ANN-Based Impedance Modeling for High-Speed PCB Trace Manufacturing
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
High volume manufacturing of printed circuit boards (PCBs) for high-speed serial links often results in variations in trace parameters, leading to impedance and loss values that may not meet tolerance levels, making it challenging to ensure consistent performance across all manufactured PCBs.
Innovation Solution
An Artificial Neural Network (ANN) is used to derive and model impedance and loss distributions for trace parameters, selecting specific models to ensure that all modeling points pass within tolerance levels, allowing for the precise definition of PCB parameters for manufacturing.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If high volume manufacturing is used to produce PCBs, then productivity increases, but manufacturing precision deteriorates due to variations in trace parameters
Solution Approach 1:
The patent applies preliminary action by using an Artificial Neural Network to pre-model and identify impedance corner cases before manufacturing. The system derives impedance distributions, determines corner cases (minimum, maximum, and intermediate impedance values), and establishes tolerance levels in advance. This allows the manufacturing process to proceed with high volume while having pre-established criteria to ensure precision, resolving the contradiction between productivity and manufacturing precision.
2Measurement precision
If traditional modeling methods are used, then device complexity is low, but measurement precision deteriorates in capturing manufacturing variations
Solution Approach 1:
The patent introduces an Artificial Neural Network as an intermediary between the physical PCB manufacturing process and the impedance modeling. The ANN acts as a mediator that learns from training data to predict impedance distributions and identify corner cases. This intermediary enables high measurement precision in characterizing manufacturing variations without requiring complex analytical models, thus resolving the contradiction between measurement precision and device complexity.
3Reliability
If impedance corner cases are not properly identified, then ease of manufacture improves, but reliability deteriorates due to tolerance violations
Solution Approach 1:
The patent implements feedback by using the ANN to derive impedance distributions, determine corner cases, and compare them against tolerance levels. The system provides feedback on whether the corner cases meet the minimum and maximum impedance requirements. This feedback mechanism ensures reliability by identifying potential tolerance violations before manufacturing, while the automated nature of the ANN keeps the ease of manufacture high, resolving the contradiction between reliability and ease of manufacture.
Data Source
AI summary
An information handling system includes a memory that stores code, and a processor that executes code stored in memory to derive a distribution of impedances for parameters of a trace within a printed circuit board (PCB). The processor further to determine impedance corners of the distribution of impedances, to select the impedance corners as first, second, and third trace models, and to derive first, second, and third distribution of losses based on the first, second, and third trace models. The processor further to store loss corners of the first, second, and third distribution of losses as modeling points, and to determine whether all of modeling points pass within tolerance levels of loss and impedance of the trace.