A new dual-tone, signal-normalisation method for the measurement and prediction of small-signal distortion

Abstract

A new method for characterising small-signal distortion is introduced. This novel gain-compression measurement approach employs a ratio of high- and low-level tones, closely spaced in frequency, which are simultaneously measured. A theoretical analysis of two tones passing through a weakly nonlinear device under test is used to demonstrate this new procedure for recovering the gain-distortion (i.e. power series) coefficients from the dual-tone measurement data. The signal normalisation procedure is also shown to effectively remove the time-varying errors associated with the linear receiver circuitry, thereby providing a high level of measurement accuracy. Test results for two amplifiers show that the gain-distortion coefficients can be used to accurately predict the amplifier's weakly nonlinear behaviour at low input power levels for different amplitude offsets between the high- and low-level tones. Although this new technique is demonstrated by using a power-series representation to estimate the small-signal gain compression at the fundamental frequency for memoryless devices, the general techniques that are developed here for this relatively simple problem can be extended to more complex problems where a Volterra series representation is required.