Sequential Decompressor FIFO Shift Register Scan Data Reuse
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing test pattern generation methods, such as XOR decompressors, are inefficient in handling varying numbers of care bits across scan slices, leading to unused scan data and inefficiencies, particularly when there are more care bits than available scan data, and struggle to build equations in a predictable manner for independent scan channels.
Innovation Solution
A sequential decompressor system using a first-in, first-out (FIFO) shift register with a XOR decompressor network is introduced, which updates bits in the linear shift register and feeds outputs to a second XOR network to solve for care bits in each scan slice independently, allowing for overscan cycles to prime the system without additional hardware or control pins, and utilizing a multi-stage decompressor network with a FIFO linear shift register to generate unique and predictable equations.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Quantity of substance
If XOR decompressors are used to compress test data, then test data volume is reduced, but scan data cannot be fully utilized when care bits exceed available scan data
Solution Approach 1:
The patent changes the fundamental parameter of the decompressor architecture from XOR-based to sequential decompressor with linear equations. This allows the system to dynamically adjust how scan data is utilized across multiple slices, transforming the approach from slice-by-slice independent processing to holistic scan data utilization across the entire test pattern.
Solution Approach 2:
The patent introduces a new dimension of time by utilizing scan data from multiple slices to solve for care bits in any given slice. Instead of being constrained to using only scan data from the same slice (XOR decompressor limitation), the sequential decompressor operates across the temporal dimension of multiple slices, allowing unused scan data from slices with fewer care bits to contribute to slices with more care bits.
2Productivity
If PRPG is used in sequential decompressor, then scan data is reused across slices, but equation predictability and independence deteriorates
Solution Approach 1:
The patent introduces an intermediary equation solver that acts as a mediator between the sequential decompressor and the test pattern generation. This solver systematically manages the linear equations, ensuring that scan data reuse across slices does not compromise equation independence. The intermediary layer provides predictability by methodically solving equations while maintaining the ability to control and track correlations.
Solution Approach 2:
The patent implements feedback mechanisms where the equation solver continuously monitors and adjusts the system of linear equations based on scan data correlations. This feedback loop ensures that as scan data is reused across slices, the system can detect and manage correlations, maintaining equation predictability and independence through iterative refinement of the solution process.
3Use of energy by moving object
If ATPG specifies care bits across multiple time slices to reduce toggling, then power consumption is reduced, but variance in care bits per slice increases
Solution Approach 1:
The patent applies dynamics by making the decompressor adaptive to varying care bit distributions across slices. The sequential decompressor with linear equations dynamically adjusts which scan data from which slices contributes to solving care bits in each slice. This dynamic approach maintains stability in the overall system by efficiently handling slices with varying numbers of care bits, whether high or low, without compromising power optimization or data utilization.
Data Source
AI summary
Systems and methods for a sequential decompressor which builds equations predictably provide a first-in, first out (“FIFO”) shift register which is fed by a first XOR decompressor and provides outputs to a second XOR decompressor.