Digital Voltage Regulator Adaptive Gain for Light-Load Stability
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Digital voltage regulators face instability at light current loads due to increased output impedance, which is not effectively addressed by existing methods that rely on voltage set points rather than load current, and require complex computing or large memory resources for adaptive gain adjustments.
Innovation Solution
Implementing an adaptive gain mechanism in the control loop of digital voltage regulators that adjusts gain based on the number of enabled current sources, using a binary shift function to reduce gain by half at lower load currents, and transitioning to pulse frequency modulation when necessary, thereby stabilizing the loop across a wide range of load conditions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If digital voltage regulators use fixed gain control, then the device complexity is low, but stability deteriorates at light current loads due to increased output impedance
Solution Approach 1:
The patent applies dynamics by transitioning from fixed gain control to adaptive gain control that dynamically adjusts the controller gain based on operating conditions. The gain is modified according to the number of enabled current sources and load current levels, allowing the system to adapt to varying load conditions and maintain stability across the full operating range.
Solution Approach 2:
The patent implements parameter changes by modifying the controller gain parameter based on the operating point. Specifically, the gain is adjusted according to the code value that determines which current sources are enabled, allowing the system to compensate for output impedance variations without requiring complex computing resources.
2Reliability
If adaptive gain adjustment uses complex computing methods, then stability improves, but the use of energy and device complexity increase
Solution Approach 1:
The patent replaces complex computational methods with a simplified digital implementation using binary shift operations. Instead of requiring floating-point arithmetic or complex algorithms, the adaptive gain is achieved through integer-based binary shifts and additions, dramatically reducing computational energy consumption while maintaining the stability benefits of adaptive control.
Solution Approach 2:
The patent uses simple digital logic operations (binary shifts and additions) that are computationally inexpensive and can be implemented with basic digital circuitry. This approach replaces energy-intensive complex computing methods with low-cost, low-power digital operations that are well-suited for integrated circuit implementation.
3Reliability
If the gain is reduced at lower load currents, then stability improves, but the control precision may deteriorate
Solution Approach 1:
The patent applies dynamics by implementing different gain values for different operating ranges. The controller dynamically selects appropriate gain levels based on the current load conditions, using higher gain when precision is critical and lower gain when stability is the primary concern, thereby balancing both requirements across the operating range.
Data Source
AI summary
Embodiments herein relate to a feedback loop in a digital voltage regulator for controlling an output voltage. To avoid instability at light current loads, a gain of the loop is reduced as a power gate code indicates a reduced number of branches in set of current sources are enabled. In an example implementation, the code is classified into one range of a number of ranges, and the gain is set based on the one range. The gain can decrease each time the code enters a lower range, as indicated by the code crossing a threshold or predetermined value. For example, the gain can decrease by half each time the code enters a lower range.