The apertures and interiors of the magnetic chamber in the CRRES MEA are of the "disk-loaded" type. Electrons impact at near normal angle when they hit the collimator or the inside of the instrument. The minimum number of scatterings needed to reach a detector is 3. For the most part, electrons are very efficiently trapped by the structure in the instrument. The response to off-angle fluxes is down by at least four orders of magnitude. (Some people fly funnels for collimators---an electron can hit walls with a grazing incidence and continue on its merry way to a detector.) Experimental tests with beams of energetic electrons incident off-angle or onto the front of the instrument (to test for bremsstrahlung response) show that typical counts in a channel are less than 1 per 10**6 incident electrons. There are two background effects which do occur: a) Electrons incident on one detector can scatter back out of that detector and be focussed into another detector. This effect is small, due to the fact that the detectors are made out of silicon, which is relatively low-Z, and typical scattering events are in the forward direction. Most of the electrons that scatter out either hit the top or bottom of the chamber or are focussed back onto the detector. But some can be focussed onto higher energy channel detectors. The scattering effect (from detector to detector) will make a monoenergetic electron beam appear to have a spectrum of the form Ne=N0E(-12). Since the softest spectrum I've ever seen (in an aurora) was E(-7), this isn't normally worth considering in data analysis--the effect is there but is completely washed out by the natural flux spectrum. b) Penetrating protons (e>120Mev), cosmic rays, and electrons with e>6.5 Mev all can penetrate through the walls of the instrument and mimic a true event in the detectors. On-board processing of the signals takes care of about 95% of these. The rest are monitored and are subtracted from each channel during data analysis. There is a problem, however. In the CRRES MEA, an extra energy channel (the 1.7 MeV channel) was added to the original design and the bremsstrahlung monitor was eliminated because its telemetry was needed. I had never found a use for the bremsstrahlung data. So far so good. But the threshold on the Proton Monitor channel was lowered to include the bremsstrahlung. This was a minor disaster because the penetrating electrons (e>6.5 MeV) can show up in this monitor while they are rejected by the rest of the particle channels. Thus, in a region where there is a very high flux of very energetic electrons and very few electrons of intermediate energy, subtracting the background counts results in an over- correction. This occurs in the highest energy channels in the outer zone under certain conditions. This very energetic electron background is a minor nuisance. In the inner zone, the background is angle dependent because the tungsten collimator stops some protons which would otherwise be able to penetrate through the magnetic chamber walls and reach the background detector. The effect varies from -20% to +15% with respect to 90 degree incidence. The nominal background correction factors are correct only for 90 degree incident protons. Thus, the survey plots, which average fluxes for 2 complete spins of the spacecraft, have an over-correction for the proton background in the inner zone. This is evident in the highest energy channels, where the true electron flux is very low. **************************************************************************** NOTE -- NOTE -- NOTE -- NOTE -- NOTE -- NOTE -- NOTE -- NOTE -- NOTE -- NOTE NOTE -- NOTE -- NOTE -- NOTE -- NOTE -- NOTE -- NOTE -- NOTE -- NOTE NOTE -- NOTE -- NOTE -- NOTE -- NOTE -- NOTE -- NOTE -- NOTE -- NOTE -- NOTE **************************************************************************** I have made no attempt to produce a background correction for the very relativistic electrons produced at 2