Peak-to-Average Power Ratio (PAPR)

PAPR is the ratio of the instantaneous peak power to the average power of a transmitted signal over a given time interval. Mathematically, it is expressed as:

• Formula: PAPR(dB) = 10 log₁₀(P_peak / P_average)
• Complex Baseband: Calculated directly from the I/Q samples where power is proportional to I² + Q².
• Complementary Cumulative Distribution Function (CCDF): The standard statistical tool for characterizing PAPR, showing the probability that a signal's power exceeds a given threshold relative to the average.

High PAPR forces a power amplifier to operate with a large output back-off (OBO) from its compression point to avoid nonlinear distortion, severely degrading efficiency.

• Back-off Requirement: A signal with 10 dB PAPR requires the PA to operate at an average power 10 dB below its peak capability.
• Efficiency Drain: This directly translates to higher power consumption and thermal dissipation in base stations.
• Linearization Synergy: Digital Predistortion (DPD) can allow the PA to operate closer to compression, but effective PAPR reduction via Crest Factor Reduction (CFR) is often a necessary pre-processing step to maximize the combined efficiency gains.

Orthogonal Frequency Division Multiplexing (OFDM) is notoriously susceptible to high PAPR because the time-domain signal is the sum of many independently modulated subcarriers.

• Constructive Interference: When subcarriers align in phase, massive instantaneous peaks occur.
• 5G and Wi-Fi: As the foundational waveform for 4G LTE, 5G NR, and Wi-Fi, managing OFDM's PAPR is a universal challenge.
• Subcarrier Count: The theoretical maximum PAPR grows linearly with the number of subcarriers (N), though practical CCDF curves show a statistical distribution.

CFR algorithms are signal conditioning techniques applied before the power amplifier to deliberately limit the signal's peak excursions.

• Clipping and Filtering: The simplest method, which hard-limits peaks but causes in-band distortion and out-of-band spectral regrowth that must be filtered.
• Peak Windowing: Multiplies high peaks by a smooth window function (e.g., Gaussian, Kaiser) to reduce spectral splatter compared to hard clipping.
• Pulse Injection: A sophisticated method that subtracts a spectrally shaped cancellation pulse from detected peaks, offering excellent Error Vector Magnitude (EVM) vs. Adjacent Channel Leakage Ratio (ACLR) trade-offs.

Reducing PAPR is not a lossless process; it inherently introduces a degradation in signal quality that must be carefully managed.

• EVM Degradation: Clipping or windowing peaks distorts the in-band constellation, increasing the Error Vector Magnitude.
• Spectral Regrowth: Nonlinear CFR operations create intermodulation products that leak into adjacent channels, potentially violating ACLR limits.
• Design Balance: The engineering goal is to find the optimal balance point where PAPR is minimized for efficiency without exceeding the EVM budget or spectral mask requirements.

PAPR is not a single static value but a statistical property best understood through its distribution.

• CCDF Curve: The primary visualization tool. A plot with PAPR (dB) on the x-axis and probability (log scale) on the y-axis. A designer reads the PAPR value at a target probability (e.g., 10⁻⁴).
• Test Equipment: Modern vector signal analyzers and oscilloscopes automatically compute and display CCDF curves for captured waveforms.
• Design Target: A typical design target might be a PAPR of 7-8 dB at a 10⁻⁴ probability for a 5G NR signal after CFR.

A dimensionless measure of a waveform's peakiness, calculated as the ratio of the peak amplitude to the root-mean-square (RMS) value. A constant-envelope signal has a crest factor of 1 (0 dB), while a complex modulated signal can exceed 12 dB. The crest factor directly dictates the power amplifier back-off required to avoid clipping distortion, making it a primary design constraint for transmitter efficiency.