Seagate boosts automated vision system with AI. 

To drive process improvement and cost reduction, the Factory IT team integrated AI into Seagate’s slider production’s enhanced automated vision system (EAVS). The slider operation team aimed to boost defect detection, maintain throughput, and implement advanced image processing.

Since the 2010s, manufacturing vision systems have rapidly advanced, with AI integration offering significant benefits. In Seagate’s slider factory, EAVS machines were initially seen as limiting throughput, measured by unit per hour (UPH) and first pass yield (FPY) for air bearing surface (ABS) and poles, relative to the predetermined overhead rate (POR). The team identified image processing time as a key area for improvement to meet POR targets.

Cost was another challenge. Enhancing EAVS detection capabilities required a strong return on investment (ROI) without compromising quality. The strict “C=0” (no defects found in the accepted sample size) quality assurance (QA) standard meant a single defect in sample lots triggered full rescreening and recleaning of shipments.

The slider team set both short-term and long-term goals. Short-term goals included replacing conventional image processing with a machine learning-integrated solution to improve defect detection in the final slider inspection. Also planned was improving outgoing defective parts per million (DPPM) and lot acceptance rates (LAR) to ensure downstream slider quality. An additional aim was to not impact EAVS machine UPH throughput or costs.

Long-term goals focused on QA and traceability improvement for downstream failure analysis. For QA, the team planned to replace physical parts handling (including physical parts inspection) by using images captured with the EAVS system. Using EAVS images for in-process quality assurance (IPQA) inspection would not only help reduce the physical handling of parts, but also improve slider cleanliness. The plan to strengthen traceability was meant to improve visibility of sampled sliders and provide coherent data for failure analysis based on customer feedback.

Before AI integration, the image processing method required extensive engineering resources for template creation and maintenance. Developing new templates took weeks, and engineers continuously fine-tuned them to sustain defect detection capabilities. Rule-based image processing struggled with detecting defects in transition width and near slider edges, leading to missed defects and lower detection accuracy. Enhancing EAVS detection capabilities required a strong ROI without compromising quality.

To address these challenges, the team replaced conventional image processing with a machine learning-integrated solution. This led to upgraded defect detection and improved DPPM and LAR rates. This transition also helped to maintain EAVS machine UPH without added costs. Seagate Research Group (SRG) developed the SliveLine Bridge (SLB) as an edge device to enable real-time inferencing, chosen for its cost advantage over high-performance workstations and its alignment with the "Seagate on Seagate" strategy.

Simplified EAVS ML solution architecture.

Enlarge

Simplified EAVS ML solution architecture.

The AI-driven solution integrated multiple technologies, including a centralized machine learning (ML) operations platform, global monitoring tools, on-premises S3 storage, data loaders, a common data service, and edge device integration with equipment. This comprehensive approach ensured immediate results for disposition decisions and minimized performance gaps between original and quantized (lower precision) models. Successful implementation required collaboration across Seagate teams in model development, IT infrastructure, system integration (tools interface and factory analytics), machine software, edge hardware, model optimization, and shopfloor execution. Additional efforts included resource planning, deployment, material builds for evaluation, and quality assessment.

Seagate’s deployment of EAVS integrated with AI significantly transformed business operations, streamlining workflows and enhancing efficiency. By implementing this solution across all 34 EAVS machines, the team completed this scaling effort by November 2024, marking a major operational excellence milestone. The integration led to substantial improvements in key performance metrics:

• Higher FPY rates. ABS FPY increased from 60% to 74%, and pole FPY from 80% to 86%, reducing rework and waste.
• Cost and space savings. The company saved $2.6 million through workforce reallocation and reduced cleaning costs, as well as reclaimed 77 square meters of clean room space.

These metrics highlight the significant impact of the AI-driven solution on Seagate operations, demonstrating improvements to productivity, cost savings, and quality control. The successful implementation of AI with EAVS has not only streamlined workflows, but also provided substantial financial and operational benefits, positioning Seagate for continued success and innovation in manufacturing processes. 

Enlarge

Ongoing model monitoring and retraining ensures adaptability to new defects and advanced air bearing (AAB) designs. ML-driven inspection has replaced manual processes, improving quality and freeing engineers for strategic initiatives.

Looking ahead, the team aims to re-deploy a 100 times magnification scope to upstream processes for early detection and quality at source. Beyond EAVS, the team plans to apply similar approaches to other use cases, such as lapping plate quality assessment, diamond density analysis, and in-process imaging-based defect detection. These expansions will further advance manufacturing precision and quality control, positioning Seagate for continued innovation and operational excellence.