TransientPropagation: Remove Bricolage

This removes some bricolage that initially was added. We should discuss if this either should be removed (this MR) or if we need to extend the logic. The background is: during development of TransientPropagation I have noticed that when looking at individual charge carriers, there is a high chance for them to immediately leave the sensor when deposited close to the surface (low electric field, diffusion in random direction).

This creates some issue for those carriers deposited very close to the readout implant which are supposed to drift all the way to the backside. They are expected to contribute significantly (drifting away from a ~1 weighting potential to ~0 weighting potential) to the signal, but are captured immediately.

However, the current logic simply avoids sensing them out of the sensor, which is really not a nice solution to this. Furthermore, the carrier type is hard-coded. :/

Input? Comments? Ideas?

Some demonstration:

(example simulation with single cmsp1 detector, electric field taken from example init file in repo, weighting potential generated via tool in repo)

master (with bricolage)

this branch (clean)