Minutes of Detector Characterization
Teleconference (September 3, 2004)
Present:
Caltech: Ron Drever, Peter Shawhan,
Patrick Sutton, John Zweizig
CSUDH: Ken Ganezer
Florida: Sergey Klimenko
HWS: Steve Penn
LHO: Mike Landry, Vern Sandberg, Daniel Sigg
LLO: Natalia Zotov
Loyola: Marc Cenac, John Whelan
LSU: Gaby Gonzalez
Michigan: Keith Riles
MIT: Jun-Wei Cao, Erik Katsavounidis
Oregon: Ray Frey, Rauha Rahkola
PSU: Shantanu Desai
Syracuse: Alessandra DiCredico
DMT Software Status & Plans:
- Needs and planning for engineering runs and S4
KR briefly reminded everyone of the tentative upcoming engineering run
schedule,
with a mini-run at LHO in mid-October and a weeklong run at LHO in
mid-November. To get ready for the mini-run, we need new or improved
DMT monitors to be up and running 24/7 in mid-September.
- Status of DMT infrastructure:
- John Zweizig: Version updates, etc
John has been helping Patrick Sutton lay the groundwork for a new DMT
calibration class (see below). He has written a generic XML parser that
can be used to read the input files Patrick needs for carrying out
calibrations. If that information can later be stored and retrieved
from frames, the new class can
be readily adapted. It includes handy utilities like "find".
John is also improving the automated building of the DMT to cope with
special cases, such as the absence of an FFTW3 library.
- Patrick Sutton:
Calibration infrastructure
Patrick provided a detailed set of calibration class templates to be
used by DMT astrophysical figure-of-merit monitors. These new classes
will include FDEasyCalibrate, which provides a user-friendly and
standardized interface to obtain calibrated power spectra, where the
mechanics of reading in reference functions and gain constants is
hidden from the user. SenseMon itself will eventually use this
infrastructure, which is based in part on the original core of that
monitor. The input XML files will need eventually to be produced
routinely by the calibration team, but for now, Patrick will create
what he
needs from S3 data. He plans to have the new classes available for use
by
other DMT authors by the next detchar telecon in two weeks.
- Status of DMT monitors under substantial development /
revamping / debugging:
- John Whelan / Mark Cenac / Brian O'Reilly: StochMon
Marc reported that the StochMon now works correctly, except for the
implementation of the overlap reduction function, which should require
another week or two of work. The monitor now provides to the DMT Viewer
two time series: Omega and 1/Omega**2, along with a frequency series of
the instantaneous power spectrum.
The code has been tested on a LHO machine and will soon be tried out at
LLO
and placed into CVS.
- Tim Bodiya / Keith Riles: SpecMon
Tim has been on vacation since the detector investigation camp where he
gave a demo on the latest version. He will be putting the new code into
CVS soon and asking John to create a standard executable for general
use. Bill Kells has been trying the program out and has given Tim
suggestions for improving the GUI interface.
- Ken Ganezer: seis_blrms
Ken is modifying the existing code to catch time step discontinuities.
- Sergey Klimenko: BurstMon
Sergey awaits Patrick's new calibration class and will ready the code
according to the new template linked above. He mentioned having trouble
keeping up with
data in real time, apparently because of the long time strides he has
been
using (10's of seconds). About 40% of the data is lost. The loss is not
critical,
since the monitor is a sampling one, but it means that trend files
would
be full of undesirable gaps. He is working with John to see if
adjusting
memory buffer sizes can eliminate the losses (CPU consumption should
not
be an issue; data flow seems to be the problem). He expects the new
code to
be placed in CVS in about a week.
He also stated that if there should be a shortage of available CPU time
during the S4 run, he recommends disabling the WaveMon monitor for
online
use and running it slightly offline instead. KR didn't like the idea of
disabling the only online DMT monitor that looks explicitly for
correlated transients in AS_Q and auxiliary channels. He reminded
everyone that a standing committee chaired by John is charged with
deciding which monitors to run during data taking.
- Steve Penn: BicoMon and BicoViewer
The new heterodyning option in BicoViewer for zooming in on regions is
now working, but a new bug has arisen that (mysteriously) prevents
reading
of offline frame files. Steve is working with John to sort the problem
out. The new "movie" option of BicoViewer is being used by high school
students in New York near Steve's campus as part of a semester-long
class project led
by their teacher who spent part of the summer working at LHO. Steve
expects the students to start confronting the data seriously in the
latter part of the coming fall semester.
- Giovanni Santostasi: PulsarMon
Giovanni was unable to attend the telecon, but reported he will be
spending the weekend at LLO to work on the monitor.
- Rauha Rahkola: SuspensionMon
Rauha is resuming work on the monitor and plans to have the
trend-writing bug diagnosed and fixed in the next two weeks.
- Natalia Zotov: PTMon
Natalia has run all of the S3 data through the monitor, using condor,
and has added detected SNR to trigger info.
- John Zweizig: LSCMon et al
No substantial work on monitors lately. .
- Next detchar telecon: Friday September 17
at 1:30 p.m. EDT (DMT software & S2/S3 investigations)