Multi-Band Adjacent Channel Leakage Ratio (MB-ACLR)

MB-ACLR is the ratio of the total power leaked into a specified adjacent channel to the total power in the desired multi-band transmit channels. The measurement is performed on the composite RF signal after the power amplifier, capturing the aggregate spectral regrowth from all carriers and their nonlinear interactions. It is typically expressed in dBc (decibels relative to the carrier).

Unlike single-band ACLR, MB-ACLR is heavily influenced by cross-band intermodulation products (IMD). When two carriers at frequencies f1 and f2 are amplified, third-order products (2f1-f2, 2f2-f1) and fifth-order products (3f1-2f2, 3f2-2f1) can fall directly into adjacent channels. MB-ACLR captures the power of these products, making it a direct measure of cross-band cancellation effectiveness.

MB-ACLR is a critical compliance metric for 3GPP 5G NR and 4G LTE-Advanced carrier aggregation scenarios. Standards bodies specify minimum ACLR requirements (typically -45 dBc or better) for each adjacent channel. For multi-band transmitters, these limits must be met simultaneously across all active carriers, which is significantly more challenging due to the increased density of distortion products.

MB-ACLR is the primary figure of merit for evaluating Multi-Band Digital Predistortion (MB-DPD) algorithms. A successful MB-DPD architecture must suppress both in-band distortion (EVM) and out-of-band spectral regrowth (MB-ACLR). The metric directly quantifies how well the predistorter cancels the nonlinear memory effects and cross-band coupling terms modeled by structures like the 2D Memory Polynomial (2D-MMP).

Measuring MB-ACLR requires a vector signal analyzer with sufficient bandwidth to capture all carriers and adjacent channels simultaneously. A key characteristic is spectral asymmetry: the ACLR in the lower and upper adjacent channels often differs due to frequency-dependent PA memory effects and bias network impedance variations. MB-ACLR specifications often require reporting both lower and upper values independently.

Multi-Band Crest Factor Reduction (MB-CFR) directly influences MB-ACLR. By reducing the peak-to-average power ratio (PAPR) of the composite multi-band signal, MB-CFR allows the power amplifier to operate with a higher average output power while maintaining the same peak power headroom. This reduces the amount of nonlinear compression, thereby improving the achievable MB-ACLR without increasing DPD complexity.

The primary physical phenomenon that degrades MB-ACLR. When multiple carriers are amplified concurrently, nonlinear mixing generates intermodulation products that fall into adjacent channels of other active bands.

• Mechanism: Third-order and fifth-order nonlinearities create spectral regrowth that overlaps with adjacent carriers
• Impact: Directly reduces the measured ACLR in each band, as leakage from one band becomes interference in another
• Measurement Challenge: Requires synchronized multi-channel spectrum analysis to isolate cross-band contributions from self-distortion

A predistortion architecture that uses a two-dimensional indexing structure based on instantaneous magnitudes of both concurrent baseband signals to synthesize correction terms.

• Key Insight: A single-dimensional DPD cannot correct cross-band distortion; the 2D model captures envelope coupling between bands
• Implementation: Uses a 2D look-up table or polynomial with cross-terms dependent on |x₁(n)| and |x₂(n)|
• MB-ACLR Improvement: Typically achieves 15-20 dB improvement in adjacent channel leakage for both bands simultaneously

The fundamental nonlinear mechanism generating unwanted frequency components when two or more signals mix in an active device. IMD products are the root cause of MB-ACLR degradation.

• Third-Order IMD (IM3): Falls at 2f₁ - f₂ and 2f₂ - f₁, often landing directly in adjacent channels
• Fifth-Order IMD (IM5): Falls at 3f₁ - 2f₂ and 3f₂ - 2f₁, affecting alternate channels
• Cross-Band IMD: Products generated by carriers in different bands falling into the adjacent channels of a third band

The signal processing procedure for estimating DPD model parameters from observed PA input and output waveforms. Joint coefficient estimation is critical for MB-ACLR optimization.

• Indirect Learning Architecture (ILA): Identifies a post-distorter from attenuated PA output, then copies coefficients to the predistorter
• Least Squares Estimation: Solves for coefficients by minimizing the error between desired and actual PA output
• Challenge: Cross-band terms dramatically increase the parameter count, requiring robust numerical methods to avoid overfitting

The broader engineering discipline focused on reducing adjacent channel leakage and improving ACLR through linearization. MB-ACLR is the multi-band extension of this single-band metric.

• Regulatory Context: 3GPP TS 38.104 defines ACLR limits for 5G NR base stations; multi-band operation adds cross-band leakage requirements
• Measurement Bandwidth: MB-ACLR measurements require wider observation bandwidth to capture leakage across all active carriers
• Trade-off: Aggressive linearization improves MB-ACLR but increases power consumption and computational complexity

Digital predistortion specifically optimized for 3GPP carrier aggregation scenarios where multiple component carriers transmit through a common power amplifier.

• Intra-Band Contiguous: Carriers within same band, simpler DPD sufficient
• Intra-Band Non-Contiguous: Carriers in same band with gaps, requires memory effect compensation
• Inter-Band Non-Contiguous: Carriers in different bands, demands full multi-band DPD with cross-band terms to meet MB-ACLR targets
• Typical Configuration: 2-5 component carriers aggregated, each requiring independent ACLR compliance