Notes:

• 1) The fitting includes the systematic errors.
• 2) When the burst includes telescope slew time periods, the fitting uses an average response file made from multiple 5-s response files through out the slew time plus single time preiod for the non-slew times, and weighted by the total counts in the corresponding time period using addrmf. An average response file is needed becuase a spectral fit using the pre-slew DRM will introduce some errors in both a spectral shape and a normalization if the PHA file contains a lot of the slew/post-slew time interval.
• 3) For fits to more complicated models (e.g. a power-law over a cutoff power-law), the BAT team has decided to require a chi-square improvement of more than 6 for each extra dof.

Time averaged spectrum fit using the average DRM

Power-law model

Multiple GTIs in the spectrum
tstart tstop
Photon index:       (       )

Cutoff power-law model

Multiple GTIs in the spectrum
tstart tstop
Photon index:       (       )

1-s peak spectrum fit

Power-law model

Time interval is from   23.200 sec. to   33.270 sec.
Photon index:       (       )

Cutoff power-law model

Time interval is from   23.200 sec. to   33.270 sec.
Photon index:       (       )

Time-resolved spectra