Intermodulation Distortion (IMD)

IMD products are classified by their order, which is the sum of the absolute values of the harmonic coefficients. Third-order products (IM3) are the most problematic in wireless systems because they fall close to the original carrier frequencies and are difficult to filter.

• Second-order (IM2): f1 ± f2
• Third-order (IM3): 2f1 ± f2, 2f2 ± f1
• Fifth-order (IM5): 3f1 ± 2f2, 3f2 ± 2f1
• Higher odd-order products typically decrease in power but extend the distortion bandwidth.

The Third-Order Intercept Point (IP3) is a figure of merit used to characterize a device's linearity. It is the theoretical point where the extrapolated fundamental signal power equals the extrapolated third-order intermodulation product power.

• Input IP3 (IIP3): Referenced to the input power.
• Output IP3 (OIP3): Referenced to the output power.
• A higher IP3 indicates better linearity and lower IMD generation for a given output power.

IMD is the physical root cause of spectral regrowth. When a digitally modulated signal with a non-constant envelope passes through a nonlinear amplifier, the intermodulation between the signal's spectral components generates new frequency content.

• This causes the signal's bandwidth to widen, spilling power into adjacent channels.
• The result is a violation of Adjacent Channel Leakage Ratio (ACLR) limits.
• Digital predistortion (DPD) works by generating inverse IMD products to cancel this regrowth.

IMD products are categorized by their location relative to the original signal's bandwidth.

• Out-of-Band IMD: Products falling outside the transmission channel, causing interference to other users and measured by ACLR.
• In-Band IMD: Products falling within the transmission channel, corrupting the signal itself and degrading Error Vector Magnitude (EVM).
• Both types must be suppressed for high-order modulation schemes like 256-QAM and 1024-QAM.

The two-tone test is the classic method for characterizing IMD. Two continuous-wave (CW) tones of equal amplitude and close frequency spacing are injected into the device under test.

• The output spectrum reveals the fundamental tones and the generated IMD products.
• The amplitude difference between the fundamental and the IM3 products gives the IMD level in dBc.
• This test provides a controlled, repeatable measurement of a device's nonlinear transfer function.

In wideband systems, IMD generation is not static; it depends on the signal's modulation history. These memory effects cause the IMD products to become frequency-dependent and asymmetric.

• Electrical memory: Caused by bias network impedances and envelope frequency-dependent components.
• Thermal memory: Caused by dynamic transistor junction temperature changes.
• Modern DPD models, such as the Memory Polynomial, must account for these effects to cancel IMD effectively.

The direct consequence of intermodulation distortion in a communication channel. When a digitally modulated signal passes through a nonlinear power amplifier, the mixing of in-band frequencies generates new spectral components that spill into adjacent channels. This regrowth is not harmonic; it appears as a broadening of the original signal's spectrum, causing interference to other users. The severity of spectral regrowth is quantified by the Adjacent Channel Leakage Ratio (ACLR).

The primary regulatory metric for quantifying the impact of intermodulation distortion. ACLR is the ratio of the transmitted power within the assigned channel to the power leaking into an adjacent channel. Strict 3GPP standards mandate specific ACLR limits to prevent network interference.

• Typical Requirement: -45 dBc or better for 5G NR base stations.
• Measurement: Requires integrating power over the assigned channel bandwidth and comparing it to the integrated power in the offset adjacent channel.
• DPD Target: The fundamental goal of a digital predistortion system is to improve ACLR to meet compliance masks.

A theoretical figure of merit used to benchmark a component's linearity. The IP3 is the extrapolated point where the power of the fundamental tone and the third-order intermodulation product (IM3) would be equal. A higher IP3 indicates better linearity and lower IMD.

• Input IP3 (IIP3): Referenced to the input power.
• Output IP3 (OIP3): Referenced to the output power.
• Rule of Thumb: For every 1 dB increase in input power, the IM3 product increases by 3 dB, making it a critical predictor of distortion at high power levels.

A critical design constraint for wideband signal linearization. When a predistorter applies an inverse nonlinearity, the resulting signal's bandwidth expands. For IMD, the bandwidth of the distortion products is wider than the original signal.

• Third-Order IMD: The distortion bandwidth is typically 3x to 5x the original signal bandwidth.
• DPD Implication: The digital-to-analog converter (DAC) and feedback analog-to-digital converter (ADC) must operate at sampling rates high enough to capture this expanded bandwidth without aliasing distortion, often requiring rates exceeding 1 Gsps for 100 MHz signals.

A phenomenon where intermodulation distortion is not solely dependent on the instantaneous input signal but also on its past values. This causes the IMD products to become frequency-dependent and asymmetric.

• Electrical Memory: Caused by bias network impedance variations over frequency, trapping effects, and matching network dispersion.
• Thermal Memory: Caused by dynamic die temperature changes with signal envelope variations.
• Modeling Impact: Simple memoryless AM/AM and AM/PM curves are insufficient. Accurate linearization requires Volterra series or memory polynomial models to capture these dynamic effects.

The mathematical framework used to analyze IMD without simulating the high-frequency carrier. A modulated signal is represented by its in-phase (I) and quadrature (Q) components. Nonlinear analysis is performed on this complex envelope, where the carrier frequency is set to zero.

• IMD Calculation: The distortion products are calculated by applying the nonlinear transfer function directly to the complex baseband signal.
• Efficiency: This avoids the computational burden of simulating gigahertz-range sinusoidal carriers while perfectly preserving the amplitude and phase of all intermodulation products.