Skip to main content
Application NoteNxp

NXP AN14887: Automatic Voltage Scaling on i.MX RT700 Using ML-Based PVTS

Guide for using Machine Learning-based PVT sensors on i.MX RT700 MCUs to optimize power consumption and extend battery life through automatic voltage scaling.

View application note

Overview

This application note describes the implementation of a Machine Learning (ML) approach for Automatic Voltage Scaling (AVS) using Process, Voltage, and Temperature sensors (PVTS) on the NXP i.MX RT700 crossover MCU. The document details how the PVTS monitors timing margins in the Compute (PVTS0) and Sense (PVTS1) domains to adjust VDD1 and VDD2 supply rails dynamically. Unlike traditional fuse-based methods, the ML model enables power optimization across any operating frequency. Implementation involves a closed-loop system where the MCU communicates with an external PMIC, such as the NXP PCA9422, via I2C to reduce digital logic power consumption by an average of 20%.

Use Cases

  • Extending battery life in wearable electronics
  • Optimizing active power consumption in low-power MCU applications
  • Implementing closed-loop voltage control in digital logic domains
  • Managing PVT variations in high-performance microcontroller designs

Topics

NXP i.MX RT700
PVTS
Dynamic Voltage Scaling
DVS
Machine Learning
PCA9422
Power Optimization
Automatic Voltage Scaling
MCU
Low Power Design

Referenced Parts

PCA9422

NXP

external PMIC as the NXP PCA9422 on the i.MX RT700 EVK board.

i.MX RT700

NXP

The i.MX RT700 is a highly integrated MCU designed for low-power applications such as wearables.

PCA9422

NXP

the sense domain using the internal LDOs or using an external PMIC as the NXP PCA9422 on the i.MX RT700