From jzweizig@ligo.caltech.edu Thu Dec 27 14:37:19 2007
Date: Thu, 27 Dec 2007 11:37:13 -0800 (PST)
From: John Zweizig 
To: Keith Riles ,  
Subject: Trigger-based side coil segments

Hi Keith and Laura,

I have a preliminary set of SIDE-SENSOR 6Hz blrms segments. These have the
same trigger threshold as the ones I generated from trends, but since the
triggers limit the times to ~5 seconds rather than 60 seconds, the dead time
is decreased by a factor of about 7. I compared the results to the
segments from the trends and I found that there are quite a few kW triggers
vetoed in the trend-based segments that are not veoted by the new
trigger-based ones. I assume that this is because thing are ahaking enough
over prolonged periods of time that the glitches seen in the ifo may not be
correlated with the peaks of motion seen in the side sensor RMS. Anyway, the
new segments may be found  in


http://www.ligo.caltech.edu/~jzweizig/S5_Data_Quality/H1-ETMX-trig_segments.txt

and


http://www.ligo.caltech.edu/~jzweizig/S5_Data_Quality/H1-ETMY-trig_segments.txt

I'm not sure what the best way to procede with these would be. Since the
dead-time is drastically reduced, it may be possible to use the new segments
as category 2 (?) and then possibly extend times around these vetoes, either
by using the trend-based segments or by adding an arbitrary dead-time at a
later point in the analysis.

I also suspect that the something interesting might come from integrating
over a longer time, but I'm not quite sure how to approach this. I have also
tried higher thresholds and this gives better efficiencies, but I though
that the dead-time was OK as is. If you'd relly like to see a decrease in
dead-time of, say, 50-60%, with some loss of vetoes, that can be arranged
easily.

Anyway, If you have a quick test for segments, you might want to try out the
segmens in the above lists to see if they are useful.

Hope you had a great Christmas, and best wishes for a wonderful New Year.
Note that I'm in Ireland through Jan 7th and I don't have particularly
fantastic network access, etc. Nevertheless, I should be reachable with some
time lag.
Best regards,

John