Transconductance Bias Circuit With Feedback Calibration in Nanometer CMOS
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional bias circuits for RF/analog integrated circuits fail to maintain constant transconductance over process and temperature variations, especially in nanometer CMOS technologies where the square-law biasing model is no longer applicable, leading to inaccuracies and increased noise.
Innovation Solution
A bias circuit that does not rely on the square-law model, utilizing a first reference current generator, replica circuits, a comparator, and digital feedback control to determine and adjust the bias current, ensuring constant transconductance while consuming less power and producing less noise.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If square-law biasing model is used, then constant transconductance can be achieved with perfect square law characteristic, but it is not applicable in nanometer CMOS technologies where I-V characteristics deviate from square law
Solution Approach 1:
The patent changes the biasing approach from relying on square-law I-V characteristics to using a feedback control system that adjusts bias current based on measured transconductance. This parameter change enables constant gm to be achieved regardless of the specific I-V characteristic model, making it applicable to nanometer CMOS technologies where square-law does not hold.
Solution Approach 2:
The patent implements a feedback mechanism where transconductance is measured and compared against a target value, and the bias current is adjusted accordingly. This closed-loop feedback system ensures constant transconductance without relying on theoretical models like square-law, thereby achieving both reliability and adaptability to different CMOS technologies.
2Reliability
If conventional bias circuits are used, then circuit operation is maintained, but transconductance varies over process and temperature variations
Solution Approach 1:
The patent employs feedback control where the measured transconductance is continuously compared to a target value, and the bias current is adjusted to maintain constant gm. This feedback mechanism compensates for process and temperature variations, ensuring both reliability and stability of transconductance across operating conditions.
Solution Approach 2:
The bias circuit performs self-adjustment by measuring its own transconductance and automatically correcting the bias current to maintain the target gm value. This self-service capability enables the circuit to maintain constant transconductance without external intervention, achieving both reliability and stability.
3Speed
If short-channel transistors are used for high ft, then unity current gain cutoff frequency is improved, but I-V characteristics no longer follow square law and cannot be easily modeled
Solution Approach 1:
The patent uses feedback control to measure the actual transconductance of short-channel transistors and adjust the bias current accordingly. This approach bypasses the need for complex I-V characteristic modeling by directly measuring and controlling the relevant parameter, enabling high ft performance while maintaining constant transconductance through empirical adjustment rather than theoretical modeling.
4Reliability
If standard constant-gm bias circuit is used, then constant transconductance is achieved with perfect square law, but power consumption and noise are not optimized
Solution Approach 1:
The patent implements a self-adjusting bias circuit that measures its own transconductance and automatically corrects the bias current to maintain target gm. This self-service approach optimizes power consumption by adjusting the bias current precisely to the minimum required for constant gm, rather than using fixed conservative biasing, while also minimizing noise through optimized operating points.
Data Source
AI summary
Methods, circuits and apparatus for biasing an amplifier to maintain consistent operational characteristics over variations in fabrication processes and operational temperature conditions are disclosed. A bias is determined by first comparing output voltages of replica circuits of the amplifier during an offset canceling phase. The output voltages are differently driven by an offset induced by a first reference current and the offset is canceled in response to the first comparing step. The output voltages are secondly compared during a calibration phase and a calibration bias current is adjusted in response to the second comparing step. The amplifier bias is determined based on the calibration bias current. The process is periodically repeated in response to operational variations.