Just for the record, plots attached to show the problem we saw on the horizontal corner1 HEPI actuator.  The first and second plot shows the unmatched response slopes to the H1 drive.  The third plot shows an acceptable result although it might be nice to normalize/overlay the traces but we are also looking at the calculated slopes so we aren't just sticking a thumb up and seeing if it looks ok, but almost.  These acceptable slopes range from 1.75 to 2.00cts/ct.  The first plot H1 slope is low at 1.35 and the second is too high at 2.6cts/ct.  The PTB haven't given an explicit criteria, so, executive decision: if you have to look closely to see if the slopes are much different then they are ok.

A magnitude 8 earthquake coming from Chile tripped all the BSC-ISIs on 04/01/14. Sheila's wd plots show that the ISIs tripped on the actuator watchdogs. We suspected that the servo loops were trying to compensate for the common mode seen by the BSC-ISIs at both ends of the arms, thus saturating the actuators. We studied the ground motion recorded by the ground STS at the corner station (BS-STS) and the ground T240 installed at EX (EX-T240) at the GPS time provided by Sheila, in order to find out.

On the attached document:

• p1:  The time series recorded along X,Y and Z* show the three consecutive waves of the earthquake, coming one after the other: P-wave, S-wave, and then the surface waves.
• p2:  The common motion is highly dominant along the degrees of freedom which are the most affected:
  • Z for P-wave. Amplitude peak corresponds to actuator trip time.
  • X and Y for S-wave
• p3: The ground sensors show a high level of coherence below 0.1Hz, We expect to see most of the common motion there.
• p4: Most of the motion is seen below 0.5Hz. X and Y are affected the most .
• p5-7: The differential motion is smaller than the common motion by up to a factor 6, in the frequency region of interest.

We could considerably reduce the amount of signal sent to the BSC-ISI actuators during a major earthquake (factor as big as 6), and thus reduce the risk of saturating the actuators, by removing the common mode from the signal we feed to the servo loops of the BSC-ISIs.

Data, plots and scripts are commited to the seismic SVN:
/ligo/svncommon/SeiSVN/seismic/Common/Data/2014_04_01__Chile_Earthquake_Data/

Note 1: X and Y had to be swapped on EX data, and a minus sign was also added to the new x-axis EX data. It is likely that EX T240 is mis-oriented.
Note 2: The calibration of EX-T240 was still the one that JeffK made, at the time we looked at. RichM updated it since.

We Left ETMX running over night (it wasn't very windy so this should probably count as a quite time, I need to make a tool that will tell me that kind of thing).

Sensor correction was installed in three degrees of freedom last night and running. This is ground to stage 1, using Ryan's filter which is targeting the first suspension mode at 0.45Hz.

I've attached three sensor correction plots on/off, blue is sensor correction off the other colors are sensor correction on with various gains. The X direction gain ended up at 1 (which is what it was designed to be) Y and Z came out to be 0.6,
which is very weird.

  There are also three performance plots from last night, with the GS13 and T240 curves.  In general we are buried in the GS13 noise above 0.5Hz (red vs green) except at the BSC pier resonance, we do not have the HEPI L4C feed forward filters installed yet. GS13 rY and rX tilt coupling (yellow line) explains most of the low frequency gs13 signal.  There is clearly some room to relax rZ loop to reduce the gain peaking below 0.2Hz

and again none of this is Stephen's fault I can't figure out how to change the login to me

Rich 

Mabey I hope that I now have the correct plots

Latest curve. 11 days shown.

loglog plot also attached - note that the BSC6 annulus was probably vented until day 8. There may be a slope change after this date.

 

   Here is a check list for setting up/working with the ground STS2s (or T240)

   Before touching a STS2 the THREE test masses should be locked using the special non-magnetic screw driver, the T240s do not have locks and are more rugged
 
   After you are done mucking with it

   Make sure that it is aligned using the swizzle stick

   Make sure that the instrument is levelled (bubble level on the sensor)

   Make sure that the feet are locked  (unlocked feet are not stable)

   For STS2s unlock the masses using the special screw driver

   Restore the thermal insulation (exactly what we need is not clear, but we know that at least an enclosure is important)

   Zero the mass positions (this happens at the electronics racks) and read mass positions to make sure that it was successful

  ------------------------------------'

  Eventually it is nice to take a transfer function from the ground instrument to a set of ISI platform instrument to check the alignment

Jim B, Filiburto, Jim W, Dave

we powered down h1seih45 computer and its IO Chassis. Filiburto installed a fourth binary cable to the output port of the second Contec card (HAM5's card). He disconnected HAM4's CN-B from the upper 32 channel input of the binary-output chassis and connected HAM5's CN-A to this port.

We powered up the system, verified that the new logging script worked.

Jim W verified binary output switching for both HAM5 and HAM4 is working correctly.

reboot log shows the startup as:

We have insulated ~100 feet of beam tube as a trial. In the near future vibration measurements will be taken in order to understand the damping performance of this configuration.

Most recent attempt at this had no effect as I had made the programming changes to the 7000V-5000V step instead of the 5000V-3000V step

So we ran overnight at ETMY with the HEPI position loops on.  I've revisited the data I posted yesterday morning.  In the attached plot I've added some of last night's overnight running with the loops on.  The dashed lines continue to be the off loop reference from ~2AM Tuesday, The solid line references(REF17+) are with the loops on but starting about 8am Tuesday, possibly a noisier time.  The current traces are from 2AM Wednesday with the loops on.  It is involved but my simple analysis says that some places are worse with it on and at another time it is better suggesting it is just dependent on the ground motion at the time.

When I started the loops yesterday I just put the current positions into the Target. From my log I put in 12 March, the Target positions should zero.  The differences are 57 to 30 ums in position,  11urads tilt, and <7 urads Rz.  The HEPI Position at ETMY is ready for controlling as needed; so if the current position is not close enough, go for it.

Partway through our accumulation on Y2 - strange CC gauge behavior on one gauge and different rate of rise on another.

We might have a small leak at our RGA setup - will check at the end of accumulation.

The default version of Matlab for the control room is now 2012b, as it was previous to Tuesday's internet outage.

So this actuator got 17+hours of bleeding, hopefully sufficient.  Shoot, I forgot to grab the serial number again.  Replacement SN is 22952

no restarts reported. 

Report from Richard McCarthy:

The site contractor was able to make repairs to the fiber and our
network services and phones have been restored as of ~3 AM today. 

(Sheila, Alexa)

We have been trying to search for the 1/f feature present at 1-50Hz in the noise plots. We tried a few different things as described below. The noise spectrum under nominal conditions is represented in REF0-2.

1. Hepa fan on/off on ISCTEX:  We turned the HEPA fans on and noticed that the 1/f feature appeaared to be upconverted; the noise improved around 1 Hz, but got worse from 5-50Hz (see REF 3-5).
2. Modulation frequency: Under the nominal conditions the modulation frequency is at 24.407363MHz with a 12dB attenuator. We changed the attenuation in this path several times and found no noticable/significant difference in the noise.
   • 6dB attenuation  --> REF 6-8
   • 0dB attenuation ---> REF9-11
   • 18dB atttenuation --> REF 12-14
3. Laser power: The nominal power after the second faraday is 53mW; the nominal power at the bottom periscope is 47mW. We adjusted the HWP before the faraday to reduce the power at the second faraday to 25mW. We found that the noise increased by a factor of 2 from around 4Hz to 50Hz as seen in REF 15-17.

All these plots can be found under /ligo/home/sheila.dwyer/ALS/HIFOX/COMM/COMM_COherences_April8.xml. In the set of three, the first reference is always the power spec, the second the RMS, and the third a coherence plot. I have attached an image which shows the nominal power spectrum and the cases with the HEPA fan on and one with the power reduced (NoiseSearchPic1).

For future improvement of the QPD servos, we adjusted the input beam onto the QPDs with the servo off, such that the beam was fairly centered onto both QPDS. We did not get a chance to work on the actual servos though.

We proceeded to do some more noise hunting, but halted the effort when we saw two new spikes in the noise spectra. One spike was at 9.8Hz and the other at 13.8Hz (see second png). Turns out these are the bounce and roll modes of the suspension that must have got rung-up during an ISI trip. (Refer to Sheila's alog).

(Rich M, Sheila, Alexa)

HEPI and ISI ETMX tripped because we were playing with the suspension watchdogs.

Sheila, Chris, Alexa, Stefan

This morning we looked for coherences with many signals with our COMM noise.  There weren't any surprises, we see the same coherence with the end station PDH control signal around 1kHz that we are used to, ISCT1 accelerometers around 70-100Hz, and a little coherence (-0.7-0.9) with the oplevs at low frequencies (pitch at 0.5Hz, yaw around 0.1).  We don't see much coherence at all with the TMS QPDs, just a small bit at low frequencies.  We lose coherence with the in loop sensor below around 100Hz.  

Chris and I went out to End X to look into the beam quality problems.  We didn't find much that was promising.  We checked rather carefully for clipping on the way into the chamber again.  We also set up the nanoscan again, and we see that there really isn't anything coming back to the WFS when the ITM and the ETM are both misaligned (so it isn't a second beam that is directly reflecting a lens or some optic on transmon.)  

We also tried Daniel's suggestion of closing down an iris in the path.  For WFS B this only made things worse.  WFS A scans will be attached to this alog soon, but we didn't get any improvement there either.  

We came back and tried out some of Rich's blends.  We were able to get relatively low noise once without using WFS, 30z down to 0.02 Hz. The attached plot shows the COMM noise (measured by the IR) with different blends on.  We are curently seeing a lot of noise on ETMX T240s at around 3Hz, (100 times more) Rich is looking into this.