How does digital predistortion linearize a PA?

From RF Design Wiki · By the RF Design Wiki editorial team · Published October 21, 2026 · Last reviewed October 21, 2026 · How we research

Digital predistortion (DPD) applies the inverse of the PA’s nonlinear transfer characteristic to the baseband signal so that the cascade is linear. An observation receiver samples the PA output, and an adaptive algorithm fits a behavioral model, commonly a memory polynomial or generalized memory polynomial, to capture AM-AM, AM-PM and memory effects.^[1]^[2]

What linearization methods exist?

Digital predistortion: dominant in base stations; adaptive, power efficient.
Analog/RF predistortion: diode or cubic generators; low complexity, limited correction.
Feedforward: an error amplifier cancels distortion; wideband and accurate, but inefficient.
Cartesian or polar feedback: loop around the PA; bandwidth limited by loop delay.
Crest-factor reduction: reduces PAPR before DPD to allow higher average power.

^[1]

What are memory effects?

The PA’s output depends on past inputs as well as the present one, due to bias-network impedances at the envelope frequency, thermal time constants and charge trapping (notably in GaN). Memoryless models fail for wide bandwidths; memory polynomials include delayed terms.

What does DPD require from the hardware?

An observation path with bandwidth several times the signal bandwidth (to capture intermodulation, commonly 3–5×), a DAC/transmit path with the same expanded bandwidth, and accurate time alignment. Results are judged by ACLR and EVM.

Frequently asked questions

Can DPD fix a saturated PA?

No; DPD cannot exceed saturation, so peak power must remain within capability.

Why is wider bandwidth hard for DPD?

Required sampling rates scale with signal bandwidth, and memory effects grow.

Is DPD used in handsets?

Simplified forms are increasingly used; base stations use full adaptive DPD.

References

Online sources accessed October 21, 2026; printed works cited by edition. See all sources.

“High-Linearity RF Amplifier Design”. P. B. Kenington, Artech House. 2000.
“RF Power Amplifiers for Wireless Communications, 2nd ed.”. S. C. Cripps, Artech House. 2006.

Categories: Amplifiers

Articles on RF Design Wiki

Start here	Main page What is RF design Learning path Design workflow
Fundamentals	Decibels and dBm RF power Wavelength and frequency Thermal noise Noise figure Noise temperature SNR and Eb/N0 Skin effect Q factor Nonlinearity and harmonics Intermodulation and IP3 P1dB compression Dynamic range and SFDR
Transmission lines	Transmission line theory Characteristic impedance Γ, VSWR and return loss Input impedance of a line Quarter-wave transformer Microstrip Stripline Coplanar waveguide Coaxial cable Waveguides RF connectors
Matching and S-parameters	Impedance matching Maximum power transfer Smith chart L-network matching Pi and T networks Stub matching Bode–Fano limit S-parameters Touchstone files Z, Y, ABCD parameters De-embedding Mixed-mode S-parameters
Filters and passives	RF filter basics Filter approximations LC filter design Microstrip filters Cavity and dielectric filters SAW and BAW filters Diplexers and duplexers Wilkinson divider Couplers and hybrids Baluns Circulators and isolators RF attenuators Passive components
Amplifiers	Gain definitions Amplifier stability LNA design Power amplifier basics PA classes Efficiency and PAE Doherty amplifier Linearization and DPD Load-pull Transistor technologies Transistor biasing ACPR and EVM
Mixers and frequency generation	RF mixers Mixer specifications Image frequency I/Q modulators Oscillator basics Crystal oscillators VCO design Phase noise Jitter from phase noise PLL basics PLL loop filter Fractional-N synthesis Direct digital synthesis Multipliers and dividers
Antennas and propagation	Antenna fundamentals Gain and directivity Dipole antenna Monopole antenna Patch antenna Arrays and beamforming Small antenna limits Polarization Friis equation Free-space path loss Link budget Multipath and fading Atmospheric and rain loss Path loss models Fresnel zones
Receivers and systems	Receiver architectures Superheterodyne Direct conversion Low-IF receiver Software-defined radio Receiver sensitivity Blocking and selectivity Transmitter architectures Modulation ADCs for RF Radar basics
PCB layout and materials	RF PCB layout RF PCB materials Grounding and vias Controlled-impedance stackup Decoupling RF shielding Thermal design
EMC and EMI	EMC and EMI basics FCC Part 15 CISPR standards Immunity and ESD EMI filters and ferrites Precompliance testing RF exposure limits
Test and measurement	Vector network analyzer VNA calibration Spectrum analyzer Noise figure measurement Phase noise measurement RF power measurement Signal generators TDR Antenna measurement Oscilloscopes for RF
Simulation	Simulation overview Harmonic balance EM simulation EM solver methods SPICE for RF
Standards and bands	Frequency bands Radar band letters ISM bands 5G NR bands Wi-Fi and Bluetooth RF GNSS frequencies Amateur radio bands Spectrum regulation
Reference	Formula cheat sheet Textbooks and references Glossary FAQ All sources
About	About Editorial policy