Mitchell Robinson, Scott Shankle, Mike Vargas, Thomas Vo and Apollo's Tyler Manifold Cryopump for ETMx - Tyler modified 20 FHCS 1/2-13 and submitted for Class A clean and bake. Reviewed specifications for 'Made in India' group of screws for acceptability. Determined parts are ANSI qualified SSTL. ECR E1200612 was submitted and approved. Scott submitted design of spacer. Part fabrication in process; aim is to Class A by installation on Tuesday. Arm Cavity Baffle for ETMx - Started upgrading Arm Cavity Baffle from BSC8 to install in BSC9. Completed Spring Blade Assembly.
(Corey, Keita)
Went out to the Lab in the afternoon. With an autocollimator, we aligned the gold-coated alignment mirror (located at input of Telescope). Next up, made sure an iris was aligned (via a second autocollimator at the "output" of the small Secondary mirror (this is on "top" of the Telescope).
We then removed the upper autocollimator and mounted 180deg, such that it can be pointed at a "laser table" nearby. Used the upper autocollimator to position an iris on the laser table. The IR laser was then used to align optics on the laser table and then point the beam to the Telescope from the laser table. The Secondary Mirror's position was adjusted to make the output beam size acceptable---this was done only roughly.
Next up, we want to finely align the irises (via multiple Watec cameras), and then continue with alignment. All of this happens as we expect to receive (perhaps in the next few days) some new mirrors which will hopefully replace our less than desireable F1 Mirror.
The Telescope Assy was covered with a "tent" for a night with a big desicant pouch inside to keep humidity down (in the hopes of preserving Mirror coatings).
Received notification from vendor that replacement F1 mirror delivery is now 7/22/13.
Daniel Halbe, Josh Smith, Jess McIver Summary: strong, semi-periodic transient ground motion is propagating up the SEI platforms and down the suspension stages at ETMY. Cause of the ground motion is not yet determined. Effect on the HIFOY signal is not yet evaluated. Glitching in the top stage of the ETMY BOSEMs was first identified by Daniel Halbe (see Spectrogram_SUS_Longitundinal_M0_BOSEM_July2.png). These glitches are seen in all DOFs of the suspensions and seismic isolation platforms, have an irregular periodicity of about every 10-20 minutes, a duration of a few minutes, a central frequency of 3-5 Hz, an amplitude in Stage 1 T240s of the ISI on the order of a thousand nm/s (~ um/s) away from baseline noise, and have been occurring since at least June 12, 2013. They are not seen in ITMY suspensions channels. For a table that traces these glitches across each DOF and up the stages of seismic isolation to the top stage of the suspension, see: https://wiki.ligo.org/DetChar/HIFOYETMYglitching > Normalized spectrograms (PSD) of the periodic glitches for 1 hour 10 min Daniel also found them in the lower stage ETMY OSEMs: https://wiki.ligo.org/DetChar/SpectrogramsOfH1AllMassQuadSUSETMY And Josh Smith traced them to excess ground motion using a representative top stage BOSEM channel (see EMTY_top_stage_BOSEM_pitch_correlation_to_excess_ground_motion.png). These glitches have a strong correlation with local ground motion and significant correlation with ground motion near the vault. There appears to be faint correlation with ground motion near MX and the LVEA that merits further investigation. (See the normalized spectrogram Top_stage_BOSEM_ETMY_longitudinal_glitching.png and compare to normalized spectrograms Ground_motion_PEM_{location}_spectrogram.png of the same time period) For additional plots of ground motion at various locations around the ifo during these glitches, see again: https://wiki.ligo.org/DetChar/HIFOYETMYglitching (If you are unable to see some of the plots on this page, please see the instructions under 'Normalized spectrograms (PSD) of the periodic glitches for 1 hour and 10 min'). Note that the reported units of counts are incorrect for all plots (a bug in ligoDV) - these channels are calibrated to nm/s for inertial sensors or to um for BOSEMs and OSEMs.
According to the summary bit of the ODC, the ETMY ISI was not in a 'good' state during this time.
From the Hanford cluster:
$ligolw_segment_query -t https://segdb-er.ligo.caltech.
Returned no results.
TJ Massinger, Jess McIver
TJ did a similar study in the H1 BS top stage BOSEMs and found glitching at a lower frequency (2.8Hz) than we've seen in the ETMY (3-5Hz).
A comparison of the top stage BOSEMs of the core optics at Hanford is attached. The glitches seen in the beam splitter BOSEMs do not seem coincident in time with the glitches in the ETMY.
ISI states at this time are below (note that if an isolation loop is not indicated to be in a good state, it may be because the 'correct state' value for the comparison to generate the ODCs was wrong/outdated for some chamber until Celine fixed it a few hours ago):
We have seen larger fluctuations at the optics 0.44Hz mode. To verify that we indeed are still on the right mode, I took another mode-scan. This time it includes the 9MHz sideband, as well as the REFLAIR_A_RF9_I signal. We are still on the 00 mode, but the alignment is somewhat worse.
As for ETMY, " 2.1 damping filters " have been installed on ITMY top mass main chain, while the cavity was locked. The loops are working fine, and are stable. From now, damping GAINS will be set to -1
Attached is a spectra showing a comparison of the new filters with ISI Isolated (today, green curve) and the old filters with ISI damped (May 2013 blue curve).
In order to characterize the performances of the filters ONLY, I will take a similar spectra when the ISI will be damped, whenever I will have the opportunity to do so.
T Sadecki, J Oberling
Today, with Jason's help, I installed the wire loop in the reaction chain and checked the relative roll between the PenRe and ERM (which was installed earlier in the week). I then suspended the ERM in a single hang and by-eye checked the relative pitch between the masses. The pitch seemed acceptable (will be checked more quantitatively when we are able to do a double hang of the chain). Currently the ERM is suspended in the lower structure for the purpose of proof testing the loop for slippage overnight. Mating of the chains and attachment to the upper structure is up for next week.
For convenience, the UGF of the HEPI isolation filters can be increased by hand. With the position control, the open loops fall is 1/f and UGF of level 2 controller is 100mHz.
So far, the blend filters used on all ISIs didn't provide isolation (even a light motion increase) around 500mHz (the first pitch mode of the QUAD at 0.44Hz causes trouble to ISC). I installed new blend filters (blend frequency at 100mHz with a notch at 440mHz). I have attached the LP blend filters used yesterday night and the new ones (called T100mHZ_0.44). They shoud provide a factor 10 isolation at 0.44Hz.
Quiet day:
- CP1 was refilled early this morning triggered Vacuum alarm
- Some parts of HAM3 were adjusted this morning, check H1:HPI-PUMP_LO_DIFF_PRESSURE or ask Greg
- Work in Y-End, ask Alexa and Sheila
- On Sitemap -> CDS -> Overview: Dave fix a DDC error rebooting the channel
- Work in X-End, ask Apollo crew, Mike, Lisa, Scoot, Mitch, and Travis
- Fred gave a tour inside the LVEA to 11 students
- There were more visitants but unfortunately no one could give them a tour
- The Pcal LLO Y-arm crate is ready to be shipped
pablo
Using advLigoRTS trunk r3409, built and installed symmetricom driver on h1dc0. Then built and installed daqd for x1dc0, x1nds1, and x1fw1. Note that daqd in the target directories for the systems is a symbolic link to daqd-r3409. Restarted data concentrator.
Here are some instructions to ramp up the HEPI and the ISI
1- SUS must be damped (SUS watchdogs are connected to ISI)
2- ISI must be ONLY damped (using the "Command window")
3- Ramp up the feedforward using a gain of 1 in X, Y, Z. No feedforward in the other directions
4- HEPI can be controlled in position or with a IPS-L4C blend
1- Position control uses the IPS only. In BLND block, FM2 filter must be engaged for position control
2- There are 3 sets of Isolation filters in the ISO Block (Level 1 are the filters dedicated for the IPS-L4C blend, Level 2 are filters for position control with UGF at 100mHz, Level 3 are filters for position control with UGF at 5Hz). USE Level 2 or level 3 for now
5- Ramping up the ISI
1- Super sensor of stage 1 must use L4Cs when ramping up (T240 might saturate due to DC position offsets) Press 750mHz
2- Super sensor of stage 2 can be ramped with GS13 blended at 250mHz
3- Go to command and press Isolate Level 3 (40Hz UGF on stage 1, 32Hz UGF on stage 2)
4- Wait few seconds (till the T240s input values are less than 2000counts)
5- Lower the Blend frequency on stage 1 and introduce the T240s in the blend. Press T250mHz in all DOFs. Additional filters (T100mHz and T100mHz_0.44) can also be used to get some isolation at 0.44Hz (QUAD first pitch mode)
6- HEPI offloading. In ISC Mon. Turn on control in Y
7- Sensor correction (tricky during the day due to install activities)
1- Ramp up the sensor correction on stage 1 (Turn on the output switch and ramp up the gain to 1 in the GND bank)
2- Ramp up the sensor correction on stage 2 (Turn on the output switch and ramp up the gain to 1)
This applies for the three BSC-ISIs (except HEPI on BSC1 )
[Stefan, Kiwamu]
The demod phases for the IMC changed again and we readjusted them.
Then after a course of adjustments and etc, we came back to the point where we can update the CARM beatnote noise spectrum.
There is some improvements at around 10 Hz but the low frequency performance was not great today probably due to seismic noise, resulting in a worse RMS.
The spectrum update:
The attached is the new spectrum which we took tonight. Note that ISS was off tonight.
Green : our best noise curve
Blue : new spectrum with a low control BW
Red : new spectrum with a high control BW
IMC demod phase rotated again :
The demod phases for the IMC locking and WFSs rotated again. It is still unclear what caused it. This is the second time to see such a spontaneous phase rotation. The first one just happened yesterday and the second happened this evening. In this early evening we found that the IMC locking loop lost a factor of about 3 for some reason, resulting in a lower UGF. At this point we were OK with the demod phase which was newly set yesterday. However after a number of investigation we found that the demod phase rotated by 90 degrees backward as if it came back to the original phase.
Before we noticed it we were simply chasing the missing gain factor. We looked at the EOM modulation source path at which we found the loose connection yesterday. The output from the distribution amplifier looked OK --- the signal at 24 MHz had 3.4 Vpk-pk. Then I entered the PSL to check the power of the same RF signal but at the end of the RF chain right before the EOM. I disconnected the SMA connector at the side panel of the EOM can and plugged it to an scope. It seemed also fine --- the RF signal had 3.11 Vpk-pk which agrees with the previous measurement done by Volker a while ago. Note that during the measurement I closed the HPO external shutter to be safe. Also I had to disconnect the SMA connector of the 45 MHz source since this was in the way of the 24 MHz connector on the side panel. All of the SMAs were put back on after the measurement. Before getting in the PSL I switched the PSL from the science mode to the comissioning mode. I did the vice versa when existing. In addition to it Stefan was manually aligning a couple of steering mirrors on IOT2L to center the beams on length diode and both WFSs. This ended before I started messing up the PSL RF connections.
After these investigations Stefan found that all the demod phases got rotated by approximately 90 degrees so that the IMC length signal showed up mostly in the Q-phase. So for the reason we had to start over the phase adjustment.
The new demod settings:
It seems they came back to where they were with a descrepancy of roughly 10 degrees compared with the values we had until yesterday. The below are the new settings for the WFS demod phase.
And the below is a picture of the new setting for the delay line for the length demodulation.
The IMC controls :
With the new demod settings we became able to lock the IMC with no major problems. A good this is that the UGF came back to the original high value of 60-ish kHz, which was 25 kHz this morning due to the missing factor of 3. However the WFSs seemed slightly unstable. We saw a slow oscillation in pitch at an order of 100 mHz which was visible in the reflection CCD. We thought this was due to a too much gain even though we are not sure why. We decreased the gain of all the DOFs in the WFS by simply disabling the 10 dB gain flat filter for each of them. Note that the overall gain H1:IMC-WFS_GAIN is set to be 1. It is working fine now without an oscillation. Good.
Alignment of interferometer :
We moved on to the global alignment of the interferometer. The Y arm cavity was tweaked to maximize the green intracavity power. We touched both ITM and ETM. The BS was aligned by using the opleve as a reference. Later on we did a fine alignment of the BS by looking at the amplitude of the beatnote and maximizing it. At this point we established the green beam path so that we had a reasonably big beatnote signal at ISCT1. Then we moved on to the alignment of the infrared path. We then tweaked both IM4 and PR2 to get a high transmitted light of the Y arm in infrared. The alignment was mainly done by touching PR2 in both yaw and pitch. As a result we got 400 counts in H1:ASC-TR_A_SUM_OUT when the 00 mode is on resonance. Since the highest value we got ever was about 500 counts we concluded that this was good enough for measurement of the noise spectrum.
HEPI turned off for the night :
We found that the cavity length moving quite a lot tonight although we didn't see an apparent seismic activity in the seismic strip chart in the control room. However this was obvious since the end PDH kept dropping its lock frequently. The fluctuation was at a time scale of 1 sec or longer and hence the HEPI offloading didn't help at all as it was too fast for it. As a test we turned off the ETMY HEPI by running the script. But this tripped the HEPI and everything else (TMSY, ISI, ETMY). We tried to bring the HEPI back on by running the script but it tripped again after immediately the control were engaged. But it turned out that the cavity was quiet if the HEPI was left off. So we decided not to turn it on for tonight. We proceeded the noise measurement with the ETMY HEPI off.
A near-future task:
The IMC guardian must wait for the refcav for a long enough time when it is being recovered so that the IMC won't kick the refcav until it settles.
If the switch is down on the delay line phase shifter, a delay gets added. The old front panel provides some more information
Today, on BSC-ISIs:
- Calibration filters were reinstalled (the minus gain of the HEPI L4C is introduced in the calibration filter)
- The feedforward controllers were retuned/reinstalled on the 3 BSC-ISIs (The gain is 1 when engaged)
- The sensor correction filters were retuned/reinstalled on the 3 BSC-ISIs
At 18h50 PT:
- HEPI BSC6 and BSC2 are controlled in position
- ISI BSC6 and ISI BSC1 are controlled with
- Feedforward ST0->ST1
- Damping
- Feedback Level 3
- Sensor correction Stage 1 X,Y,Z
- Sensor correction Stage 2 Y
ISI-BSC2 controlled with
- Feedforward ST0->ST1
- Damping
The DAQ went 10 years into the future for about 10 minutes this morning due to an IRIG-B error. Unfortunately this bogus time was then written into the raw minute trends. You will find if you ask for minute trend data which spans this time (approx 9:22 - 9:50 local time this morning) you will only get data up to 09:22. To get the rest of the data you will have to ask a second time for data from 09:50 onwards.
We stopped writing raw minute trends for a short while this afternoon and switched them to a new disk location. This will allow us to fix the now static broken files, most probably tomorrow.
Mark Barton, Gerardo Moreno, Arnaud Pele, Scott Shankle, Mike Vargas, Thomas Vo Magnets were added to the plates. B&K Testing performed; results are being reviewed. Purchased ~20 18-8 SSTL 1/3-13 x 1.25" and 1.5" Flat Head Cap Screws (Made in USA) locally. Additional were Made in INDIA, but will not be used. Screws will be cut to size and Class A cleaned tomorrow. Alignment Tooling and Baffle/Ring Spacer Bar have been Class B cleaned. Forceps need to be shaved and re-cleaned. We will begin working Arm Cavity Baffle for BSC9 and return to Manifold Cryopump baffle for ETMx on Tuesday (7/23)
Kyle HAM1 currently vented but not nominally being purged (HIFO-Y noise), HAM2 under vacuum -> Closed HAM1 ion pump gate valve -> Opened HAM1 purge valve and exhaust of HAM1 turbo resulting in flow from bottom to top of HAM1 -> Returned to as found state after 90 minutes (purge valve closed, turbo exhaust capped) -> We are trying to systematically eliminate all adjacent external-to-vacuum potential helium reservoirs. Also, verified YBM leak detector calibration using external calibrated leak -> 5 x 10-8 torr*L/sec leak registered as 7 x 10-8 torr*L/sec value on detector -> Not a significant discrepancy for our current purposes -> left stored calibration factor alone
DAQ Re-start this morning at 9:22am(Dave & Jim)
Kyle working on HAM1 purge air flow, he finished before lunch
HAM6 cable running (Filiberto & Interns)
Running cable & bending conduit around HAM1 started this morning (Mark/Ed)
EY Field Rack Transfer Functions (Alexa/Daniel)
Prepping for Oplev work (Thomas)
Alarms in Labs--probably should raise alarm levels to 500 counts?
Stefan working on MC Electronics
Vincent restarted the DAQ for his work in the afternoon.