Wireless Transceiver Architecture

5 Nonlinearity

The nonlinearity of the functions implemented in a transceiver is traditionally seen as an additional source of distortion for the signal present in the line-up. In most cases, we need those functions to be as linear as possible in order to preserve the modulation characteristics of the signal being processed. But in practical electronic implementations, such linearity is always achieved at some cost. For instance, improving the behavior of an RF amplifier with respect to compression always results in an increase in the power consumption. In this particular example, there is thus necessarily a trade-off between the power consumption of the solution and a reasonable degradation of the radio link.

We thus need to review and quantify the impact on the wanted signal of the different kinds of nonlinear behavior that may occur in a practical implementation. This requires us to be able on the one hand to characterize the nonlinearity of a device in a suitable way for analytical derivations, and on the other hand to quantify its impacts, mainly in terms of SNR or EVM degradation. Only this kind of characterization can enable us to minimize the implementation cost by not over-dimensioning the requirements in terms of linearity.

However, it is also interesting to see that the general statement that the functions implemented in a transceiver need to be linear may be wrong in some particular cases. One such case is when dealing with constant amplitude bandpass RF signals. ...