Review Committee Meeting Monday 26 June 2006 09:00 PST / 12:00 EST
Minutes: Monday 26 June 2006 09:00 PST / 12:00 EST
Agenda and Contact Info
- Continue discussion of comments on the draft paper for the S2-S3-S4 GRB-GWB search [ PDF ]
- Continue discussion of SGR1806-20 search [ HTML ]
AccuConference teleconferencing service: Phone: 1-800-704-9896, participant code: 038621# International callers ++1-404-920-6472 with same code
- Discussion of Isabel's document on initial results of polarization
study [ PDF ].
- Bug in simulation code discovered: wrong sky positions passed to function that computes F+ and Fx. Simulations need to be re-done.
- Find that attenuation of signal is essentially equal to sqrt( F+^2 + Fx^2 ) for circularly-polarized waves; a factor of ~5 greater attenuation for hrss90 for linearly-polarized waves.
- Discussion of Jolien's document on estimation of the search
sensitivity [ PDF ].
- Predicted distribution of max-cc agrees well with values that the search gives (detailed earlier in Soumya's notes).
- From this distribution can estimate sensitivity of search in terms of expected hrss90 limit. Agreement is not too bad for the circularly polarized results. Isabel noticed an error in first draft of these notes: had the wrong Fave in the example.
- Isabel will investigate simulations of GRB populations to show how methods of detection are complementary and to justify binomial method.
- Discussion of Isabel's document on initial results of polarization study [ PDF ].
- SGR: Discussion of Luca's comparison of Pqpo/Pavg^2 vs. Pqpo-Pavg
[ HTML ].
- "Difference" method is 40% more sensitive than "Ratio" method. This is likely due to fluctuations of Pavg adding noise to the ratio. (Suspect that this would be diminished if longer intervals than the 0.25 seconds were used.)
- Optimization of the veto threshold. How does the sensitivity depend on the threshold level and the threshold window (100 ms vs 1 s)? Request sensitivity estimates as a function of these parameters to be part of tuning the threshold (not just livetime lost).
- Safety: investigate by injecting sine-Gaussians of the correct central frequency. Explore safety as function of hrss and Q. Expect that signals will self-veto if they have large hrss and low Q. Need safety study as part of threshold tuning, though do not expect safety to be a problem.
- Conclusion: abandon "Ratio" method for now in favor of "Difference" method + veto (unless safety or other concerns arise). Still need to tune the veto.
- Sharmilla is keeping up with code review.