BSC-ISI common model was updated. I modifided the master model and a couple of MEDM screens. Changes were commited (rev 4230 for model and re 4231 for MEDM screen).
A correction was made on the T240 calibration gain and binary outputs block was modified to control STS-2s centering proccess. Overview and BIO screens were modified as well.
New codes were installed on BSCs 1 2 3 6.
Since further work continued in the chamber last week, the final chamber closeout had not been completed. After checking for task completions for the final, final time, I went in and
1) checked for tools (none found, good)
2) removed the in-chamber stool
3) took a few more pix
4) re-inspected the naked BS optic (yep, dirty @ ~1-3 particles per square inch in many regions of both HR and AR face - some there prior to last week)
5) wiped my way out of chamber
6) placed verticle 1" witness optic under the BS suspension.
For the 35W beam.
The ISI-BSC2 (tested at the end of last week) is functional. The testing report can be found at LIGO-E1300292-v1.
I had to restart the IOC for the dust monitors in the LVEA due to communication errors. Cause is unknown.
Found cdsfs1 completely unresponsive, no display on console, no ability to log in remotely. Powered off computer, removed and reinserted power cords, then powered back up. After usual reboot, system came back, disk verify shows parity mismatches which are being repaired.
Belated alog for Friday activity after Cheryl was gone.
A good news is that Cheryl told me that the IM alignment was not good when we (Kiwamu, Sheila and me) found in HAM1 that the IFO refl beam was clipping on Thursday. On Friday Cheryl restored IMs and Sheila couldn't see any clipping of the beam going to HAM1.
But the bad news is that PRM still has a huge offset.
I think it is essential to establish that the beam is actually at a good ballpark position on PR2. If the beam going from the last IM to PR2 has a reasonable angle and yet PRM is using up more than 90% of DAC range for PIT offset, one can conclude that PRM has a big natural tilt (though we don't know why). OTOH if the beam is not hitting PR2 or something, we have a bigger problem.
Cheryl told us that it should be possible to see the beam in front of PR2 if we have 00 mode flash. Since she also told that the straight shot beam (through MC1bouncing on MC2 and coming through MC3) looked good, I and Sheila tried to align MC using only MC1 and MC3 without touching MC2. We didn't do a thorough job, in the end we were still off in PIT (looked as if 10, 01 and 11 were dominant, but PIT had larger order modes flashing) but got 00 once in a while, but we couldn't see anything in HAM3.
Before giving PRM to SUS, the following should be done:
While Keita Cheryl and I were trying to see 00 flashed from the mode cleaner yesterday afternoon/evening, I ripped the door cover on the west side of HAM2. It is a small rip (about an inch long) right by the clip on the north side of the door. Since everyone had gone home and we didn't know where to find new door covers, we just covered back up when we left.
Attached are plots of dust counts requested from 5 PM April 18 to 5 PM April 19.
Both HAM2 and HAM3 ISIs were locked this morning for alignment work. It was a good occasion to take look at the time series of the ISIs' Capacitive Position Sensors, and make sure that the locked/unlocked shifts were within requirements.
Maximum shift observed < 800cts*.The locked/unlocked shifts are within requirements**. Results are attached.
*:Calibration: 32768cts/mm
**: Acceptance Criteria (E1000309-v13, p28)
Absolute values of the difference between the unlocked and the locked table must be below 1600 cts
April 10th between midnigth and 1am (UTC) (cf dataviewer trend attached)
M1 PRM Pitch offset = 2000
M1 PRM Yaw offset = 50
M3 pitch = 2000
M3 Yaw = 115
Today with the same offsets (cf medm screen attached)
M3 Pitch = 1164
M3 Yaw = -199
Keita and I looked at the PRM in both the old aligned position, p=2000 and y=50, and found that EQ stops looked good as expected based on Arnaud's spectra. Then we drove the PRM to p=4135 and looked again for an EQ stop that was touching, and found none. I looked at ISI trends, and the vertical sensor change from the locked for installation and rebalanced and refloated show that there was a tilt in the HAM ISI table and PRM saw something like 1800 counts change. From this, I believe that PRM is not touching an EQ stop, and that the shift in alignment is due to the tilt of the HAM2 ISI table, when it was rebalanced and refloated after changing it's payload. I left today with Keita and Sheila still possibly working on the IMC, the low beam in HAM3, and PRM alignment, so maybe this is all solved. IF not, then it's time to have SUS come take a look at PRM to see how healthy it is, and to relieve the pitch bias. Recall, it's aligned position has 2000 of pitch already, and we needed an additional 2000 to get REFL through to HAM1.
PRM spectra is plotted on the pdf attached for 3 different dates, with 3 different alignment offsets.
*in blue April 15th with Pitch and Yaw set to zero
*in green today (April 19th) at noon with Pitch=2000 and Yaw=50
*in red, today (April 19th) at 5pm with Pitch=3900 and Yaw=350
Except the noise floor that is higher for the last one (red), all curves match (in terms of resonnances), showing that PRM is/looks fully suspended (transfer function would be the only way to be sure, but spectra is already a very good indicator).
the conclusion is that even with large offsets in Pitch (taking more than 90% of the DAC range) PRM doesn't seem to touch any EQ.
PS : as the picture shows, when the ADC reads a DC signal greater than 10,000 counts (red number), the IOP watchdog trips (eg when a large offset is applied), meaning you'll have to bypass it if you need to keep the offsets on.
Alarms: Dust Monitor, IOP Watchdogs, CDS Front end Laser Hazard 08:06 Washington Crane on site for crane inspection Pablo and Michael working in H2 Laser enclosure for PCAL 08:36 Transition to Laser Safe 10:30 Sno Valley on site to work on chillers Japanese National Broadcasting film crew on site being escorted by Fred Raab Crane inspection in LVEA and at End X Aaron & Filiberto working at End Y 11:15 and 11:40 Dale escort two high school tours of control room 13:08 Washington Crane on site for crane inspection 13:20 Kyle pumping at End Y 14:00 Cheryl transition to Laser hazard 14:25 Alarm on air handlers at Mid Y, Kyle to check out.
The H1 DMT computers were moved within the equipment rack in the MSR to make room for additional CDS hardware. The DMT system has now been powered up and is available for use.
X1boot was powered down for a disk clone, which was successful. It has been powered up again and is available for use.
The two steering mirror assemblies, D1101851, are installed in BSC2. They went in fairly smoothly. One hitch is where the support bracket attaches to the chamber's lower bracket; the weld from the lower bracket to the chamber wall kept the two surfaces from mounting flush. The two mirrors are NOT installed yet. They oxidized during their bake and I am still working on figuring out if anything should be done about that. The mirrors shed a little, but it seems as though they are not shedding anywhere near the level that the chamber itself is.
We set up Viton Pads under HAM3-ISI balancing masses. The size of those pads is the same as the size of LLO's. However we set those pads under 1/2 the mass LLO uses, in order to rise the corner frequency of the mass+Viton stack.
Comparison before/after is attached.
The Viton Pads improve the 310Hz peak in V efficiently, by a factor > 5 without any drawback to it. Results are satisfying enough for now. We may try other configurations next time the chamber is open.
This Viton Pads Efficiency Test was plotted in the cartesian basis. Results are attached.
Best improvements can be seen along RX and RY, in the 310Hz region. Suspension modes can be seen there and they are known to be coupling with the ISI.
Plots show that the Viton Pads we used helped limiting the impact of SUS/SEI couplings, on the ISI's Transfer Functions.
Having those viton pads should also facilitate the design of control loops, and eventually allow pushing the performance a bit further, as more gain margin will be available.
Beamsplitter transfer functions have been running over the weekend in order to fix any issues during the day if needed.
The attached pdfs show :
Vertical degree of freedom is noisier than expected between 1.5 and 3Hz (3rd page of each pdf).
Gerardo went into the chamber, checked the EQ stops, and nothing is apparently touching. When the chamber will be free tomorrow during the day, I will run quick DTT measurements for the vertical degree of freedom, between 1 and 3Hz and see if it's still noisy. There were no need of taking the night for full measurements since the other DOFs are fine. I let the night to Vincent for ISI tfs if needed.
I ran a DTT transfer function on the vertical degree of freedom of the beamsplitter this morning
New TF looks nicer than the old one from April 12th, meaning the beamsplitter is healthy and passed Phase 3a testing
Will run a full set of TF tonight, for our record.
Data of the DTT measurement has been saved in a matlab structure named :M1VtoM1V_20130422.mat under SusSVN/sus/trunk/BSFM/H1/BS/SAGM1/Data and has been commited to the svn
Last measurement on the beamsplitter has been taken monday night April 22nd with damping on and off, before pumping down BSC2 chamber.
One thing to notice :
Vertical degree of freedom remains noisy around 2Hz, as on April 12th, even though dtt transfer function from aLOG 6158 doesn't show any issues. It could be that when taking TF with matlab, the drive for that frequency band is not high enough. I will try with higher gains in the MATLAB_TFs.m script and see if it changes anything.