Envelope Frequency Heating

The instantaneous power dissipated in a transistor channel is not constant but varies with the squared magnitude of the signal envelope. This creates a dynamic heat source whose frequency content is directly tied to the modulation bandwidth, not the RF carrier. When the envelope frequency falls below the device's thermal cutoff frequency, the junction temperature can track these power variations, leading to a time-varying thermal profile that modulates electrical characteristics.

A device's thermal impedance acts as a low-pass filter on the power dissipation waveform. The thermal bandwidth defines the frequency range over which junction temperature can fluctuate in response to envelope variations. Key factors include:

• Thermal time constants of the die, attach, and package layers
• Thermal capacitance of the semiconductor material
• Die attach thermal resistance as a primary bottleneck Envelope frequencies above this bandwidth are thermally filtered, while those within it cause significant dynamic temperature swings.

As junction temperature fluctuates with the envelope, critical transistor parameters shift dynamically:

• Threshold voltage decreases with rising temperature, altering the quiescent bias point
• Carrier mobility degrades, reducing transconductance and gain
• Parasitic capacitances change, introducing a dynamic phase shift This produces thermal AM-AM distortion (envelope-dependent gain compression) and thermal AM-PM distortion (envelope-dependent phase rotation), both of which exhibit memory due to the thermal lag.

Unlike short-term electrical memory effects caused by bias network resonances, envelope frequency heating creates a long-term memory spanning microseconds to milliseconds. This duration is governed by the thermal relaxation time—the characteristic time for the junction to return to equilibrium. The resulting distortion cannot be corrected by memoryless predistorters and requires specialized thermal-aware predistortion or thermal-induced memory polynomial models that incorporate the device's thermal impulse response.

The dispersive phase response of thermal memory creates an imbalance between the upper and lower intermodulation sidebands. This thermal-induced spectral asymmetry is a hallmark of envelope-driven heating effects and cannot be compensated by standard memory polynomial models that assume symmetric nonlinearity. The asymmetry arises because the thermal phase lag varies with envelope frequency, causing different phase shifts for the upper and lower distortion products relative to the carrier.

In Gallium Nitride (GaN) transistors, envelope frequency heating interacts with charge trapping mechanisms in surface states and buffer layers. Trapping is often thermally activated—elevated temperatures can accelerate detrapping time constants. This creates a complex, coupled electro-thermal memory where:

• Self-heating modifies trap occupation dynamics
• Trapped charge alters the electric field and power dissipation profile
• The combined effect produces a nonlinear thermal convolution that demands advanced electro-thermal models for accurate linearization.

A distortion mechanism where the device's temperature history, caused by signal envelope variations, alters its instantaneous electrical behavior. This creates a long-term nonlinear memory that standard memoryless predistorters cannot correct. The effect manifests as a slow drift in gain and phase response tied to the thermal time constants of the semiconductor structure.

A measure of a material's resistance to heat flow, defining the dynamic relationship between power dissipation and the resulting temperature rise. Key characteristics include:

• Units: °C/W or K/W
• Transient behavior: Governed by distributed thermal resistance and capacitance
• Measurement: Extracted via pulsed thermal response testing
• Modeling: Represented using Foster or Cauer network topologies

The characteristic time required for a device's junction temperature to reach approximately 63.2% of its steady-state value following a step change in power dissipation. This parameter directly dictates the memory duration of thermal effects. Multiple time constants exist in real devices, corresponding to different thermal interfaces:

• Die-to-substrate: microseconds
• Package-to-heat sink: milliseconds to seconds

A co-simulation technique that couples semiconductor device physics with dynamic heat generation and dissipation equations. This approach predicts temperature-dependent electrical nonlinearities by solving the heat equation concurrently with device transport equations. Essential for:

• Predicting thermal AM-AM and AM-PM distortion
• Designing thermal-aware predistortion algorithms
• Validating thermal mitigation strategies before fabrication

A behavioral model structure that augments standard memory polynomials with additional terms specifically designed to capture low-frequency, long-duration thermal lag effects. The model separates short-term electrical memory from long-term thermal memory by incorporating envelope-dependent temperature kernels. This enables predistorters to compensate for both trapping and self-heating phenomena simultaneously.

A digital linearization technique that incorporates real-time temperature sensing or electro-thermal models into the predistorter. This compensates for dynamically shifting amplifier nonlinearities by:

• Indexing correction coefficients by estimated junction temperature
• Applying thermal convolution to predict instantaneous thermal state
• Adapting LUT contents based on thermal boundary condition changes Critical for GaN-based transmitters where self-heating is pronounced.