Optimal placement for partially populated system EMI testing

Abstract

Electrical systems like servers, IO cages, storage appliances, and others can contain many sources of radiated emissions. During product development the demand for hardware can outpace the supply, especially when the sub-components are at the forefront of technology themselves. Necessity may require testing with less than a full configuration. If so, the available hardware should be positioned to maximize emissions rather than minimize them. It is better to overestimate when establishing margins with respect to the regulatory limits. This paper describes a method of finding optimal placements for the limited hardware using genetic algorithms, full-wave models, and far-field calculations.