2.6. Angular Positions
BATSE is capable of estimating on its own the direction to a burst. It is composed of eight detectors that are pointed towards different directions in the sky. The relative intensity of the counts in the various detectors allows us to measure the direction to the burst. The positional error of a given burst is the square root of the sum of squares of a systematic error and a statistical error. The statistical error depends on the strength of the burst. It is as large as 20° for a weak burst, and it is negligible for a strong one. The estimated systematic error (using a comparison of BATSE positions with IPN (Inter Planetary Network) localization) is 1.6° . A different analysis of this comparison [112, 113] suggests that this might be slightly higher, around 3°.
The location of a burst is determined much better using the difference in arrival time of the burst to several detectors on different satellites. Detection by two satellites limits the position to a circle on the sky. Detection by three determines the position and detection by four or more over-determines it. Even in this case the positional error depends on the strength of the bursts. The stronger the burst, the easier it is to identify a unique moment of time in the incoming signals. Clearly, the accuracy of the positional determination is better the longer the distance between the satellites. The best positions that have been obtained in this way are with the IPN3 of detectors. For 12 events the positional error boxes are of a few arc-minutes .
BeppoSAX Wide Field Camera (WFC) that covers about 5% of the sky located a few bursts within 3' (3). BeppoSAX's Narrow Field Instrument (NFI) obtained the bursts' positions to within 50". X-ray observations by ASCA and ROSAT have yielded error boxes of 30" and 10" respectively. Optical identification has led, as usual, to a localization within 1". Finally VLBI radio observation of GRB970508 has yielded a position within 200 µarcsec. The position of at least one burst is well known.