A problem has just been identified that affects the normalization of
some RHESSI images and feature-based spectra.

The problem is that up to this point all imaging fluxes have been
computed assuming the instrument had the thin shutter in.  That
attenuator state had been hard coded into the relevant software with a
note to correct it once a method was available to determine the shutter
state.  Although the method became available some time ago, the note was
overlooked.  I have just uploaded a fix for this and the fluxes will now
be computed correctly for the shutter position at the start of an image
accumulation.

What does this mean for any results you have obtained up to now?

The morphology and location of your images are unaffected.
Spatially-integrated spectra that were obtained using SPEX are
unaffected. If your observations were only in atten_state 1 (thin
shutter in, thick shutter out), then your results are unaffected.  If
you obtained fluxes from integrating over image kernels with both
attenuators out (state 0) then the results will be dramatically
decreased at the lowest energies.  If the data were obtained with both
attenuators in (atten_state 3) then the new flux will be greater than
previously obtained at lower energies.  In these cases the normalization
of individual images are in error by an energy-dependent factor.  This
means that each point in feature-based spectra (derived, by using the
IMSPEC tool for example) will be off by a different factor and so the
feature-based spectra for observations with attenuator states 0 or 3
will be incorrect.

For example, analyses of 20-feb-2002 are not affected since RHESSI was
in atten_state 1. For April 21 the start and end of the flare are in
atten_state 0 so your fluxes will decrease at low energy.  For the July
23 flare after 00:26 UT the atten_state shifts from 1 to 3.  The
atten_state 3 photon fluxes will increase at low energy.  At energies
above 50-60 keV the differences between the attenuator state
transmissions become less important so spectral slopes extending from
low to high energy based on image kernels in attenuator states 0 and 3
should also be recalculated.

The attenuator states are indicated by bars at the top of the more
recent quicklook plots.  If you are worried about making changes to your
galley proofs for the Solar Physics papers you may want to consult
either richard.schwartz@gsfc.nasa.gov or ghurford@ssl.berkeley.edu about
the magnitude of the effects.

There will be a slight difference in your image morphology in attenuator
states 0 and 3 because there is a small difference in the attenuator
transmission function between the different detectors and those
differences vary with the attenuator state.

As always, you should avoid making images (or accumulating spectra) that
integrate across changes in attenuator or decimation states.

The correction has been made to hsi_calib_eventlist__define.pro (in the
process method) and it has been uploaded to hxrbs to the atest directory
of the hessi software.  You will receive it the next time your software
is updated.

My apologies

I'm falling on my sword now,

Richard