Technology as Foundation | Lyric's AOI Visual Inspection Equipment: The Premier Quality Inspection Solution for PCBs and Consumer Electronics

As the surging tide of AI server computing power sweeps the globe, and as foldable smartphones unfold one after another in the palm of your hand, every leap forward in high-end manufacturing places ever more exacting demands on quality inspection. On a high-end HDI board, circuit traces are measured in micrometers; on a foldable screen lens, a hairline scratch alone can render the entire device scrap. Traditional manual visual inspection, constrained by operator fatigue, variations in experience, and efficiency bottlenecks, can no longer keep pace with the cycle time and precision required by modern production lines. As a result, AOI (Automated Optical Inspection) has evolved from an optional add on to a standard necessity, becoming the first line of defense in quality control for the electronics manufacturing industry. In this industrial transformation from the "human eye" to the "AI eye," Lyric leverages its deep-rooted expertise in equipment manufacturing and self-developed vision algorithm technologies to deliver AOI appearance inspection systems that span both the PCB and consumer electronics sectors. With intelligent, high-precision, and high-efficiency inspection capabilities, Lyric builds a robust and reliable quality risk control system for product excellence.

Multi-Dimensional Vision Algorithms: Forging a Strong Technological Moat in Inspection

The core competitiveness of Lyric's AOI equipment lies in its dual‑engine technology architecture that integrates traditional vision algorithms with deep learning. The system captures product images using high‑resolution industrial cameras and multiple sets of customized optical lighting. For each inspection surface, it acquires channel information from no fewer than four imaging dimensions, combined with qualitative analysis of height and depth, to build a multi‑dimensional perception capability that far surpasses the human eye. Building on this foundation, Lyric's self developed AI deep learning model intelligently screens massive volumes of images, precisely capturing real defects while significantly reducing false positives. The system achieves an overall over kill rate of less than 3%, a miss rate as low as below 0.3%, and a 100% detection rate for critical defects such as cracks and breakages. The first pass yield per scan reaches ≥99.98%, effectively resolving the industry's long standing dilemma between over killing and missing defects.

The algorithm model can be trained using OK samples to distinguish qualified products from defective ones, and can also adopt NG samples for defect classification and position marking. The equipment establishes a full-link data traceability system, enabling one-click binding of the unique product ID, inspection images, defect data and equipment operating status. It supports seamless connection and data upload to MES systems or industrial clouds, eliminating data silos on production lines and guaranteeing traceable and reviewable data throughout the full product lifecycle. Inspection images are stored traceably for no less than one year, delivering solid data support for process optimization and quality review. The model conversion training cycle for standard products can be controlled within 7 person-days. Equipped with an in-house developed few-shot learning framework, the model can achieve rapid convergence with only a small number of defective samples, significantly cutting customers’ sample preparation costs and model commissioning cycles for product changeovers.

Universal Compatible Configuration Covering Two Major Application Fields

It is worth noting that our AOI visual inspection equipment caters to both the PCB and consumer electronics sectors. It does not mean a single unit can be deployed across different scenarios without modification. Instead, we build a differentiated product portfolio based on a unified technology foundation. Targeted adaptations to hardware platforms, optical configurations and algorithm software enable one set of core technologies to serve these two major application fields. In terms of functional configuration, Lyric's AOI leverages a common self-developed vision technology platform, demonstrating strong adaptability across all scenarios. For PCB inspection, the machine supports rigid PCBs, flexible FPCs, IC substrates, and advanced HDI boards deployed in AI servers and switches. It accurately detects dozens of process defects including open circuits, short circuits, notches, copper residues, pinholes and circuit deformation. For consumer electronics applications, the equipment targets phone cover glass, foldable screen lenses, metal middle frames, hinge structural parts, heat dissipation modules and display modules, identifying surface defects such as scratches, indentations, discoloration, bumps, dents, cracks and wrinkles. The overall equipment dimension is controlled within 6 m × 1.5 m × 2 m. Its compact footprint fits production line layouts in various cleanrooms. The delivery leadtime for standardized complete machines is merely 90 days, balancing fast delivery and flexible on-site deployment.

Boost Quality & Cut Costs Across the Whole Line to Address Pain Points of Smart Manufacturing Lines

For end manufacturers, Lyric AOI equipment effectively resolves long-standing production line pain points in the industry:

1. The inspection throughput of a single unit exceeds 30 PPM, capable of identifying over 30 types of defects. The product damage rate attributable to equipment is ≤0.05%. It greatly shortens quality inspection tact time while minimizing secondary process loss.

2. Fully automatic visual inspection replaces manual visual checking, eliminating missed detections and misjudgments stemming from human fatigue and subjective bias. It stops defective products from escaping at the source and frees labor from repetitive, high-intensity inspection work.

3. Supporting inspection of extra-large panels with full compatibility, the equipment can achieve automatic linkage with upstream and downstream processes to realize unmanned operation covering full procedures including code scanning, multi-surface inspection, and automatic sorting of good and defective products, substantially raising the automation level of production lines.

4. Built with modular non-stop maintenance architecture. Maintenance work will not halt mainline production to cut downtime losses and stabilize line OEE.

Furthermore, the built-in AI algorithm conducts real-time statistics, classification and root cause analysis on defect data. It automatically pushes optimization recommendations to upstream processes and builds a positive closed loop: Inspection – Analysis – Optimization for steady yield growth. Micro defects including short circuits, open circuits, scratches, contamination, edge chipping and circuit flaws are captured in milliseconds. Downstream rework, customer complaints and compensation risks are greatly reduced, bringing quantifiable quality and capacity benefits to customers.

Centered on self-developed AI vision algorithms and supported by precision optics and intelligent mechanical engineering, Lyric continuously raises the industry benchmarks for accuracy, efficiency and stability of domestic appearance inspection equipment, helping manufacturers build competitive advantages in micron-level quality control. As large-model vision technology further integrates with industrial production lines in the future, Lyric will keep iterating its AOI software and hardware systems. With powerful visual inspection capabilities, the company empowers the steady upgrading of the PCB and consumer electronics industrial chains, enabling every circuit board and precision structural component to meet the stringent quality standards of the global market.