Digital Predistortion
Understanding DPD
DPD is the enabling technology that allows modern wireless systems to achieve both high efficiency and high linearity. Without DPD, a PA must operate 8-10 dB below its compression point to meet linearity requirements for 64-QAM or 256-QAM OFDM signals. With DPD, it can operate 3-5 dB closer to compression, nearly doubling efficiency.
How DPD Works
- Observe the PA output via a coupler and feedback receiver.
- Compare the output to the ideal (pre-PA) signal to extract the distortion model.
- Apply the inverse distortion to the digital input signal.
- The PA distortion and pre-distortion cancel, producing a linear output.
- Continuously adapt the model to track PA changes with temperature and aging.
DPD Impact
- ACLR improvement: 15-25 dB typical.
- Efficiency gain: 2-3x improvement by operating closer to compression.
- EVM improvement: 5-15 dB.
Frequently Asked Questions
What is DPD?
DPD pre-distorts the digital input signal with the inverse of the PA distortion so that the PA output is linear. This allows operating the PA closer to saturation for higher efficiency while meeting linearity specs. It typically improves ACLR by 15-25 dB.
Why is DPD needed?
Modern wideband signals (OFDM, QAM) have high peak-to-average power ratios (8-12 dB). Without DPD, the PA must be backed off far below compression to maintain linearity, wasting efficiency. DPD allows near-compression operation while maintaining spectral mask compliance.
Where is DPD implemented?
DPD is implemented in the digital baseband processor (DSP, FPGA, or dedicated DPD IC). It requires a feedback receiver path from the PA output for adaptation. Every modern cellular base station uses DPD. It is also increasingly used in small cells and military radios.