Abstract
This paper presents an adaptive digital predistorter (DPD) that significantly reduces the baseband distortion of digital pulse width modulation (PWM), typically used in switching (class-D) amplifiers. A generalized Hammerstein structure (also known as power filter) composed by static nonlinearities and FIR filters is used to model the baseband behavior of PWM. We show that the contribution of the higher order terms of the nonlinearity are negligible and therefore, for practical applications, only the first three or four odd powers should be retained. The convergence of the DPD is studied and the performance is demonstrated and compared with other approaches. Good results are obtained for typical signals even when using low-order FIR filters in the DPD. Measurement results obtained using a digital signal processor are also presented aiming to validate the proposed approach.
Original language | English |
---|---|
Pages (from-to) | 562-571 |
Number of pages | 10 |
Journal | Signal Processing |
Volume | 120 |
DOIs | |
Publication status | Published - 1 Mar 2016 |
MoE publication type | A1 Journal article-refereed |
Keywords
- Baseband
- Digital predistorter
- Indirect learning
- Pulse width modulation