This paper addresses the problem of energy detection of unknown deterministic signal of a primary user in a cognitive radio environment. As an extension to the previous works, we focus on equal gain combining technique when the wireless channel is modeled as Nakagami-m. We derive series form exact expressions for probability of detection and false alarm when the number of diversity branches are 1, 2, 3 and L ges 4. Finally, performance variation is shown against the number of diversity branches and the time bandwidth product in decision statistic with the aid of numerical results.