Baseband Equivalent

The complex baseband representation is a lowpass equivalent signal that captures both amplitude and phase modulation information while omitting the high-frequency carrier. It is mathematically derived by shifting the spectrum of a bandpass signal down to zero frequency.

• Represents a real bandpass signal as a complex-valued lowpass signal
• The real part is the in-phase (I) component, the imaginary part is the quadrature (Q) component
• Eliminates the carrier frequency from analysis, dramatically reducing simulation sample rates
• Essential for modeling AM-AM and AM-PM distortion without simulating the RF carrier

Amplitude-to-Amplitude (AM-AM) distortion describes the nonlinear relationship between the instantaneous input signal amplitude and the output signal amplitude of a power amplifier. It causes gain compression at high power levels and gain expansion in some amplifier classes.

• Quantified by the deviation from an ideal linear gain curve
• The 1 dB compression point (P1dB) is a key metric marking the onset of significant nonlinearity
• In baseband equivalent models, AM-AM distortion is captured by the magnitude response of the nonlinear transfer function
• Directly contributes to spectral regrowth and in-band signal degradation

Amplitude-to-Phase (AM-PM) distortion is the nonlinear conversion of input signal amplitude variations into output phase shifts. This impairment is particularly critical in spectrally efficient modulation schemes like QAM and OFDM.

• Causes constellation rotation that varies with instantaneous signal power
• Cannot be corrected by simple gain adjustment; requires phase compensation
• In baseband equivalent models, AM-PM is captured by the phase response of the complex nonlinear transfer function
• Becomes more pronounced as amplifiers approach saturation

Memory effects describe the dependence of a power amplifier's current output on past input values, not just the instantaneous input. These effects arise from thermal, electrical (bias network), and trapping phenomena in semiconductor devices.

• Short-term memory: Electrical effects from bias circuit impedance and matching networks, typically in the nanosecond to microsecond range
• Long-term memory: Thermal effects and charge trapping with time constants from microseconds to milliseconds
• Memory effects cause the amplifier's nonlinear behavior to become frequency-dependent
• Baseband equivalent models must incorporate memory through Volterra series, memory polynomials, or neural network architectures with temporal processing

A behavioral model is a 'black-box' mathematical framework that maps input signals to output signals of a nonlinear device based purely on observed data, without requiring knowledge of internal device physics.

• Contrasts with physical models that simulate semiconductor equations
• Built by fitting model parameters to measured input-output data pairs
• Operates entirely in the baseband equivalent domain for computational efficiency
• Common structures include memory polynomials, Volterra series, and neural networks
• Enables rapid simulation and predistorter design without proprietary device information

Spectral regrowth is the appearance of unwanted frequency components in adjacent channels caused by the intermodulation distortion of a band-limited signal passing through a nonlinear power amplifier.

• Quantified by the Adjacent Channel Power Ratio (ACPR) or Adjacent Channel Leakage Ratio (ACLR)
• Results from the nonlinear mixing of in-band frequency components that creates out-of-band products
• Baseband equivalent modeling must accurately predict spectral regrowth to design effective digital predistortion
• Regulated by bodies like the FCC and 3GPP to prevent interference with neighboring channels