jenne.driggers@LIGO.ORG - posted 10:40, Friday 25 March 2022 - last comment - 11:34, Monday 28 March 2022(62350)
Nominal Low Noise at 50W
[Varun, Jenne, Georgia, Craig, Brad, Austin, Camilla, Keita, Richard]
We're at Nomial Low Noise at 50W, and only needed to do one thing by hand :)
- We did an intitial alignment this morning, since the IFO didn't want to relock nicely last night. The alignment went very smoothly, to the point that I'm not sure we needed it.
- The updated POP A offset used in DRMI locking did bring POP18 down a bit, but we'll see how the rest of the locking sequence goes. If it goes smooth and easy, we may want to leave the offsets there since it's better for the fully locked IFO, even though it makes DRMI look a little funny.
- Now that I type that, I note that since DRMI ASC gets offloaded, this may make next lock's acquisition more challenging. We'll see.
- Once we finished the CARM offset reduction sequence, POP18 came back up a bit. Not really sure why, but it seems good.
- PRC2 and ADS 3 are nice and close to zero, so that was probably good. I stoped at PREP_ASC, which may be good since INP and CHARD are still quite far from zero.
- Just letting the IFO do ENGAGE_ASC seems to have worked. It's probably been more of a problem lately since ADS3 was so far away. So, I'll leave the POPA offsets in.
- So far, the IFO has just locked itself all the way to 10W. Very nice!
- I put the A2L gains from last night into lscparams and loaded the guardian so that move_spots will go to there.
- I also uncommented from guardian the staying at the center locations (which we had done before we had refound good high power positions).
- I changed the NLN power in lscparams to be our nice round 50W, and then reloaded ISC_LOCK, IFO_ALIGN, LASER_PWR, and IMC_LOCK.
- Went up to 50W with guardian, seems like we can just click that state and the IFO will lock right up!
- ETMY Mode1 violin seems to be high. Not sure why. Not yet show stopper, but could be if it continues to increase. Varun is looking at damping it by hand with a lower gain.
- We're thermalizing verrrrryyyyy slowly, and improving most of our buildups over time.
- LOWNOISE_ESD_ETMX caused CSOFT P to ring up again, as it did last night. But, Varun found an old alog 58818 saying that in the past we reduced the CSOFT P gain to 20. So, today I reduced it further to 15 (and then to 10) and the osc is gone.
- Varun is putting this gain change into LOWNOISE_ASC.
- LOWNOISE_LENGTH_CONTROL worked. We need to check coherence with MICH and SRCL to see how much we need to tune.
- After that, we clicked NomLowNoise, and the guardian did laser noise suppression then went to NLN. Hooray!!
- The noise from goings-on in the LVEA, including cleanrooms, are obstructing most of our low frequency noise. We know from yesterday that turning off the mega cleanroom improved our noise, so it's now more urgent that we help other teams wrap up their LVEA work so that we can get those turned off.
Next up: On/Off tests with some of the cleanrooms (when okay with other teams), and other low frequency noise improvements (such as checking and tuning the MICH and SRCL LSC FF if needed).
Images attached to this report
Comments related to this report
Richard has turned many cleanrooms off for a few minutes, which has dramatically improved things, and revealed some other scattered light shelves and / or acoustic peaks (note that the pump carts are still on). So, we've clearly got some work to do in working on our low freq noise, but it's still very exciting that we're close enough to start looking for these things!
Images attached to this comment
[Georgia, Craig, Austin, Varun]
- Richard went out on the floor and turned off the cleanrooms. Georgia made the plot attached to compare the DARM spectra with and without the cleanrooms on. The HAM 3 and HAM 5 cleanrooms are responsible for peaks in DARM between 400 and 500 Hz.
- Craig reduced the ADS gain and increased the CLK gains as the ADS lines which were burried under the scatter shelves.
- We notice the scatter shelves at 18-20 Hz, which is closely in line with the OFI scatter observed at LLO alog 58287 57890.
- Violins again: The ETMY mode 1...
- was rung up with a gain of -0.1 and filters FM1 FM4 and FM6
- We tried damping at a reduced gain of -0.01 but the mde still kept ringing up.
- We changed the filter setting from by 60 degrees to filters FM1 FM4 and FM5 and turned on damping with a gain of -0.01 and the mode again rang up.
- We changed the filter setting back to filters FM1 FM4 and FM6 and turned on damping with a gain of 0.01 and the mode is damping down. The gain is now increased to 0.03.
- Craig and Georgia are now stepping the pointing to AS_C to check the OFI alignment. The alignment looks better with the peaks being lower. Plot by Georgia attached here.
- Next, Georgia moved SR3 to see if the alignment into the OFI can be improved. Original SR3 slider offset values: P 437, Y -147.3.
- Keita turned on the OFI temperature servo. SITEMAP > SYS > EtherCAT overview > SqzCorner > Ioo > Ofi > Tec > Loop.
- Until now the temperature loop was not closed and the OFI was sitting at ~21.6C (H1:IOO-OFI_TERMISTOR_1_TEMPERATURE). The loop is closed and the OFI is at 25C now.
- Craig is moving the OFI yaw alignment. We have not seen any improvements in DARM yet. We have left the OFI yaw offset at 3000.
- Richard turned off the mega clean room for us between1332282938 - 1332284030. The alignment into the OFI has improved the scattering shelves from this morning, see plot.
75 Mpc with the cleanroom off!
HAM5 and HAM6 ISI shaker test results coming soon in Georgia's alog!
Images attached to this comment
BOOM!
Follow up from yesterday's power trend plots in alog alog 62347, addressing the short lock we had from 2021. Now the 2021 lock is six hours of data, and is compared to six hours of data from today. 0) We have about the same amount of light incident on the PRM (40 W). 1) It does seem like POP18 decays with thermalization far more than it used to. 2) REFL LF never achieves the heights it once did. 3) Thermalization agrees with carrier buildup. 4) POP90 looks better than before.
Non-image files attached to this comment
Because we were seeing scattering shelves, I wrote a uncalibrated DARM spectrogram maker in python quickly,
It's in https://git.ligo.org/aligo_commissioning/labutils/-/tree/master/spectrograms with some instructions for how to run it.
On CDS machines it's in /ligo/gitcommon/labutils/spectrograms/.
I only plot between 10 and 450 Hz, because otherwise the violins drown out everything.
The main takeaway from these plots is how stationary the scattering shelves are. The peaks rise and fall by about a factor of 2 or so, but are always present.
We have been thinking about output optics mostly, because of the OFI problems at Livingston. We should perhaps be suspicious of our input optics and the new PSL.
Posted first is an example from when Georgia was moving AS_C offset in pitch. We went to a bad place, as seen by some of the glitches across the band, then she moved back and recovered the "quiet" noise.
Posted second is from when McCarthy turned the mega clean room back on. The noise humps above 100 Hz returned, but the low frequency stuff remains about the same.
Non-image files attached to this comment
- We tried shaking HAM5 and HAM6 using the same exciatation as in this template. We saw no coupling to DARM but perhaps we need to tailor our injection to produce fringe wrapping like what Anamaria used at LLO.
- We tried exploring more SR3 pointings by making the AS_C pit offset more negative (approaching -1) but we started to fall off the diode and also saw extra glitching in DARM so we backed off. AS_C offsets are now at (P, Y) = (-0.7, 0.75). We started at (-0.79, 0).
- I started a very simple script on zotws26 which steps the SR3_Y alignment slider offset by +0.1 counts every 30 seconds (it's slow because SRC2 needs to keep up). In 4 hours it will stop and we can see if this was a good idea. Varun quotes the original SR3 location above. It looks like as we step the slider we slightly recover POP RF18. If we need to stop the script we can take the interferometer to NLN_CAL_MEAS (it checks the guardian state is in NLN before stepping the slider).
I just turned the SR3 yaw walk around since it seems like it's making the range worse. It will step down to an offset of -247.3.
We lost lock during the SR3 yaw move, it's hard to tell exactly what happened but there looks like a 0.43Hz ringup in the test mass oplevs.
While moving SR3 the noise in DARM got a bit worse (see attached spectrum). When we're locked later in the week maybe we cna try a SR3_P walk. SR3 is back in it's origianl position.
Images attached to this comment
