Understanding electrical energy consumption in a robotic system leads to the ability to minimize energy consumption for a given task. This is particularly important for mobile robots and redundant manipulators where extended operating times and non-optimized movement patterns lead to increased operating costs. However, current research shows conflicting formulas for predicting energy consumption in robotic joints driven by DC motors, specifically when negative work is involved. A breakdown of energy consumption for DC motors is introduced with respect to different operating states and phases of positive and negative work. Additionally, the energy consumption of a two degree of freedom manipulator is simulated and verified experimentally. The same task - lowering the manipulator from point a to b in a vertical line - is completed in both elbow up and elbow down configurations to illustrate the difference in energy consumption during a task that consists of mostly negative work. Finally, this energy expenditure equation is extended to a multi degree of freedom simulated humanoid robot to demonstrate validity and generality.