GCN Circular 3805

D. C. Morris, D. N. Burrows, J. A. Kennea, J. L. Racusin (PSU), and N. 
Gehrels (GSFC) report on behalf of the Swift XRT team:

We have analysed the first five orbits of data for GRB050814 (GCN 3799
Loading...
 
 , 
Retter et al., 2005).  Using xrtcentoid, the refined position is:

RA(J2000) = 17h 36m 45.7s
Dec(J2000) = +46d 20' 20.5"

with an uncertainty of 8 arcsec. This is 5.4 arcsec from the original 
XRT position (GCN 3800
Loading...
 
 , Morris et al., 2005). The initially distributed 
XRT position was based on the first orbit of PC data and may have been 
effected by a hot column at the position of the source. The refined 
position is calculated by analyzing several orbits of PC data, each 
taken at slightly different positions on the detector, so that the hot 
column in the first orbit of data is no longer a problem for the 
centroiding algorithm.

The XRT began taking data at 11:41:15UT, 138 seconds after the BAT 
trigger. Both the windowed timing and photon counting data from the 
first orbit show a fading lightcurve consistent with a decay index of 
~4. Photon counting data from subsequent orbits shows a flattening of 
the lightcurve near T+1000s to a decay index of ~0.1.

The windowed timing spectrum (data from +165s to +383 after the BAT 
trigger) is well fit (reduced chi-sq of 1.22 for 72 dof) by a power law 
with neutral hydrogen somewhat larger than the galactic column density 
of 2.57e20:

gamma=2.1 � 0.1
NH=6.94e20 � 2e18

The photon counting spectrum (from +383s through orbit 5) is well fit 
(reduced chi-sq of 0.61 for 31 dof) by a power law also with excess 
absorption:

gamma=1.8 �  0.2
NH=7.82e20 �  7e18

The count rate at 20000s after the trigger is ~0.15 cts/s which converts 
to an unabsorbed flux of 4.76e-12 ergs cm^-2 s^-1.

It should be noted that the early versions of the SDC data contained 
processing errors, likely due to earthlimb contamination, which may yet 
be contributing to some peculiarities in the PC lightcurve.