Techniques for Predicting Nonlinear Distortion Effects in Ka-band Uplink Gateways

Abstract In order to achieve optimum service availability, Ka-band gateways must deliver very large effective isotropic radiated power (EIRP) per signal. For most signal types, and for all multicarrier uplinks, uplink power is limited by distortion in the final amplifier. Such distortion may be defined by a number of metrics, including adjacent-channel power ratio (ACPR), noise power ratio (NPR), and two-tone intermodulation distortion (IMD). Accurate methods for predicting the effects of distortion are critical for sizing the gateway antenna/amplifier hardware and for planning service quality. This paper reviews the method of distortion simulation using complex envelope transformation by the AM-AM and AM-PM transfer functions. Provided that the quasi-static approximation holds, as is normally the case for traveling wave tube (TWT) amplifiers, these transfer functions fully characterize the device, and thus the simulation technique provides accurate predictions of all distortion properties, including two-tone intermodulation, noise power ratio, and singleor multicarrier modulated signal distortion. By combining the transfer functions, the effects of cascaded stages (such as upconverter-amplifier or linearizer-amplifier combinations) can be computed. The transfer functions can be measured using swept-power vector network analysis or with the spectral-null pulse technique, which also allows upconverters and amplifiers with limited continuous power to be characterized. The functions may also be generated by circuit-level simulation, or more practically, by a parameterized behavioral model. In this paper, we describe a model that allows computation of the sensitivity of distortion metrics (such as NPR and two-tone IMD) to variations in device parameters (such as peak AM-PM and gain compression knee sharpness). The paper also presents examples of how the relationships between various distortion metrics depend on the shapes of the AM-AM and AM-PM transfer functions. Theoretical limits on linearizer performance are derived, and practical methods for testing the validity of the quasi-static assumption are described.