the time of the best recycling gain the build ups were 1646 and 1732 relative to the single arm, which would indicate a recycling gain of 45. Gabriele pointed out that the reflectivity of PRM is 3.1% according to galaxy, which means that this gives a recycling gain of 53.0
Another option for estimating the circulating power would be to try this method which has been done at LLO: 36745
WOW! Amazing!! Congratulations!!
Here's a quick computation of the power recycling gain as a function of the round trip losses in the arm (in ppm, sorry I forgot the units in the plot), assuming Tprm = 3.1%.
The two marked points correspond to the values of recycling gain estimated with POP_LF and with the arm transmission.
Dark offsets for TR_X and TR_Y are not very well set. The table below shows the TR values in different configurations, and the corresponding estimated recycling gain. Taking into account the dark offsets now the X and Y arm transmission give consistent recycling gains.
| Dark | Single arm | Full IFO | Recycling gain | |
|---|---|---|---|---|
| X | 0.045 | 1.00 | 1648 | 53.5 |
| Y | -0.007 | 0.98 | 1736 | 54.5 |
The REFL_A_LF also goes up by about 5% (relative to unlocked so it measures the ifo reflectivity) as expected. Great job!
From the level of reflected power (w.r.t. the off resonance state) we can estimate again the round trip losses.
The plot below shows that the REFL level we measured after re-aligning (8.3%) corresponds to about 61 ppm round trip losses, and a recycling gain of 53.4, which is very close to the recycling gain we can estimate from the arm power buildup.
The recycling gain improvement is seen in the CARM offset reduction sequence too. I ran the script that Sheila and Gabriele posted in alog-43344 on data from yesterday (starting at t = 1220674303).
This plots the reflected power (normalised) vs the arm build up (transmission normalised to a single arm). As we step through the CARM offset reduction, the TR_X increases and the REFL_DC decreases. The slope is an indication of the power recycling gain.
As of yesterday we're going better than the previous measurements, including compared to O1.
In addition to Georgia's analysis, here is a similar plot, with also some fit lines. The new curve is again consistent with no losses in the PRC and arm round-trip losses of about 66 ppm.
