Spectral Regrowth

Spectral regrowth originates from the nonlinear transfer function of power amplifiers operating near saturation. When a modulated signal passes through a device with a nonlinear AM-AM (amplitude-to-amplitude) and AM-PM (amplitude-to-phase) response, the output spectrum is no longer a scaled replica of the input. The nonlinearity generates intermodulation products that spread energy into adjacent channels. Key contributors include:

• Gain compression at high instantaneous power levels
• Phase distortion that varies with envelope amplitude
• Odd-order nonlinearities, with third-order products (IMD3) being the dominant source of regrowth
• Memory effects that make the distortion frequency-dependent The severity of regrowth is directly proportional to the amplifier's back-off from its 1dB compression point (P1dB).

ACLR is the primary regulatory metric for quantifying spectral regrowth. It measures the ratio of transmitted power within the assigned channel to the power leaking into adjacent channels, expressed in dBc. Regulatory bodies such as 3GPP and FCC mandate minimum ACLR values to prevent interference. Typical requirements include:

• LTE/5G NR: -45 dBc or better for adjacent channels
• WCDMA: -45 dBc at 5 MHz offset, -50 dBc at 10 MHz offset
• Measurement performed using gated spectrum analysis with root-raised-cosine filtering
• ACLR degradation directly correlates with Error Vector Magnitude (EVM) increase in-band
• Modern digital predistortion (DPD) systems can improve ACLR by 15-25 dB

A spectral mask defines the maximum allowable power spectral density as a function of frequency offset from the carrier. It serves as a regulatory envelope that transmitter emissions must not exceed. Spectral regrowth that breaches the mask results in compliance failure. Mask characteristics include:

• In-channel region: Allows full transmit power within the occupied bandwidth
• Transition region: Specifies the required roll-off slope, typically demanding sharp filter roll-off
• Out-of-band region: Sets absolute emission limits, often as low as -60 dBm/Hz or below
• Spurious emission limits: Cover frequencies far from the carrier, protecting distant services
• Compliance verification requires power spectral density (PSD) measurements with specified resolution bandwidth

Memory effects cause a power amplifier's instantaneous nonlinear behavior to depend on past input states, making spectral regrowth frequency-dependent and asymmetric. These effects arise from:

• Thermal memory: Junction temperature variations with signal envelope, causing slow gain and phase drift
• Electrical memory: Bias circuit impedance at envelope frequencies modulating the transistor's operating point
• Trapping effects: Charge capture and release in semiconductor materials, particularly in GaN HEMT devices
• Long-term memory: Time constants from milliseconds to seconds, affecting wideband signals
• Short-term memory: Sub-microsecond effects within the signal bandwidth Memory effects produce asymmetric spectral regrowth, where the upper and lower adjacent channels exhibit different ACLR values. This asymmetry cannot be corrected by memoryless predistortion alone.

The peak-to-average power ratio (PAPR) of a modulated signal directly determines the severity of spectral regrowth. High-PAPR signals force power amplifiers to operate with significant power back-off to avoid nonlinear operation. Key relationships include:

• OFDM signals exhibit PAPR of 10-13 dB, requiring 8-10 dB back-off for acceptable linearity
• Single-carrier QAM has lower PAPR (4-7 dB), enabling more efficient amplifier operation
• Crest factor reduction (CFR) techniques reduce PAPR before amplification, improving efficiency
• Clipping distortion from insufficient back-off generates severe spectral regrowth
• Peak windowing and iterative clipping and filtering (ICF) offer softer PAPR reduction with better spectral containment The trade-off between power efficiency and spectral purity is fundamental to transmitter design.

Intermodulation distortion (IMD) is the mathematical mechanism underlying spectral regrowth. When a nonlinear device processes a modulated signal, it generates sum and difference frequency products of all spectral components. Critical IMD characteristics:

• IMD3 (third-order): Falls at 2f₁ - f₂ and 2f₂ - f₁, landing directly in adjacent channels
• IMD5 (fifth-order): Falls at 3f₁ - 2f₂, contributing to alternate channel interference
• Third-order intercept point (IP3) quantifies IMD3 generation; higher IP3 indicates better linearity
• IMD asymmetry arises from baseband impedance and memory effects
• Two-tone testing remains the classical method for characterizing IMD, though modulated stimulus is more representative of real signals Understanding IMD generation is essential for designing effective Volterra series and memory polynomial predistortion models.

The primary regulatory compliance metric for spectral regrowth. ACLR quantifies the ratio of transmitted power within an assigned channel to the power leaking into adjacent frequency channels.

• Typical requirements: -45 dBc for 5G NR base stations
• Measurement: Integrated power ratio using spectrum analyzer
• Directly degraded by AM-AM and AM-PM distortion
• Improved through digital predistortion and crest factor reduction

Nonlinear signal products generated at sum and difference frequencies when multiple signals pass through a nonlinear device. Third-order products (IMD3) are the most problematic for adjacent channel interference.

• IMD3 falls directly in adjacent channels
• Two-tone testing characterizes amplifier nonlinearity
• Third-order intercept point (IP3) extrapolates linearity from IMD measurements
• Memory effects cause frequency-dependent IMD asymmetry

A signal conditioning technique that reduces the peak-to-average power ratio (PAPR) of a transmitted waveform before amplification. CFR enables higher average power operation without clipping-induced spectral regrowth.

• Peak windowing applies smooth time-domain windows for superior spectral containment
• Clipping and filtering iteratively removes out-of-band distortion
• Tone reservation uses reserved OFDM subcarriers for peak cancellation
• Trade-off: PAPR reduction vs. in-band EVM degradation

The two fundamental nonlinear conversion mechanisms in power amplifiers that cause spectral regrowth. AM-AM distortion causes gain compression where output amplitude deviates from linearity. AM-PM distortion introduces phase shifts that vary with instantaneous signal envelope.

• AM-PM is a critical source of spectral asymmetry
• Memory effects make these conversions frequency-dependent
• Digital predistortion must compensate both simultaneously
• Doherty amplifiers exhibit complex AM-PM characteristics

Power amplifier memory effects cause the current output to depend on past input states due to thermal, electrical, and trapping dynamics. This creates frequency-dependent nonlinear behavior that complicates spectral regrowth cancellation.

• Thermal memory: Slow time constants from self-heating
• Electrical memory: Bias network and matching circuit impedance variations
• Trapping effects: Charge capture/release in GaN and GaAs devices
• Volterra series and memory polynomial models capture these dynamics

Regulatory-defined power spectral density envelopes that limit maximum allowable out-of-band emissions. Spectral masks define the ultimate pass/fail criteria for transmitter spectral regrowth performance.

• 3GPP TS 38.104 defines 5G NR base station emission limits
• FCC Part 15 governs unlicensed transmitter emissions
• Guard bands provide spectral buffers between channels
• Spurious emission (SEM) limits protect distant spectrum users