S2 / S3 Investigation Reports

• S3 Data Quality Segment Repository - Version 0 (Keith Riles)

KR showed a preview of the version zero data quality segment repository for the S3 run. The repository info is based directly on the segments defined by operators hitting the Go button during the run and compiled by Peter Shawhan's real-time scripts. Although preliminary information on DAQ artifacts, timing, and elog-noted anomalies is becoming available, most of that information will not be incorporated into segments until version 1 is released.
 
• Interchannel correlations   (Nelson Christensen)

Nelson reported on a large variety of lines seen in LHO and LLO data. He has made tentative identifications of LHO lines from correlating with Mike Landry's line compilation (although a subsequent e-mail from Robert Schofield suggested caution on some of these identifications). Nelson will try comparing the LLO lines with a separate list of lines available on london.
 
• IFO Performance measured by SenseMonitor (Patrick Sutton)

Patrick has released a summary of  the S3 performance of the three IFO's, as quantified by the SenseMonitor inspiral ranges. The best performance seen (briefly) was a range of about 6 Mpc on H1.

DMT Software Status & Plans:

• Keith's DMT Wish List

KR presented a wish list for DMT development between now and the S4 run:
• A variety of astrophysical figures of merit, some of which are already receiving attention by the data analysis groups
• Improvements, bug fixes, and completion of several existing monitors
• General wishes on documentation, better exploitation of infrastructure, and performance measures (see John's report below on new infrastructure for measuring performance). Daniel Sigg added his own wish that there be a "quick-start" manual written for beginner DMT programmers, with code examples shown for performing various tasks.

 
• Status of DMT infrastructure
• John Zweizig: Version updates, etc & simulations

DMT version 2.7 has been running at the sites nearly unchanged since the start of S3. Trigger archiving in the database was turned off at the end of the run, but the monitors continue to run 24/7. John is working on version 2.8 on saiph and other machines at Caltech. The new version allows writing triggers to user-specified files and has new support for simulations work that requires version 1.0 FrameCPP. The new version will be tagged as soon as LDAS 1.0 is released (in about two weeks).

John also summarized the status of DMT support for signal/artifact simulation and for monitor performance evaluation. A new program called MatchTrig allows easy measures of monitor efficiency vs control parameter. For example, a glitch monitor can be tested against a family of waveforms as a function of monitor threshold parameter. A number of sample output plots were shown of timing resolution on glitches, efficiency vs threshold, time offset vs threshold, efficiency vs signal phase/amplitude/Q etc.
KR note: this should be an extremely useful tool for vetting and tuning monitors; DMT authors are strongly urged to exploit it.
 

• Szabi Marka; Peter Shawhan: Alarms and web interface; OGIL

Prior to S4 Szabi will upgrade DMT alarms to include additional silencing options, an "I am alive" blinking confirmation, and the ability to relay messages from the alarm handler about the status of monitors. He also plans to upgrade the web interface with an increased refresh rate, and more parallelization of tasks for better speed. Daniel wants to see more documentation of how operators and scimons should respond to particular alarms. He feels more education is needed, or alarms will start to be ignored. Szabi felt that the infrastructure of alarm severity and audibleness is sufficient, but that better judgement (and more work) is needed by monitor authors to use the exisiting infrastructure correctly. KR added this issue to the list of what should be addressed at engineering runs between now and S4.

Peter talked about a new script-based GUI program OGIL (LIGO spelled backward, but meant to connote a loving look at the data) that allows one to monitor the status of the three IFO's remotely from any computer with internet access. The program queries conlog and the DMT web interface to get information on the present states of the IFO's (science mode, injections, etc.)  and their recent histories, inspriral ranges, etc. A related program called S3stats can be used for monitoring coincidence-science-mode statististics. Peter developed the programs during S3 to assist in the injections work and plans soon to install them in LIGOtools under the category of Runtools. He mentioned that it would be straightforward to include GEO/TAMA status if those experiments made the analogous quantities web accessible. KR remarked he looked forward to trying out OGIL on his desktop in Michigan.
 

• New DMT Monitor - DeltaMuSigma (Daniel Sigg, Nathan Hindman)

Daniel summarized an elog entry with plots posted earlier in the week concerning the results of a new DMT monitor designed to quantify the effect of micro-seismic noise on the required dynamic range of suspension controllers. Unlike the seismic blrms monitor, this one takes into account explicitly the coherent difference between seismometer measurements at the end stations and the LVEA. Calibrations are used to allow direct comparison with IFO feedback control signals. About a month of S3 data has been examined so far, and there do appear to be periods when locking troubles correlate with high differential microseismic activity (better correlation than seen with seismic blrms). Different levels of coherence along the arms indicate different sources of the microseismic waves, e.g., Alaska vs Washington coast. In software one can definitely reduce the rms on the control signal by exploiting the seismometer readings, but it's not so easy to implement the signals in a real-time servo. It's also not clear that enough benefit can be gained from implementing the servo to pursue the issue, given other work in the queue. (Following up on his remark above, Daniel stated progress on the monitor was slowed initially by the lack of a quick-start manual for Nathan to follow in learning the DMT infrastructure. Note from KR: development of a useful new monitor in five days ain't shabby!)
 
• Status of DMT monitors

 
• Dave Chin / Tim Bodiya / Keith Riles: LockLoss and ServoMon/SpecMon

KR reported that Tim has produced a new version of SpecMon for testing. It is still not yet ready for primetime because of socket communication troubles between the foreground GUI and the background engine.
 
• Ken Ganezer: seis_blrms

Ken reported by e-mail that he is working on a couple of bug fixes (see also KR's DMT wish list above).
 
• Masahiro Ito: glitchMon

Masahiro is working on a minor upgrade that allows selective disabling of trending.
 
• Sergey Klimenko: LineMonitor and WaveMon

No update.
 
• Szabi Marka: IRIG-B, TimeMon, and ShapeMon

Data from the timing monitors has been used to produce the recent S3 timing report (see last telecon minutes). No upgrades are currently planned.
 
• Tim Olson: SpectrumArchiver

Tim reported by e-mail that he is working on a change to allow config file specification of channels to be archived (see also KR's DMT wish list above). John Zweizig remarked that he would like to see the SpectrumArchiver source code in the CVS archive.
 
• Steve Penn : BicoMon

No update.
 
• Brian Stubbs: SuspensionMon

Brian is finishing up the trending code and investigating a problem with the settling time of statistics when the monitor is first started up. He is also working on the online documentation.
 
• Patrick Sutton / Mike Ashley: SenseMonitor / LineAmp and RayleighMonitor

Patrick has agreed to provide an encapsulated class containing the SenseMonitor underlying machinery for computing calibrated strain spectra from AS_Q. This new class will allow other astrophysically-anchored figures of merit to be displayed in the control room (see KR's DMT wish list above).  Patrick is also working on a nagging aliasing problem observed some time ago in RayleighMonitor where high-amplitude lines in AS_Q are seen at much lower frequencies in the RayleighMonitor grid displays. The aliasing mechanism is not obvious.
 
• Natalia Zotov: PTMon

Natalia is focussing her work on ivana analysis, but plans soon to investigate an apparent systematic underestimate of SNR for certain PTMon-detected glitches.
 
• John Zweizig: BitTest, DataQual, HistCompr, LSCMon,PSLMon , SatMon, and SegGener

Nothing new to report on most, other than the glitch monitor evaluations discussed above.  LSCMon is being used to investigate the effect of flaky S3 DAQ on AS_Q (see last telecon minutes), but there remains a problem with faithful emulation of the LSC front-end code that may ultimately prevent success.
 

A.O.B

• Next detector characterization telecon: Friday February 13