We describe a technique for obtaining the minimum distance vector from a query point to a conic curve segment, and evaluate its performance and precision. Salient features of our approach include a symmetry-based means for guaranteeing sound results, a bracket-restricted Newton solver that guarantees convergence, and a query-locality-based optimization for improving performance.