Scan Chain Diagnostic Accuracy Using HVM Functional Testing
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for diagnosing broken scan chains in high-volume manufacturing (HVM) are inefficient, leading to poor 'one-cell' resolution and complex physical failure isolation, which hampers process yield improvement.
Innovation Solution
A scan chain diagnostic system that utilizes existing HVM functional testing content to run chain diagnosis before wafer fabrication, enabling physical failure analysis and improving diagnosis resolution without hardware changes at the wafer or die-package level.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If existing HVM functional testing content is used for chain diagnosis before wafer fabrication, then diagnosis resolution is improved, but hardware modifications are avoided
Solution Approach 1:
The patent applies preliminary action by performing chain diagnosis before wafer fabrication using existing HVM functional testing content. This allows broken cells to be identified and isolated during the design phase rather than requiring hardware modifications after fabrication, thereby improving diagnosis resolution while avoiding costly hardware changes.
Solution Approach 2:
The patent uses copying by leveraging existing HVM functional testing content and scan chain structures to perform diagnostic functions. Instead of creating new dedicated diagnostic hardware, the system copies and reuses existing testing infrastructure to achieve chain diagnosis, improving measurement precision without adding hardware complexity.
2Measurement precision
If broken scan chains are diagnosed using conventional methods, then physical failure isolation is achieved, but diagnostic accuracy is poor
Solution Approach 1:
The patent applies segmentation by dividing the scan chain into individual cells that can be independently diagnosed. By segmenting the chain and using functional testing content to target specific cells, the system achieves one-cell resolution and improves diagnostic accuracy, enabling precise identification of broken cells without relying on conventional low-accuracy methods.
Solution Approach 2:
The patent uses feedback by analyzing the results of functional testing to identify broken cells in the scan chain. The system feeds back diagnostic information from the functional testing content execution to determine which cells are broken, thereby improving diagnostic accuracy through iterative testing and analysis rather than relying on conventional single-pass methods.
3Measurement precision
If hardware modifications are made to improve diagnosis resolution, then one-cell resolution is achieved, but manufacturing cost increases
Solution Approach 1:
The patent applies copying by reusing existing HVM functional testing content and scan chain infrastructure to achieve one-cell resolution. Instead of adding new dedicated diagnostic hardware that would increase manufacturing cost, the system copies and leverages existing testing structures, thereby maintaining ease of manufacture while improving measurement precision.
Solution Approach 2:
The patent applies universality by making the existing HVM functional testing content serve dual purposes: both functional testing and chain diagnosis. This multi-functionality allows the system to achieve one-cell resolution without adding separate diagnostic hardware, thereby avoiding increased manufacturing costs while improving diagnostic capabilities.
Data Source
AI summary
This disclosure describes systems, methods, and devices related to diagnose a broken scan chain and isolate the broken cell. A device may perform a functional test on a plurality of central processing unit (CPU) cells in a chain. The device may propagate data through a combinatoric logic. The device may capture results in sequential flip-flops associated with scan. The device may utilize the results in a next combinatoric logic. The device may utilize shifted-out data to isolate a first broken cell based on the functional test.