RAPID enhancements in digital signal processors (DSPs) have made it possible for designers to use digital predistortion to perform real-time correction of distortion in RF power amplifiers (PAs). Unfortunately, unlike purely analog-based solutions, this technique puts stringent requirements on several components in the transmission chain, including predistortion circuits, data converters, reconstruction filters, and the RF transmitter (Tx) components (Fig. 1). These constraints arise because there must be a perfect transmission path between the predistorter and the nonlinear PA, since any errors in this path will result in increased adjacent-channel power (ACP). Although an adaptive digital predistortion ssystem can compensate for some of the nonlinearities in the forward path not attributable to the PAs, many other imperfections cannot not be overcome. Therefore, a successful implementation of digital predistortion requires an understanding of potential system-level limitations and the means of overcoming them. This article presents an empirical overview of circuit errors in non-adaptive systems. Since many of the components in the correction loop exhibit characteristics similar to those in the forward path, the information presented is also applicable to a fully adaptive, digital predistortion system.