Tombstone Example tomb-F for the Surface Evolver

[Click for the tomb-F.fe datafile in a second window.]

This datafile adds rotation about the lower front edge of the chip. It is derived from tomb-E.fe, so the chip contact is represented by facets. This is done for confidence in energy and volume calculations. When the contact facets are removed in tomb-FF.fe, we can compare the numbers to make sure tomb-FF.fe is implemented correctly.

Tilted to 30 degrees.	Lowered to pad.
Evolved to 9516 facets.	Undercutting flaw at top corner.

Notable features:

• The tilt of the chip is specified by the parameter tilt, in degrees.
• The chip moving commands are read from the file "change_hairy.cmd", which contains a more complicated method of interpolation between the chip and pad. Interpolation in z simply won't work for a tilted chip, particularly when we lower the front edge to rest on the pad. Hence the interpolation attribute dist is calculated as the ratio of distance to the pad to the sum of distances to the pad and the chip.
• The command change_tilt in the file "torque.cmd" changes the tilt. It is invoked by first setting the variable new_tilt to the new tilt angle.
• The read section uses the change_tilt command to set the initial tilt at 30 degrees. It is much easier to do it this way than change all the vertex coordinates up above in the datafile.
• There is a command calc_tilt_torque to calculate the torque, which is preserved in the variable tilt_torque.
• The chip's initial position is tilted and lowered to the pad before evolving. This is because liquid solder will pull the chip down (evidence by downward z_force). The chip-pad interpolation routine has a problem here when the front lower edge of the chip is on the pad and there are interface facets of the solder on the chip, since there are vertices with zero distance to both chip and pad. (modify this?)
• After gofar, there is a slight flaw at the top corners; the outer fluid surface penetrates into chip instead of the corner vertex pulling down, as seen in the close-up image above. No vertex is inside the chip, which makes this tough to handle with the mechanisms in place. The flaw is probably so small there is no significant effect on the forces. If it is deemed a problem, either move the corner vertex down by hand, or put in diagonal constraint plane (this is a general technique for handling a surface bumping into a ridge; have vertices on a constraint plane running through the ridgeline.)