Cross-Band Cancellation

The core mechanism involves generating a correction signal that is equal in amplitude but 180 degrees out of phase with the predicted cross-band distortion product. When this synthesized anti-signal is injected into the transmit path, destructive interference occurs, effectively nullifying the unwanted spectral regrowth. This requires precise magnitude and phase alignment across the entire bandwidth of the distortion product.

Effective cancellation must account for cross-band memory effects, where the nonlinear behavior in one frequency band is influenced by the past envelope history of a signal in a different band. This is caused by:

• Thermal dynamics: Die temperature changes with aggregate signal power
• Bias circuit modulation: Shared DC supply impedance coupling
• Charge trapping: Semiconductor carrier capture and release

Models like the 2D Memory Polynomial (2D-MMP) incorporate cross-band envelope lag terms to predict and cancel these time-dependent interactions.

Cross-band cancellation specifically targets inter-band intermodulation distortion (IMD) products that fall in the frequency gaps between transmit bands or overlap with adjacent carriers. Unlike conventional DPD that only corrects in-band distortion, this technique addresses:

• Lower IMD3: 2f₁ - f₂ products
• Upper IMD3: 2f₂ - f₁ products
• Cross-modulation: Envelope transfer between bands

This is critical for carrier aggregation scenarios where guard bands are minimal.

For hardware-efficient real-time cancellation, a 2D Look-Up Table (2D-LUT) is commonly employed. The table is indexed by a two-dimensional address derived from the instantaneous magnitudes of both concurrent baseband signals: |x₁(n)| and |x₂(n)|. Each table entry stores a complex gain correction value. Adaptive update mechanisms refresh these entries using:

• Least Mean Squares (LMS) algorithms
• Recursive Prediction Error methods
• Linear interpolation between table entries for smooth correction

Accurate cancellation depends on joint coefficient estimation, where all predistorter parameters—including cross-band coupling terms—are identified simultaneously in a single optimization step. This contrasts with sequential estimation, which can leave residual distortion. The Multi-Band Indirect Learning Architecture (MB-ILA) is a common closed-loop method: a post-distorter is trained on the attenuated PA output, and its coefficients are copied to the forward-path predistorter.

The primary performance metric for cross-band cancellation is Multi-Band Adjacent Channel Leakage Ratio (MB-ACLR). Effective cancellation can achieve:

• 15-20 dB improvement in inter-band ACLR
• Compliance with 3GPP spectral emission masks for carrier aggregation
• Reduced guard band requirements, increasing spectral efficiency

This metric is measured independently for each carrier and for the inter-band gap regions where cross-band IMD products fall.

The physical phenomenon that necessitates cancellation. When multiple carrier signals are amplified concurrently, the power amplifier's nonlinearity generates intermodulation products that fall directly on top of or spectrally adjacent to the desired transmit bands. These unwanted emissions, known as cross-band distortion, cannot be filtered out and must be actively cancelled by a predistorter that generates an inverse-phase replica.

A foundational behavioral model for synthesizing the cross-band cancellation signal. The 2D-MMP extends the standard memory polynomial by incorporating cross-terms that depend on the instantaneous envelope magnitudes of both concurrent bands. This captures the critical interaction where the signal in Band 1 influences the distortion generated in Band 2, enabling the predistorter to generate the precise anti-phase signal required for neutralization.

A predistortion topology where a single, unified predistorter block processes the composite multi-band signal before upconversion. This architecture inherently performs cross-band cancellation because the predistorter sees the full signal envelope and can generate correction terms that account for all inter-band interactions simultaneously, rather than treating each band independently.

The root cause of the problem that cross-band cancellation solves. IMD products are generated at frequencies that are integer combinations of the input carrier frequencies. In a dual-band system, third-order intermodulation (IM3) and fifth-order intermodulation (IM5) products are the most problematic, often falling directly into the adjacent receive band or the other transmit band, requiring precise cancellation.

A closed-loop adaptation method for estimating the coefficients of a cross-band predistorter. The MB-ILA identifies a post-distorter model from the attenuated PA output and then copies it to the forward path. This architecture is particularly effective for cross-band cancellation because it trains on the actual composite output, capturing all inter-band distortion products without requiring a priori knowledge of the amplifier's nonlinear characteristics.

A complex long-term memory phenomenon where the nonlinear behavior in one frequency band is influenced by the past envelope history of a signal in a different band. Thermal and electrical memory effects can cause the distortion in Band 1 to depend on the signal that was present in Band 2 microseconds earlier. Effective cross-band cancellation must account for these lagging interactions using models with memory depth and cross-band envelope coupling terms.