System tests for Abins are modified to calculate up to second-order
excitations only (as third and fourth-order are currently disabled.)

In addition, the test data for two other tests needs updating;
BezeneBinWidthCASTEP and TolueneScale have small changes in the
results. This is associated with a more conservative criterion for
discarding small values compared to the previous system; results
similar to the previous setup can be obtained by modifying the
threshold used in
SPowderSemiEmpiricalCalculator._calculate_s_over_threshold