Adjacent Channel Leakage Ratio (ACLR)

The primary physical mechanisms degrading ACLR. AM-AM distortion compresses the signal envelope at high instantaneous power, creating spectral regrowth shoulders. AM-PM distortion introduces input-amplitude-dependent phase shifts that asymmetrically distort the spectrum. In Doherty amplifiers, the carrier-to-peaking transition creates a pronounced nonlinearity 'kink' where both AM-AM and AM-PM conversion spike, generating significant adjacent channel leakage. Digital predistortion must independently model and invert both distortion components to restore linearity.

Modern communication signals (OFDM, 5G NR) exhibit PAPR values of 8-12 dB, forcing power amplifiers to operate at significant back-off from saturation. Higher PAPR pushes the signal envelope deeper into the nonlinear region during peaks while spending most time at lower power. This dynamic range stress exacerbates ACLR because the amplifier must handle both linear low-power operation and nonlinear peak clipping. Crest factor reduction techniques are often applied before the PA to reduce PAPR and ease the ACLR compliance burden.

Dynamic nonlinearities where the amplifier's output depends on past signal states, not just the instantaneous input. Electrical memory effects arise from bias network impedance variations and envelope frequency-dependent matching. Thermal memory effects stem from self-heating in GaN HEMT transistors, causing slow gain and phase drift. Trapping effects in semiconductor materials introduce low-frequency dispersion. These memory mechanisms create asymmetric spectral regrowth that cannot be corrected by static (memoryless) predistortion, requiring Volterra-series or memory polynomial DPD models.

The active load-pull mechanism central to Doherty efficiency directly impacts ACLR. As the peaking amplifier transitions from off-state (Class-C bias) to active conduction, the impedance presented to the carrier amplifier changes dynamically. This load modulation trajectory introduces a complex, power-dependent nonlinearity profile. Gain mismatch between carrier and peaking paths, phase misalignment at the combiner, and imperfect impedance inverter design all distort the ideal load modulation, creating additional intermodulation products that leak into adjacent channels.

Wider signal bandwidths (e.g., 100 MHz for 5G NR carriers) increase ACLR challenges through multiple mechanisms. Broader bandwidths expose frequency-dependent AM-AM/AM-PM characteristics across the channel. Memory effects become more pronounced as the modulation bandwidth approaches the inverse of thermal and trapping time constants. Multi-carrier and carrier aggregation configurations generate cross-modulation products that fall into adjacent channels. The DPD linearization bandwidth must typically be 3-5x the signal bandwidth to capture and cancel third and fifth-order intermodulation distortion.

ACLR is sensitive to the amplifier's quiescent operating point. Gate bias voltage determines the conduction angle and class of operation—shifting from Class-AB toward Class-A improves linearity at the cost of efficiency. Drain voltage variations alter the saturated output power and compression characteristics. Junction temperature changes from ambient conditions and self-heating shift the transistor's gain, threshold voltage, and trapping behavior. These environmental factors require adaptive DPD systems that continuously update predistortion coefficients to maintain ACLR compliance across operating conditions.

The nonlinear generation of out-of-band frequency components caused by AM-AM and AM-PM distortion in a power amplifier. When a digitally modulated signal passes through a nonlinear device, the spectral sidebands regenerate, spilling power into adjacent channels. This is the physical phenomenon that ACLR quantifies.

• Primary cause: Third-order intermodulation products
• Mitigated by: Digital Predistortion (DPD) and amplifier back-off
• Critical in: OFDM signals with high PAPR

A modulation quality metric measuring the vector difference between the ideal reference constellation point and the actual transmitted symbol. While ACLR measures out-of-channel interference, EVM measures in-channel distortion. Both degrade simultaneously due to amplifier nonlinearity.

• Expressed as: Percentage (%) or decibels (dB)
• Relationship: Poor EVM often correlates with poor ACLR
• Standards: 3GPP specifies EVM limits per modulation scheme (e.g., 3.5% for 64QAM)

The ratio of the instantaneous peak power to the long-term average power of a communication signal. High PAPR forces power amplifiers to operate at significant output back-off (OBO) to avoid clipping. This back-off reduces efficiency, but insufficient back-off causes severe spectral regrowth and ACLR degradation.

• Typical values: 8-13 dB for OFDM signals
• Impact: Higher PAPR demands more linear (and less efficient) amplifier operation
• Mitigation: Crest Factor Reduction (CFR) techniques

The two fundamental nonlinear transfer characteristics of a power amplifier. AM-AM distortion is the nonlinear relationship between input envelope magnitude and output envelope magnitude (gain compression). AM-PM distortion is the input-envelope-dependent phase shift. Both generate intermodulation products that cause spectral regrowth and degrade ACLR.

• AM-AM measured via: Gain compression curve
• AM-PM measured via: Phase shift vs. input power
• DPD corrects: Both simultaneously with inverse nonlinearity

Dynamic nonlinear distortions where the amplifier's current output depends on past signal values, not just the instantaneous input. Thermal memory effects (self-heating) and electrical memory effects (bias network impedance, trapping) create hysteresis in AM-AM/AM-PM curves. These effects broaden the spectral regrowth shoulders and make ACLR harder to correct with memoryless DPD.

• Time constants: Nanoseconds (electrical) to milliseconds (thermal)
• Modeling: Requires Volterra series or memory polynomial models
• Critical for: Wideband signals where memory is pronounced

The primary linearization technique used to improve ACLR by applying an inverse nonlinearity to the signal before the power amplifier. A predistorter intentionally distorts the input signal such that the cascade of predistorter and amplifier behaves linearly. Modern DPD uses adaptive learning architectures (indirect or direct) to track changes due to temperature, aging, and frequency.

• Improvement: Typically 15-25 dB ACLR reduction
• Implementation: FPGA or ASIC in the transmit path
• Adaptation: Real-time coefficient updates via observation receiver