Grid scale in time at large v||

[David Dickinson]

This is more of a comment/note than an issue or task.

Doing a similar analysis as in issue #160 but now considering the limit v|| → infinity. We then find that for fexp = 0.5

g(ig+1, it+1) = g(ig, it) - g(ig + 1, it) + g(ig, it+1)

Considering g=0 at the boundary we can see that the point next to the boundary actually gives us g(ig+1, it+1) = - g(ig+1, it), in other words we get something which is grid scale in time. This can be observed in the code when we also use nonad_zero = T (i.e. when the boundary value may not be 0).

We also know that large dg/dt at the boundary can induce grid scale in the parallel direction when v|| is small (effectively what issue #160 shows). Is it possible that our fastest particles can give a rapidly varying change in g near the boundary, inducing a rapid change in phi here, thereby leading to grid scale in theta at low v||?