Pitch-Angle Comparisons

At the science team meeting in February 1999 at Caltech, I showed some plots illustrating the way that variations in the pitch angle seen by the instruments from orbit to orbit could cause major changes in the flux observed at particular values of L, depending on which longitude sector was being traversed in a given orbit. Xinlin Li gave me a list of periods when he had data from several satellites and when SAMPEX was orbiting at different local times and in various attitude-control modes; I have assembled here a comparison of pitch angles and HILT high-resolution SSD1 observations in four of these periods, and PET ELO and EHI data for two (plus the whole set, including magnetic field variations, for two additional periods of interest to Xinlin, 98236-98246 and 99127-99137). HILT hi-res SSD1 has a 1 MeV electronic threshold, corresponding to about 1 MeV for electrons and 5 MeV for protons; the PET channels are corrected using the PHA events to be a "pure" electron signal as much as possible, but as I showed at the science team meeting, there remains some proton contamination of both rates in the SAA. Also, I have been unable to make a satisfactory livetime correction in the outer zone for much of the early part of the mission, so I only include the PET plots for the last two time periods below.

The following table provides links to the pages for the four time periods (plus two later additions), and lists the attitude control mode and orbital MLT in each period. (I had intended to include examples from the "j-perp with ram avoidance" attitude control strategy, but HILT was off for much of this period and it wouldn't look much different from periods under the "j-perp without ram avoidance" strategy anyway.) For examples of the variation of SAMPEX attitude under differing conditions of ACS strategy and MLT, there are movies of the spacecraft's behavior elsewhere on this site. On each of the pages linked below are eight or sixteen (or twenty) plots in four sets, one set for each of the traversals of L from 1 to 8 (or 8 to 1) in the four quadrants of the orbit as defined by the magnetic latitude. Each set of plots contains L vs. time "spectrograms" for the pitch angle (with 180°-90° folded over onto 0°-90°, and missing data plotted as -10°) and for the logarithm of the HILT countrate, and the last three also contain "spectrograms" for the cleaned PET ELO and EHI event rates (and the two extra time periods also display the magnetic field magnitude).

There is no commentary on the pages linked above, since all illustrate pretty much the same point: there is a strong longitudinal dependence of the outer-zone as well as the inner-zone countrates. The inner-zone variation is expected, since at low L the SAMPEX orbit only reaches B / B₀ small enough to see trapped particles in the South Atlantic Anomaly region. However, there is also a noticeable variation in the outer-zone electron observations due to this effect; this is seen in that the outer zone typically appears "hotter" in the southern hemisphere around the longitudes where the inner-zone response (L of about 1 to 2.3) is seen, regardless of variations in the pitch angle. In addition, though, particularly when comparing the ascending and descending passes through the same hemisphere at the same time when one leg is exposed to a much different set of pitch angles than the other (for example, the southern hemisphere passes on 93304-93314), it is clear that the observed pitch angle also modulates the observed flux strongly.

In conclusion, we need not look for any physical reason for why Dan, Shri, and Xinlin sometimes observe one or both hemispheres' outer-zone "ovals" to be incomplete; there is plenty of variation due to the attitude of the spacecraft to explain this. I am not sure how we could reduce or eliminate this artifact; extending the "j-perp" constraint to higher values of B (higher latitudes) would help, but at high L this would interfere with studies of precipitating particles (for which you want to be looking up the field lines rather than normal to them). The effect is not specific to the "j-perp" modes, as it is observed even under the original pointing strategy. I suspect, though, that since all longitudes are scanned at two opposed local times each day (twelve hours apart) and since the "j-perp" mode and its associated more violent jerking back and forth result in variations that depend on both longitude and local time, the daily-averaged hemispheric plots produced by Shri Kanekal simply were able to average out the differences between the two local-time samples of each longitude under the original pointing mode, but this doesn't work under the "j-perp" modes.

Back to top of Data Pages

new 21 April 1999, revised 13 September 1999