1: \begin{abstract}

2: We introduce a family of various finite volume discretization schemes

3: for the Fokker--Planck operator, which are characterized by different

4: weight functions on the edges. This family particularly includes the

5: well-established Scharfetter--Gummel discretization as well as the

6: recently developed square-root approximation (SQRA) scheme. We motivate

7: this family of discretizations both from the numerical and the modeling

8: point of view and provide a uniform consistency and error analysis.

9: Our main results state that the convergence order primarily depends

10: on the quality of the mesh and in second place on the quality of the

11: weights. We show by numerical experiments that for small gradients

12: the choice of the optimal representative of the discretization family

13: is highly non-trivial while for large gradients the Scharfetter--Gummel

14: scheme stands out compared to the others.

15: 

16: \end{abstract}

17: