Crest Factor Reduction (CFR)

The most direct method for reducing PAPR. The signal envelope is hard-limited to a threshold, which generates severe out-of-band spectral regrowth. A subsequent frequency-domain filter removes the distortion in adjacent channels.

• Hard Clipping: Simple amplitude truncation; causes sharp discontinuities and high ACLR.
• Iterative Clipping and Filtering (ICF): Repeatedly clips and filters the signal to progressively converge on a lower PAPR while controlling regrowth.
• Trade-off: Low complexity but introduces significant in-band distortion (EVM degradation) that must be corrected by the DPD system.

Instead of abruptly truncating peaks, peak windowing multiplies the signal by a smooth window function (e.g., Gaussian, Kaiser, or raised-cosine) centered around each peak exceeding the threshold.

• Spectral Containment: The smooth envelope transition produces far better ACLR than hard clipping.
• Implementation: Requires detecting peaks and applying overlapping window functions, which can be pipelined efficiently in FPGA fabric.
• Design Parameter: The window length controls the trade-off between spectral regrowth suppression and the duration of the induced distortion pulse.

A sophisticated technique that subtracts a pre-computed cancellation pulse from the signal at each detected peak location. The pulse is designed to cancel the peak while minimizing spectral regrowth.

• Pulse Design: The cancellation pulse is typically a filtered impulse or a sinc-like function whose spectrum is confined to the signal's occupied bandwidth.
• Multi-Stage Architectures: Cascaded pulse injection stages can target different peak amplitude ranges for finer control.
• Advantage: Offers excellent PAPR reduction with minimal out-of-band leakage, making it ideal for strict spectral mask compliance.

A distortionless technique specific to OFDM systems. A subset of subcarriers is reserved and does not carry data. A peak-canceling signal is computed in the time domain using only these reserved tones.

• Mechanism: The cancellation signal is orthogonal to the data subcarriers, so it reduces PAPR without introducing in-band distortion or EVM degradation.
• Computational Cost: Requires solving a convex optimization problem or using gradient-based iterative algorithms to find the optimal cancellation signal.
• Overhead: Reserving tones reduces net data throughput, typically by 1-5% of available subcarriers.

Another distortionless method for QAM-modulated OFDM. Outer constellation points are intelligently moved outward within their decision boundaries to create a peak-canceling signal.

• Principle: Only constellation points that can be extended without crossing decision thresholds are modified, ensuring zero symbol error rate degradation.
• Iterative Projection: The algorithm iteratively clips the time-domain signal and projects the correction back onto the allowable constellation extension region.
• Limitation: Effectiveness diminishes for dense constellations (e.g., 256-QAM) where extension margins are small.

A non-uniform quantization technique borrowed from speech processing. The signal envelope is compressed at the transmitter (reducing PAPR) and expanded at the receiver.

• μ-law Companding: Applies a logarithmic compression curve to signal amplitudes, reducing the dynamic range.
• Distortion Penalty: Unlike TR or ACE, companding introduces signal distortion that degrades BER, especially at low SNR.
• Application: Best suited for systems where receiver-side expansion can be coordinated, such as proprietary point-to-point links rather than broadcast standards.

The fundamental metric that CFR targets. PAPR quantifies the ratio of a signal's instantaneous peak power to its average power, expressed in dB. Modern modulation schemes like OFDM inherently produce high PAPR signals (often 8-13 dB), forcing power amplifiers to operate with significant power back-off to avoid nonlinear distortion. Without CFR, the amplifier must be sized for rare peaks rather than average power, severely degrading power efficiency.

The primary nonlinear impairment that CFR prevents. When a power amplifier is driven beyond its saturation point, waveform peaks are abruptly truncated—a process called hard clipping. This generates severe out-of-band spectral components that violate emission masks. CFR techniques like peak windowing replace hard clipping with smooth amplitude limiting, trading controlled in-band Error Vector Magnitude (EVM) degradation for dramatically reduced spectral regrowth.

The regulatory compliance metric that CFR directly improves. ACLR measures the ratio of transmitted power within an assigned channel to power leaking into adjacent channels. 3GPP specifications mandate minimum ACLR values (typically 45 dB for base stations). CFR enables higher average transmit power while maintaining ACLR compliance by reducing peaks before they reach the nonlinear amplifier region, preventing the spectral regrowth that degrades this critical measurement.

A sophisticated CFR algorithm that applies a smooth time-domain windowing function to signal peaks exceeding a predefined threshold. Unlike hard clipping, peak windowing multiplies the signal around each peak by a shaped envelope (e.g., Gaussian, Kaiser, or raised-cosine windows), producing softer transitions with superior spectral containment. The window duration trades off between PAPR reduction effectiveness and introduced in-band distortion that degrades EVM.

The modulation quality metric that trades off against CFR performance. EVM measures the vector difference between ideal constellation points and actual transmitted symbols. CFR techniques intentionally introduce controlled in-band distortion to reduce peaks, which directly increases EVM. System designers must balance:

The traditional but inefficient alternative to CFR. Without peak reduction, power amplifiers must operate at average power levels far below their 1dB compression point (P1dB) to accommodate signal peaks without clipping. This back-off directly reduces power-added efficiency (PAE) from a potential 40-50% down to 15-25%. CFR recovers much of this lost efficiency by reducing the peak excursions that necessitate back-off, enabling operation closer to compression.