Class AB Transconductor Linearization for Low-Noise I/Q Modulators
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing I/Q modulator designs based on Gilbert cell mixers suffer from high shot noise and inefficiency due to the use of class A transconductor stages, which are dictated by worst-case peak current requirements, leading to nonlinear outputs and spectral splatter, especially in telecommunication modes like WCDMA and CDMA where peak-to-average-ratio (PAR) varies.
Innovation Solution
A low noise, linearized class AB transconductor stage with embedded noise filtering and negative feedback, utilizing a common source amplifier stage with constant common-mode gate-to-source voltage biasing and field effect transistors following a square law characteristic, along with a dominant pole low-pass filter and optional resistive digital step attenuator to reduce noise and power consumption.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If a class A transconductor stage is used to ensure sufficient DC current for worst-case peak requirements, then linearity is maintained, but shot noise increases and power efficiency decreases
Solution Approach 1:
The patent transitions from a static class A biasing scheme to a dynamic class AB biasing scheme where the bias current is modulated according to the signal envelope. This allows the transconductor to maintain sufficient current during peak signals for linearity while reducing current during low-signal periods to minimize shot noise and power consumption.
Solution Approach 2:
The patent changes the operating parameters of the transconductor stage by implementing envelope tracking of the bias current. The DC bias current is dynamically adjusted to match the envelope of the modulating signal, allowing the system to operate at optimal current levels for each signal condition rather than maintaining a constant high current level.
2Power
If a class A transconductor stage is used with high DC current to meet peak requirements, then maximum signal current capability is achieved, but power consumption increases
Solution Approach 1:
The patent implements dynamic biasing where the DC current supplied to the transconductor stage varies in real-time according to the envelope of the modulating signal. This ensures that high current is only supplied when actually needed for peak signal conditions, while power consumption is reduced during low-signal periods.
Solution Approach 2:
The bias current is modulated periodically in sync with the signal envelope, creating a time-varying current supply that matches the instantaneous power requirements of the transconductor stage, thereby reducing average power consumption while maintaining peak current capability.
3Reliability
If a class A transconductor stage is used with high quiescent current, then sufficient current margin is available, but noise filtering becomes less effective
Solution Approach 1:
The patent reduces the quiescent current by implementing class AB operation with envelope tracking, which dynamically adjusts the bias current to match signal requirements. This lower, variable quiescent current improves the effectiveness of noise filtering mechanisms while maintaining sufficient current margin during peak signal conditions.
Data Source
AI summary
A transconductor stage is a linearized class AB amplifier having embedded noise filtering that enables a biasing of an in-phase/quadrature (I/Q) modulator core with a low quiescent current. Linearization of the transconductor stage is increased by introducing a small amount of negative feedback into the transconductor stage via a feedback circuitry and an error amplifier. A dominant open loop pole in a path between the error amplifier and an output stage of the transconductor stage forms a dominant pole low-pass filter. A low-pass filter transfer function created when a loop including the feedback circuitry is closed attenuates noise introduced by baseband circuitry that supplies baseband signals to the transconductor stage. A master output stage biases a plurality of slave output stages that are in parallel with the master output stage. Each slave output stage is coupled to an individual modulator core such as a Gilbert cell mixer core.