# Why do we sometimes say that we can use momentum conservation if time period is small and other we don't?

In this stack and this other stack which I had asked, there was discussion on when you can apply momentum conservation in collisions. The conclusion was that you can do it if time period of collision is small.

However recently I came across a question in which momentum conservation is not obeyed even if time period of collision is small.

So, in the set up of image above, a block of mass M is dropped from top of a ramp where it collides perfectly inelastically with a second ramp at 'B'. Supposedly, the momentum is conserved in direction along the surface and not in the direction perpendicular to it.

I was confused by this because, we used the collisions have small time period argument to argue about momentum in the first case but not here. So, why exactly can't I use the argument that collision time period is small to say that momentum should be conserved in all directions here?