Multi-Band Crest Factor Reduction (MB-CFR)

Unlike independent per-band CFR, MB-CFR operates on the composite baseband signal before upconversion. It identifies peaks in the combined time-domain waveform and generates cancellation pulses that simultaneously reduce the crest factor across all constituent carriers.

• Composite envelope detection: Monitors the instantaneous magnitude of the summed multi-band signal
• Coordinated pulse injection: Applies spectrally shaped cancellation pulses aligned to each carrier's frequency offset
• Cross-band awareness: Prevents peak regrowth in one band when cancelling a peak in another

PAPR quantifies the ratio between the instantaneous peak power and the average power of a transmitted signal. High PAPR forces power amplifiers to operate with significant back-off from their compression point, drastically reducing efficiency.

• Multi-band penalty: Concurrent carriers produce higher composite PAPR than single-carrier signals due to constructive phase alignment
• dB scale: Expressed as 10·log₁₀(P_peak / P_avg)
• Efficiency impact: Every 1 dB of PAPR reduction can translate to 1-2% improvement in PA efficiency

The core MB-CFR algorithm applies amplitude clipping to the composite signal followed by frequency-domain filtering to remove out-of-band distortion. The filtering is performed per-band to ensure spectral mask compliance.

• Iterative clipping: Multiple stages of soft clipping with progressively tighter thresholds
• Per-band spectral shaping: Filters aligned to each carrier's center frequency and bandwidth
• Error signal processing: Extracts and conditions only the clipping distortion, preserving the original signal structure

Instead of hard clipping, peak windowing multiplies the signal by a time-domain window function centered on each detected peak. This produces spectrally contained cancellation with minimal out-of-band leakage.

• Window functions: Kaiser, Hamming, or custom-designed windows with controlled spectral roll-off
• Overlapping window management: Handles closely spaced peaks without excessive cancellation overlap
• Configurable window length: Shorter windows preserve more signal fidelity; longer windows provide better spectral containment

MB-CFR introduces a deliberate in-band distortion that degrades modulation accuracy. The design challenge is balancing PAPR reduction against EVM budget allocation across all carriers.

• Per-carrier EVM allocation: Different modulation schemes (QPSK, 256-QAM) have different EVM tolerances
• Joint optimization: MB-CFR distributes distortion across bands to meet each carrier's EVM limit
• DPD co-design: The residual distortion from CFR must be correctable by the downstream digital predistorter

MB-CFR is typically implemented in FPGA or ASIC fabric within the digital front-end, operating at high sample rates to process wideband composite signals. Resource efficiency is critical.

• Polyphase architectures: Reduce clock rate requirements through parallel processing paths
• CORDIC-based magnitude computation: Efficient rectangular-to-polar conversion for envelope detection
• Coordinated with DPD: CFR precedes DPD in the transmit chain; the predistorter must not undo CFR's peak reduction

The fundamental metric that MB-CFR seeks to minimize. PAPR is the ratio of the peak signal power to the average signal power, expressed in dB. High PAPR signals force power amplifiers to operate with significant back-off from their compression point, drastically reducing power efficiency. In multi-band scenarios, the composite signal's PAPR is typically higher than any individual carrier due to constructive phase alignment. MB-CFR algorithms must reduce this composite PAPR without causing excessive Error Vector Magnitude (EVM) degradation or spectral regrowth.

A primary source of spectral regrowth that MB-CFR indirectly mitigates. IMD products are generated when multiple carriers pass through a nonlinear amplifier, creating sum and difference frequencies. In multi-band transmitters, cross-band IMD can fall directly into adjacent receive bands, causing desensitization. By reducing the peak excursions that drive the amplifier into its nonlinear region, MB-CFR lowers the overall IMD floor. However, the CFR algorithm itself can generate clipping noise, which must be shaped to fall outside critical frequency bands.

A linearization architecture specifically designed for 3GPP carrier aggregation scenarios where multiple component carriers are transmitted simultaneously. MB-CFR is a prerequisite for these systems because the composite aggregated waveform exhibits extreme PAPR characteristics. The CFR must handle non-contiguous intra-band and inter-band carrier configurations while preserving the EVM budget for each individual component carrier. Joint CFR and DPD optimization is critical to meet the stringent ACLR requirements of carrier aggregation without sacrificing power efficiency.

An advanced efficiency enhancement technique where the power amplifier's drain bias voltage dynamically tracks the instantaneous signal envelope. MB-CFR plays a crucial role in envelope tracking systems by limiting the envelope bandwidth and peak excursions. The CFR algorithm must be designed with awareness of the envelope tracker's slew rate and bandwidth limitations. Excessive peak reduction can actually reduce the efficiency gains of envelope tracking by flattening the envelope variation that the tracker exploits.

The ultimate regulatory compliance objective. Spectral regrowth is the broadening of the transmitted spectrum caused by amplifier nonlinearity, leading to Adjacent Channel Leakage Ratio (ACLR) violations. MB-CFR reduces regrowth by preventing the amplifier from operating in its highly nonlinear saturation region. Modern CFR algorithms employ peak windowing and pulse cancellation techniques that shape the clipping noise spectrum to minimize out-of-band emissions, effectively acting as a first stage of spectral regrowth control before DPD.