AI-Driven Analog Circuit Retuning for PVT Stability

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Conventional analog circuits face operational detuning due to unpredictable Process, Voltage, and Temperature (PVT) conditions, leading to failures and high costs associated with ASIC re-spins, as they cannot accurately adapt to varying conditions post-fabrication.

Innovation Solution

An Integrated Circuit (IC) with a Process, Voltage, Temperature (PVT) characteristics monitor, a tuning memory, and an Artificial Intelligence (AI) engine, where machine learning models predict and adjust the electrical characteristics of tunable components to maintain nominal values, using PVT sensors and change control bits to re-tune the circuit on-the-fly.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If conventional analog circuits are designed to operate within target PVT conditions, then the circuit meets design specifications during simulation, but the electrical characteristics drift widely when fabricated and subjected to varying PVT conditions

Engineering Contradiction:
Improvecircuit performance stabilityVSAvoidadaptation to PVT variations
Core Design Contradiction:
ReliabilityVSAdaptability or versatility

Solution Approach 1:

The patent implements dynamic retuning of analog circuit parameters during runtime based on monitored PVT conditions. The system continuously adjusts circuit characteristics to compensate for PVT variations, transforming a static circuit design into a dynamically adaptive system that maintains optimal performance across varying conditions.

Inventive Principle:
Principle #15Dynamics

Solution Approach 2:

The patent employs feedback mechanisms where PVT sensors monitor actual operating conditions and feed this information back to the retuning system. This closed-loop feedback enables the circuit to automatically adjust its parameters in response to measured PVT deviations, ensuring continuous compliance with design specifications.

Inventive Principle:
Principle #23Feedback

2Manufacturing precision

If extensive simulation testing is performed to verify all PVT conditions, then design coverage improves, but the number of simulations required becomes unmanageably large (100,000 or more)

Engineering Contradiction:
Improvedesign verification accuracyVSAvoidsimulation time
Core Design Contradiction:
Manufacturing precisionVSLoss of time

Solution Approach 1:

The patent performs retuning parameter calculations and optimal configuration determinations during the design phase, storing these pre-computed parameters in lookup tables or memory structures. This preliminary action eliminates the need for extensive runtime simulations, as the system directly applies pre-determined retuning parameters based on monitored PVT conditions.

Inventive Principle:
Principle #10Preliminary action

Solution Approach 2:

The patent creates simplified models or lookup tables that represent the complex PVT behavior and retuning relationships. Instead of performing full circuit simulations for every PVT condition, the system uses these copied representations to quickly determine appropriate retuning parameters, dramatically reducing computational requirements.

Inventive Principle:
Principle #26Copying

3Ease of manufacture

If analog circuits are designed with fixed parameters, then the design is simple and fabrication is straightforward, but the circuits cannot adapt to unpredictable PVT variations and may fail

Engineering Contradiction:
Improvefabrication simplicityVSAvoidoperational reliability under PVT variations
Core Design Contradiction:
Ease of manufactureVSReliability

Solution Approach 1:

The patent introduces tunable elements into the analog circuit design that allow parameter adjustment after fabrication. These dynamic elements enable the circuit to adapt its characteristics in response to PVT variations while maintaining a relatively simple fixed architecture for the majority of the circuit, balancing manufacturing ease with operational reliability.

Inventive Principle:
Principle #15Dynamics

Data Source

PatentUS20230153597A1Design and fabrication methods of runtime self-tuning analog integrated circuits using machine learning
Publication Date: 2023.05.18 ANALOG INTELLIGENT DESIGN INC
  • US20230153597A1 patent drawing
  • US20230153597A1 patent drawing
  • US20230153597A1 patent drawing

AI summary

An Integrated Circuit with an automatically re-tuning analog circuit is provided. The Integrated Circuit comprises (a) an analog circuit comprising a plurality of tunable components each configured to respond to a plurality of change control bits, (b) a Process, Voltage Temperature (PVT) characteristics monitor comprising a plurality of PVT sensors, (c) a tuning memory embedded with a machine learning (ML) model of the analog circuit and (d) an artificial intelligence (AI) engine configured to receive a PVT signal input from the plurality of PVT sensors and the machine learning model embedded in the tuning memory. Each tunable component is configured to change its electrical characteristics such that together each of the tunable components is enabled to retune the analog circuit to attain a predefined set of electrical characteristics.