I was thinking. It's kind of like an overshooting cloud top.
A negatively bouyant parcel has so much downward momentum it shoots past its equilibrium temperature and continues to the surface. Naturally, it should stop when the temps are equal, but it has momentum and continnues to the surface.
[/b]
And remember that buoyancy is force, which means that buoyancy acts to ACCELERATE parcels (we'll neglect other complications like VPPGF, water loading, turbulence, etc). So, as long as the parcel is more dense than the surroundings, it will actually continue to accelerate downward. Therefore, it should always take some time to bring it's velocity back to 0m/s given that positive buoyancy must act to provide positive acceleration (meaning to slow the parcel's downward ascent before beginning to move the parcel "upward").
In an ideal situation, therefore, a parcel should "overshoot" the tropopause to a height at which the negative buoyancy (CINH ABOVE the equilibrium level) equals the positive buoyancy (CAPE). Of course, for updrafts which have an EQ near the tropopause, there tend to be other factors which affect vertical acceleration (e.g water loading, entrainment/detrainment, turbulence, etc). So,
ideally and where buoyancy is the force considered, you should be able to find the height of an overshooting top based on the sounding... A parcel moving upward through a profile with 5000 j/kg CAPE should not reach 0 m/s (i.e. should not stop moving upward) until the parcel has ascended through 5000j/kg CINH (or -5000 j/kg CAPE if you wish). To "stop" a downward moving parcel is similar to "stopping" an overshooting top, just occurring near the ground (where there's a stable layer / inversion) rather than at and above the tropopause (where there's also an inversion).