move_can produces negative radii about cancrusher HOT 5 CLOSED

The bug's not fixed, but getting a track on it. The can movmenet, radius change, is incorrect.

from cancrusher.

It appears that when the current goes negative, the force rather than continuing to be compressive reverses sign. This creates a large discontinuity in the velocity of the can. Looking into how many steps of the algorithm were reported in the AJP article.

from cancrusher.

Moving the can makes the current go run away instead of ringing as before.

from cancrusher.

By step 119, the middle two radii have gone negative [ 0.01219539, 0.00294055, -0.00130072, -0.00130072, 0.00294055, 0.01219539]

from cancrusher.

The main issue was that the dbcoilbz function was not dependent on the current in the coil. Once this was fixed, the results look just about like the results from the AJP article. You still can't run further than 70 microseconds because the can has split in two by this point.

Here's a look at the can compressing:

from cancrusher.