Gain Compression

The P1dB is the most critical figure of merit for quantifying the onset of gain compression. It is defined as the output power level at which the amplifier's actual gain has dropped by exactly 1 dB from its ideal linear small-signal gain.

• Significance: P1dB marks the transition between quasi-linear operation and the strongly nonlinear saturation region.
• Back-off reference: System designers typically back off the operating point from P1dB by a specified amount (e.g., 6-10 dB) to meet linearity requirements.
• Measurement: Determined by sweeping input power while monitoring the deviation of the gain curve from its constant small-signal value.

Gain compression manifests as two interrelated but distinct distortion mechanisms that DPD must correct simultaneously.

• AM-AM Conversion: The amplitude-dependent variation in gain. As the instantaneous input envelope increases, the amplifier's gain compresses, causing amplitude distortion that flattens waveform peaks.
• AM-PM Conversion: The amplitude-dependent variation in phase shift. In compression, the amplifier introduces an unwanted phase rotation that varies with the instantaneous signal envelope, causing spectral regrowth.
• Complex correction: Effective predistortion requires a complex-valued correction that expands amplitude while applying an inverse phase rotation.

Beyond the compression region lies hard saturation, where the output power becomes completely independent of input drive.

• Clipping mechanism: The transistor's output voltage swing reaches the supply rail limits, physically preventing further increase.
• Spectral consequences: Hard clipping generates severe spectral regrowth into adjacent channels, dramatically degrading ACLR.
• Irrecoverable distortion: Once a signal is hard-clipped, information is permanently lost. Crest Factor Reduction (CFR) intentionally clips peaks before the PA in a controlled manner, but DPD cannot recover signals already driven into hard saturation.

Gain compression is not a static, memoryless phenomenon. The amplifier's nonlinear behavior depends on the history of the signal envelope.

• Short-term memory: Electrical memory effects caused by bias network impedance variations at the modulation bandwidth. The compression characteristic shifts dynamically with envelope frequency.
• Long-term memory: Thermal memory effects where die temperature changes modulate gain over milliseconds. Self-heating during high-power bursts causes transient compression shifts.
• Modeling requirement: Accurate DPD requires memory polynomial or Volterra series models that capture both the instantaneous compression curve and its dependence on prior signal states.

Gain compression defines the fundamental engineering tension between power efficiency and signal fidelity.

• Peak efficiency: Power amplifiers achieve maximum power-added efficiency (PAE) when operated deep in compression near saturation.
• Linearity cost: Operating at peak efficiency introduces severe nonlinear distortion that violates spectral emission masks.
• Back-off penalty: Traditional linear operation requires significant power back-off from P1dB, often reducing PAE from 50%+ to below 25%.
• DPD's role: Digital predistortion enables operation closer to the compression point while maintaining linearity, recovering much of the efficiency lost to back-off.

Doherty power amplifiers exhibit a unique dual-stage compression characteristic that complicates linearization.

• Carrier amplifier: Compresses first as it approaches voltage saturation, providing the initial efficiency peak at back-off.
• Peaking amplifier: Turns on and compresses later, creating a second inflection point in the composite AM-AM curve.
• Composite nonlinearity: The combined transfer function has a distinctive double-hump gain expansion/compression profile that requires specialized DPD model structures.
• Efficiency sweet spot: Properly linearized Doherty PAs can achieve high efficiency at 6-8 dB back-off, ideal for modern signals with high PAPR.

The P1dB is the most common figure of merit for quantifying the onset of gain compression. It is defined as the output power level at which the amplifier's gain has dropped by 1 dB from its ideal linear value. This metric provides a standardized boundary between linear and nonlinear operation, serving as a critical specification for component selection and system link budget analysis.

Gain compression manifests as two distinct but simultaneous distortion mechanisms. AM-AM distortion describes the nonlinear relationship between input amplitude and output amplitude—the classic compression curve. AM-PM distortion is the unintended phase shift introduced as a function of instantaneous input power. Both must be corrected by a predistorter to fully linearize the amplifier's complex transfer characteristic.

The IP3 is a theoretical metric extrapolated from low-power two-tone measurements that predicts the amplifier's nonlinear behavior. While gain compression describes the large-signal saturation of the fundamental tone, IP3 characterizes the small-signal generation of third-order intermodulation products. A higher IP3 relative to P1dB indicates a more linear amplifier with a softer compression knee.

Not all amplifiers compress identically. Soft compression describes a gradual, progressive gain reduction that begins well before saturation, typical of class-AB amplifiers. Hard compression is characterized by an abrupt, sharp saturation knee, common in switching-mode amplifiers like class-E designs. The compression characteristic directly influences the complexity of the DPD model required to linearize the device.

Gain compression is not purely instantaneous. Memory effects cause the amplifier's current nonlinear behavior to depend on previous signal states due to thermal dynamics, bias circuit impedance, and trapping effects in semiconductor materials. This transforms a simple static compression curve into a dynamic, history-dependent nonlinear system that requires memory polynomial or Volterra series models for accurate predistortion.

The most straightforward method to avoid gain compression is power back-off—operating the amplifier at an average power level significantly below its P1dB. While this ensures linearity, it drastically reduces power-added efficiency (PAE). The economic and thermal cost of back-off is the primary motivation for adopting digital predistortion, which allows operation closer to compression while maintaining signal integrity.