Hiroyoshi Ishikawa

A highly efficient adaptive digital predistortion amplifier for IMT-2000 base stations

We have developed the worlds most compact, energy-efficient transmitter amplifier for IMT-2000 base station systems that uses an adaptive digital predistorter (DPD). The power efficiency is improved by reducing the output back-off (the margin from the saturation point) of power amplifiers and by suppressing the resulting increase in nonlinear distortion with the DPD. This method uses a predistortion technique that estimates the distortion characteristics of amplifiers by using the least mean square (LMS) algorithm, and it adds a complementary signal to the digital baseband signal with a reverse-characteristic of the distortion. The power amplifiers we have developed achieve twice the power efficiency of conventional types. In this paper, we describe the DPD equipment and the experimental results of a DPD transmitter prototype. Currently, IMT-2000 base station systems equipped with the DPD are being shipped worldwide.

A novel adaptive digital predistortion for wideband power amplifiers with memory effects

In order to provide broadband communication, the signal bandwidth has continuously increased in existing 4G and will keep increasing in upcoming 5G communication systems. Therefore, as the bandwidth of the transmitting signal increases, memory effects arise in high power amplifiers (HPAs). A GaN Doherty HPA is known as the device that can amplify the RF signals with high efficiency. However, the performances of the linearizer are severely degraded due to memory effects in GaN active devices as the bandwidth of the input signal increases. This paper proposes a novel scheme that improves the linearization performance of the digital predistortion (DPD) for wideband transmitting signals with 100 MHz continuous bandwidth and 3.5 GHz GaN Doherty HPA. It is shown that the knowledge of the phase trajectory between adjacent transmitting samples gives an extra degree of freedom for DPD coefficient selection compared with traditional approaches and results in enhancement of the linearization performance. Thus, the proposed DPD based on the memory polynomial is able to compensate for strong nonlinearity with memory in GaN Doherty HPA by considering the phase shift between adjacent transmitting samples. The experimental results obtained for a transmitting signal with 100 MHz continuous bandwidth show that the proposed approach achieves an 8 dB adjacent channel leakage ratio (ACLR) performance improvement over standard memory polynomial DPD.