This paper provides an automated design to provide robust PIDs with fixed control gains, suitable to be applied in power converters whose parameters belong to real intervals. Differently from conventional PIDs which use only a nominal model to obtain the fixed control gains and, a posteriori, verify robustness, the proposed approach ensures, a priori (i.e. during the design stage), robust performance for a set of plant parameters. To illustratethe proposed procedure, two conventional PID controllers are given, to achieve phase margin and crossover frequency for a nominal model of a buck converter. An objective function based on frequency domain specifications is proposed. A particle swarm optimization algorithm is then used to find PIDs, in a large search space that include stable and unstable controllers, allowing to optimize this function for all cases of combinations of plant parameters. A case study for the buck converter illustrates the improvements of performance with the proposed method when compared to the conventional PID controllers. Additionally, the design is used in a more challenging application, for a buck-boost converter suitable for small satellites application, becoming a simple alternative for benchmarks for robust control of power converters.