Test Data Compression via Background Scan Cell Segmentation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The increasing complexity of integrated circuits leads to high test data volumes during manufacturing, resulting in increased costs due to the need for various test pattern sets for fault models, which existing technologies have not adequately addressed.
Innovation Solution
The method involves selecting background scan cells based on specified bit distribution, merging test cubes with compatible main portions, and using background chains with segment control to reduce test data volume by shifting main and background test cubes into scan chains, allowing for efficient test data compression.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If multiple test pattern sets are applied to detect various faults, then fault detection coverage is improved, but test data volume increases
Solution Approach 1:
The patent segments scan cells into background scan cells (specified by majority of test cubes) and non-background scan cells. Test cubes are divided into main portions (non-background cells) and background portions (background cells). This segmentation allows the majority-unspecified bits to be compressed using LFSR while maintaining fault detection coverage through selective reconstruction of specified bits.
Solution Approach 2:
The patent merges incompatible test cubes by combining their main portions (non-background bits) to form a parent test cube. Incremental test patterns and control patterns are merged with the parent test cube to reconstruct the original incompatible test cubes during testing, enabling compression while preserving test effectiveness.
2Quantity of substance
If test data compression is applied using LFSR encoding, then test data volume is reduced, but encoding capacity is limited
Solution Approach 1:
The patent employs dynamic reseeding where compressed test stimuli are injected at both current and previous cycles to encode the test stimulus needed at the current shift cycle. The LFSR seed is dynamically updated based on control patterns and incremental test patterns, enabling greater encoding capacity than static reseeding approaches.
Solution Approach 2:
The patent introduces background chains as an intermediary structure between the tester and scan chains. Background chains store and transfer background test cube data, acting as a mediator that enables the LFSR-based compression scheme to handle both compressed and uncompressed portions of test data efficiently.
3Quantity of substance
If vector clustering is applied to improve encoding efficiency, then test data volume is reduced, but device complexity increases
Solution Approach 1:
The patent segments background chains into multiple segments that can be independently controlled. Each segment can be bypassed or disabled separately through control signals, allowing flexible configuration and reducing the effective complexity for different test scenarios while maintaining the ability to handle large test data volumes.
Solution Approach 2:
The patent creates a universal compression architecture where background scan cells and background chains can serve multiple functions: storing background test cube data, transferring data between tester and scan chains, and being selectively activated or bypassed. This multi-functionality reduces the need for separate dedicated structures for different test requirements.
Data Source
AI summary
Background scan cells are selected from scan cells in a circuit based on specified bit distribution information for a plurality of test cubes generated for testing the circuit. A main portion and a background portion are then determined for each test cube in the plurality of test cubes. The background portion corresponds to the background scan cells. Test cubes in the plurality of test cubes that have compatible main portions are merged into test cube groups. Each test cube group in the test cube groups comprises a main test cube and background test cubes. A main test cube, supplied by a tester or a decompressor, may be shifted into the scan chains. A background test cube may be shifted into background chains and be inserted into the main test cube in the scan chains based on control signals.