EPAM instrument status:

There was a sudden increase to very high levels in the lowest energy channels
in one of the 7 EPAM solid state detectors.  The time of onset of the noise
was DOY 78 at 23:54:56 UT.  To the best of our understanding, this noise
increase is in the detector and not the electronics.  After careful
deliberation, a command was issued to change the F strobe threshold to
minimize dead time from processing noise pulses on the F detector system.  The
command was issued on Wednesday August the 12th 1998 (DOY 224) at ~13:30.  The
instrument response after the command is as expected.  It is expected EPAM
will operate in this mode indefinitely.

The noise increase started suddenly, in less than the 1.5 s readout time for a
sector. This is very unusual for solid state detectors. We usually see
failures come on slowly. We have previously seen this sudden onset only when
there was physical damage to the detector, such as the micrometeoroid damage
on Voyager 1. We believe that the problem is in the detector, but we can not
completely rule out an electronics problem in the pre-amp.

The noisy detector (detector F) points 150 degrees from the spin axis and
looks through a thin foil. It responds to electrons from about 35 keV to 400
keV and ions from 400 keV to 5 MeV. The noise in the detector is confined to
the three lowest channels (less than 150 keV).

This noise affects one of the five EPAM apertures; it is one of the three
apertures that measure electrons. All species and all energies can still be
measured from the other apertures. However, this noise prevents us from seeing
low energy electrons from the anti-Sun/Earth direction.

The science impact of this problem is a loss of low-intensity upstream
electron events. Higher intensity and higher energy events at 150 degrees from
the spin axis are unaffected. This aperture is not used in the Real Time
Solar Wind data.

The threshold in the F detector was raised so it now responds to electrons
above about 60 keV. The 60 to 100 keV channel still has a few thousand counts
per second of noise and the 100 to 180 keV channel has about 100 counts of
noise. Except for these high backgrounds, the data look correct.