Statistical Circuit Robustness Analysis Separating Global and Local Variations

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Integrated circuit manufacturing faces challenges in predicting and controlling the robustness of circuit modules due to uncontrollable local variations, which can lead to failures in modules like SRAM cells, despite global variations being manageable.

Innovation Solution

A system and method that differentiate between global and local variations in statistical analysis, using separate evaluations to model the behavior of circuit modules and determine their robustness, incorporating global variation data and local variation data to simulate and predict the impact of local variations on module performance.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Manufacturing precision

If global variations are controlled through process monitoring, then manufacturing precision is improved, but local variations remain uncontrollable and cause module failures

Engineering Contradiction:

Improveglobal variation controlVSAvoidmodule robustness

Core Design Contradiction:

Manufacturing precisionVSReliability

Solution Approach 1:

The patent segments the statistical analysis into two distinct components: global variation analysis (affecting all elements uniformly) and local variation analysis (affecting individual elements). By separating these variations and analyzing them independently with appropriate statistical methods, the system can control global variations through process monitoring while accounting for local variations through statistical prediction, thereby improving both manufacturing precision and module reliability assessment.

Inventive Principle:

Principle #1Segmentation

2Device complexity

If statistical analysis treats all variations uniformly, then analysis simplicity is maintained, but prediction accuracy of local variation impact deteriorates

Engineering Contradiction:

Improveanalysis complexityVSAvoidyield prediction accuracy

Core Design Contradiction:

Device complexityVSMeasurement precision

Solution Approach 1:

The statistical analysis is divided into separate global and local components. Global variation statistics are computed across all modules, while local variation statistics are computed individually for each module. This segmentation enables accurate prediction of local variation impact on yield without requiring overly complex unified analysis, balancing simplicity and accuracy.

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

The patent applies full statistical analysis only where necessary - global variation analysis is performed once for all modules, while local variation analysis is performed selectively for individual modules based on their specific characteristics. This partial application of detailed analysis reduces overall complexity while maintaining prediction accuracy for critical cases.

Inventive Principle:

Principle #16Partial or excessive action

Data Source

PatentUS8380478B2Statistical evaluation of circuit robustness separating local and global variation

Publication Date: 2013.02.19 TEXAS INSTRUMENTS INC

AI summary

Semiconductor fabrication using statistical analysis (400) to determine the robustness or reliability of a fabricated integrated circuit module given global and local variations of operating parameters of elements, such as transistors, of the module. Multiple sequences of statistical simulations (408, 414) are run to ascertain (416) the robustness of the module to local variations in an environment of global variations.