In 2004, SGR 1806-20 underwent a period of intense and long-lasting burst activity that included the giant flare of 27 December 2004 -- the most intense extra-solar transient event ever detected at Earth. During this active episode, we routinely monitored the source with Rossi X-ray Timing Explorer and occasionally with Chandra. During the course of these observations, we identified two relatively bright bursts observed with Konus-Wind in hard X-rays that were followed by extended X-ray tails or afterglows lasting hundreds to thousands of seconds. Here, we present detailed spectral and temporal analysis of these events observed about 6 and 1.5 months prior to the 27 December 2004 Giant Flare. We find that both X-ray tails are consistent with a cooling blackbody of constant radius. These spectral results are qualitatively similar to those of the burst afterglows recorded from SGR 1900+14 and recently from SGR 1550-5418. However, the latter two sources exhibit significant increase in their pulsed X-ray intensity following the burst, while we did not detect any significant changes in the RMS pulsed amplitude during the SGR 1806-20 events. Moreover, we find that the fraction of energy partitioned to the burst (prompt energy release) and the tail (afterglow) differs by an order of magnitude between SGR 1900+14 and SGR 1806-20. We suggest that such differences can be attributed to differences in the crustal heating mechanism of these neutron stars combined with the geometry of the emitting areas.

We report on data obtained with the Chandra, XMM-Newton, Suzaku and Swift X-ray observatories, following the 2006 outburst of the Anomalous X-ray Pulsar CXO J164710.2-455216. We find no evidence for the very large glitch and rapid exponential decay as was reported previously for this source. We set a 3 sigma upper limit on any fractional frequency increase at the time of the outburst of Delta nu/nu < 1.5 x 10^{-5}. Our timing analysis, based on the longest time baseline yet, yields a spin-down rate for the pulsar that implies a surface dipolar magnetic field of ~9 x 10^{13} G, although this could be biased high by possible recovery from an undetected glitch. We also present an analysis of the source flux and spectral evolution, and find no evidence for long-term spectral relaxation post-outburst as was previously reported.