Some RHESSI detector gains changed noticeably following the anneal.
A post-anneal gain calibration was done but was not properly
propagated in the Solarsoft tree as the default; this was recently
discovered by Iain Hannah.  Thus, spectroscopy in the front segments
-- particularly 6 and 9 -- may be off by up to 1 or 2 keV if done
on flares that occurred since the anneal (November 2007) and were
analyzed before March 18, 2008. The new gain solution was put in
its proper place in the Solarsoft tree on March 17.  Spectroscopy
performed before March 18 on recent flares should be redone.
Further improvements to the post-anneal gain solution (running from
November to the present) are intended to appear within the next three
weeks, and will be announced via another email to this list.  If you
are not sure whether to proceed with any reanalysis now or wait for
the next refinements, contact David Smith (dsmith@scipp.ucsc.edu).