3 SXC flight operation

During normal flight operations, the CCDs are read out in the "fast'' mode described above. Each row of CCD data is examined for X-ray events (the method is very similar to the one used with the MIT-developed X-ray CCD instrument on the ASCA satellite), and a "running difference'' correlation map is continuously calculated. This map is calculated as the difference between the correlation map using the last N photons and one using N photons detected $\sim 2N$ photons earlier, where N is typically 1000. This method has enormous advantages: first, the contribution of steady background sources is continuously subtracted, and, second, there is always a current "best answer''. Thus, if a GRB is detected on board, any statistically-significant peak in the running-difference correlation map is likely to be due to an X-ray counterpart to the GRB.

If there is no statistically-significant peak in the cross-correlation function, the SXC alone will not be able to provide a precise localization for the GRB. However, a peak which is not statistically significant when the entire correlation map is examined becomes more significant if only a fraction of the map containing that peak is examined. We plan to use localization information given by the WXM on board in real time to reduce the region of the correlation map examined: in this way, smaller peaks in the specified region may become statistically significant and give us a more precise localization.

The aspect of each SXC is measured using the boresight cameras comounted with the SXC CCDs. The boresight camera pixel size is $\sim 1'$, but measurement of many stellar images allows the orientation to be calculated to a precision of a few arc-seconds. Since the boresight camera and SXC are co-mounted, we expect the relation between the SXC and optical coordinate systems to be highly stable.