More Tests

Radius	BoundaryRadius=0.35
Radius-4	BoundaryRadius=0.35
BoundaryRadius-4	BoundaryRadius=0.35
Radius-4	BoundaryRadius=0.2		(note that this one is just wrong - R* is ~0.32)

Doesn't seem to much matter whether the radius or the boundary radius is the inner edge of the wind. Only first case has no inflow at center (still running BoundaryRadius for similar run).

1. 

2. 

3. 

4. 

Comparisons + Mass over time

White contour is the boundary of the star. The scale height of the disk (c/omega - for units of meters) is much smaller than the resolution - it's about 4 millionths of a computational unit.

Just for comparison are weighted sum of density for whole box over time, weighted sum for cylinder, and weighted sum for prism:

Y-axis is in computational units; multiply by 1.29236x10²⁰ for units of g.

3x3x1

​frames

​frames

Looks like the boundaries are still too close - increase half a CU on either side in x and y and 1 CU in z (for the strange wave pattern - is this a problem?)? Reluctant to increase the dimensions too much more - these simulations are only 2 levels of AMR, and already take ~3 days to run - but I could run a much larger simulation for a long time once, just to see the extent that we do need to get.