X1DAQ DTS and H2 DAQ system install and test on LHO DTS test stand.

Rolf and Alex, August 18-27 2010.

Rolf gave a summary of the work which is minuted

here

executive summary:

9 new 2U diskless front end machines were installed. A new boot and NFS
server was installed. The front ends were loaded up with models to
approximate roughly 50% of an IFO. Data rates between the front ends and
the H2 DAQ were tested.

EX and EY systems were simulated. IOP comms was performed using 5565 RFM
NIC and switching equipment.

Dolphin and RFM IOP used to distribute timing to the front ends which did not
have an IO Chassis attached.

EDCU channel limit was found to still be evident, Alex to work on removing this
limit. The system as built has 48,000 epics channels, in line with the 
100,000 per ifo estimate.

awgtpman found to take many cpu cycles, needs to refinement.

dtt and dataviewer need some work to get them operational.

Attached are PDFs of the 3 of the DOF TFs of the first QUAD.

I have installed a 32-bit version of awgstream, multiawgstream, and awgexc_run at /opt/apps/awgstream on seiteststand.  I also compiled awgexc_conf.c using mex.  

Modified ~controls/.bashrc to include paths /opt/apps/awgstream/bin and /opt/apps/awgstream/matlab_tools in $PATH.  

Time is now the same on workstation and seiteststand, so awgstream and dtt apps can be used on workstation again.  

Modified awgstream.m to include path to awgexc_run, this needs to be changed to get the path without hard coding!!!

Tested awgstream as standalone executable, and matlab script that uses awgstream.m, awgexc_run, and so on.  Did not verify that signals are actually injected and viewable with dataviewer, but no errors were reported.

During the course of testing awgstream, noticed that the time was ~20 seconds different between workstation and seiteststand.  On workstation, modified ntp.conf to match that of seiteststand, restarted ntpd.  Time (as reported by tconvert now) is the same.

First file is a pdf of the actuator to disp. sensor and GS-13 transfer functions on unit 2 from 800 to 0.025Hz.  Second file is the same measurement on unit 1.  

Data stored in:
'~/seismic/HAM-ISI/X1/Data/unit_2/TransferFunctionData/'

files named: 'colocTF_100824_L2L_*freq_band*_test1.mat'

Plots located in:
'~/seismic/HAM-ISI/X1/Data/unit_2/Figures/'

Plots named: 'colocTF_100824_L2L_0p025to800Hz_*DoF*'

This directory contains the .pdfs and the .fig files for use in Matlab.

Script to plot data:
'~/HAM-ISI/X1/Scripts/DataAnalysis/unit_2/TFanalyze_100824_L2L_0p025to800hz.m' 

Added Adobe Reader to bscteststand in ~controls/Adobe, command is 'acroread' from an xterm on the iMac.

Modified ~controls/.bash_profile to include /home/controls/Adobe/Reader9/bin to include acroread.

Added bookmark to aLIGO LHO Logbook to firefox on bscteststand to enable logbook entries from bscteststand without copying files between machines.

The computer 'workstation' attached to the Seismic test stand can now talk to the outside world.  This means it can get licenses for Matlab and you can use firefox to view and modify the aLIGO logbook.  Please use your real login name!

Also set up ssh to not require re-entry of the password from seiteststand to workstation or the other way around...

Bookmarked aLIGO LHO Logbook in firefox
Installed Adobe Reader 9, the command to run it is 'acroread' from a terminal window.  
Modified ~controls/.bash_profile to add path to /opt/Adobe/Reader9/bin for acroread.

The first file contains plots of IFO-coordinate-basis transfer functions from actuator to displacement sensor and GS-13 for each 6 degrees of freedom on our unit_2 assembly.  The excitation band was from 0.025 to 800 Hz.  

The second pdf file contains the results of the same measurement from unit_1 testing. The third file is a text file of the parameters used for the measurement.

Data stored in:
'~/seismic/HAM-ISI/X1/Data/unit_2/TransferFunctionData/'

Plots located in:
'~/seismic/HAM-ISI/X1/Data/unit_2/Figures/'
This directory also contains the .fig files for use in Matlab.


Scripts to plot data:
'~/HAM-ISI/X1/Scripts/DataAnalysis/unit_2/TFanalyze_100820_M2M_0p025to800hz.m'

Build the bottom half of the Beam Splitter Top Masses, 1, 2, and 3 for creak bake. 

Assembly Notes: 
(1). Install stability legs after pulling down blades, as the base plate cannot be clamped to optics table with the stability legs in place. 
(2). Use 6kg for blade pull down mass. A 7kg mass pulled the blade past flat, putting reverse bend into the blade. There is no base line spec for the blade tip height adjuster. Therefore lowering the blade tip height may allow for a 7kg pull down mass, if necessary. 
(3). Mount the pull down wire clamp as far out on the blade tip as possible to clear stop bridge when removing wire clamp.  
(4). Used a 2mm side gap between the blade and stop bridge.
(5). Take care when moving the blade side to side, as the blade twists easily.   

The H2 CDS drawing


is posted in the DCC

Page 2 is the current combined H2DAQ-DTS system in the computer users room.

Page 1 is the proposed complete H2 aligo networks drawing.

I am maintaining the 


H2 DAQ Rack as-built drawing

For this week's test the H2 DAQ is part of the DTS system, and has
temporary DTS IP addresses. The DTS has been enlarged with many
more front end systems to simulate an IFO. This will allow us
to perform load testing of the H2 DAQ.

Last night I ran a Local to Local actuator-to-gs13 response transfer function measurement.  Plots are attached from last night along with unit_1's measurement from July.  A text file of the parameters used in July and last night is posted as well.

Scripts to Collect data located in:
'/opt/svncommon/seisvn/seismic/HAM-ISI/X1/Scripts/DataCollection/unit_2/'
  with titles: 
'TFcollect_100819_L2L_0p025to0p5Hz.m'
'TFcollect_100819_L2L_0p5to5Hz.m'
'TFcollect_100819_L2L_5to200Hz.m'
'TFcollect_100819_L2L_200to800Hz.m'


Data file location:
'/opt/svncommon/seisvn/seismic/HAM-ISI/X1/Data/unit_2/TransferFunctionData/'

Data file titles:
'100819_L2L_0p025to0p5Hz_test1.mat'
'100819_L2L_0p5to5Hz_test1.mat'
'100819_L2L_5to200Hz_test1.mat'
'100819_L2L_200to800Hz_test1.mat'

Scripts to Plot data located in:
'~/X1/Scripts/DataAnalysis/unit_2/'

Plots located in:
'/opt/svncommon/seisvn/seismic/HAM-ISI/X1/Data/unit_2/Figures/'

It looks like our resolution isn't as good as we'd like in the lower frequency band.  The same features at lower frequencies are not present in this measurement as compared to unit_1. The gs-13 phase is also 180degrees from unit_1. I will set up a new measurement over the weekend with more repetitions and higher resolution.  

As for now, our damping filters have been implemented and have been running stably.

Around 30 July, the Assembly was moved from the Test Stand to the Transportation container.  I'll attach a couple photos here; find several more in my web space here.

We are currently boiling nitrogen into the container and logging the humidity.  After a couple hours the humidity exhausting from the container has dropped to 16% from 32%.  We expect to get a container volumes worth(not accounting for the ISI) through in less than four hours at our flow rate.  Before the end of the day we'll throttle back a go slow through the weekend.  Monday we'll download the humidity log and close all the valves.  The last photo below shows the N2 Dewar attached to the container.  The exhaust is plumbed into a bag containing our humidity logger on top of the box.

Before moving the Stage1 PreAssembly onto the Stage0 Base on Tuesday, we checked the Base top surface bolted to the Test Stand for level.

Attached is my logbook copy of the observations.
Bottom line: Of 22 observations, a total range of 7.5mils.  Discard 1 reading nowhere near a support post and 21 observation are +-2.5mils.  

We (Rolf, Alex, Dave, Richard) started building the H2 aLIGO DAQ system
this week. Since it is being constructed and tested on the DTS,
this week's alogs will be filed under the DTS section but in part cover
the H2 DAQ system.

Rolf and Alex's visit started yesterday, August 18th. 
This visit will extend through to next Wednesday 25th.

Just a quick alog to note the start. The DAQ test stand
is being reconfigured as a H2 simulation to load test
the new H2 DAQ system. 

The H2 DAQ rack was moved from the MSR into the DTS room,
and the h2daqdc0 and h2daqnds0 were built for the
initial testing.

More detailed alogs with pictures to follow. As-built drawings
will be posted to the DCC.

Dave and Alex,

Using Mark's full list of 64 fast channels which should be acquired
by the DAQ we increased L1QTS.ini from 25 chans to the required 64 chans.

All fast channels are being acquired at 2048Hz.

We attempted to turn on the second and minute trends, but found that starting
minute trends crashed the DAQ. Alex took a look at the code, but was unable
to find the cause. So for now, we are only acquiring full frame and second 
trends. 

I did the math, and figured that with the disk space available we can have:

14 days of second trends
11.5 days of full data

We can extend the second trend look back at the expense of full frames
if this is required. 

The DAQ was restarted many times between 17:30 and 18:30 for this work.

Hugh R., Eric A., Jeff G.

We believe we have found the solution to the odd GS-13 behavior from the transfer function measurements.  After several failed attempts to implement the damping loops, the path from each GS-13 to the feedthrough board was traced to see if any cables had been swapped since unit_1 testing. 

 Hugh and Eric noticed the "clean side" of the feedthrough for H1&V1 had been swapped with H2&V2. Since both GS-13s for each location (i.e. 1 & 2) are on the same cable up to the feedthrough, H1 & V1 were being mistaken for H2 & V2 and vice versa.  After the swap, a quick test from 200-800Hz showed the Horizontals and Verticals were fairly close in magnitude in this bandwidth, similar to unit_1.  Full measurements of the full bandwidth to be posted soon.

Shortly thereafter, the damping loops were implemented successfully with gains of -1.  I will set up long collocated and coordinate Transfer Function measurements for tonight.