Estimated reading time: 5 minutes

Happy’s Tech Talk #46: Data Management for AI and Automation

I am sure you have read about the importance of digitization and automation for the future of printed circuit fabrication. (If not, there are several excellent articles in the December 2025 edition of PCB007 Magazine). There are numerous pressures on fabricators today, including higher component density and pitches, every trace being a transmission line, the need for improved yields, and a skilled workforce shortage, with trained employees increasingly harder to find.

However, there is a predecessor challenge that remains something on which we must focus and improve to achieve digitalization. It involves the collection and management of our process data to achieve enduring performance benefits. This is often referred to as the manufacturing data dilemma. Although there are multiple data sources comprising significant data streams within PCB manufacturing, they have typically been difficult to both access and utilize because they are contained in data silos. This data is the prerequisite for data analysis, from sensing and connecting to predicting, and data analysis translated into actionable intelligence is the core of digitalization.

Data From the Factory Floor
The first link in planning your smart factory is understanding the fundamentals of “Where is all my work?” and “Who has been working on it?” This is the role of work in process (WIP) tracking or manufacturing execution systems (MES). ERP systems are also suitable.

An MES has many features that one can customize for PCB fabrication. The data that needs to be captured depends on the type of manufacturing you perform and the number of different products you support. The data can vary (Figure 3) from management and financial to technical and environmental, as well as from other sources, like utilities.

Data From Sensors
Collecting and monitoring process data for SMT assembly is easier than for PCB fabrication if the assembly equipment supports the IPC-CFX factory standard. However, many PCB processes are manual or use machines that have no available networking protocol. In these cases, you need to purchase numerous sensors or build them in your maintenance department. The most common sensors are specific gravity, temperature, conductivity, metal content, and ampere-hours for plating¹.

Once you install the sensors, you can wire them for alarms or replenishment or connect them to a data acquisition system to feed a database. For SMT assembly or PCB fabrication machines, you can connect latent equipment to networks through Seica Spa’s ShoeBox, which, aside from older SMT and fab machines, can connect to many sensors.

Data Collection and Monitoring
Equipment data collection is the foundational pillar for analyzing the effectiveness of production lines. Manufacturers capture production data, alarm information, and specific parameters in the manufacturing process. For SMT assembly, the IPC-CFX standards can help with data collection. You can then identify areas for improvement and optimize production processes. But in PCB fabrication, there are no established standards. Programmable logic controllers (PLCs) use protocols such as PROFINET (Process Field Network), Modbus TCP, or OPC-UA, along with the SEMI process standard of SECSII/GEM. Once established, detailed data analysis provides valuable insights into production performance, facilitating continuous improvement and enhanced operational efficiency.

Condition Monitoring
As an important part of the smart factory, condition monitoring involves real-time tracking of the equipment’s performance through IIoT data, allowing real-time assessment of the equipment’s health. By analyzing this data, you can quickly detect performance deviations and act to minimize failures and yield losses. Condition monitoring enables predictive-proactive maintenance actions, giving you equipment that operates at peak efficiency and reduces unplanned downtime.

Traceability
Quality parameters monitoring: Quality control is a key focus for manufacturing operations, maintaining detailed records of production parameters to trace quality issues to specific production situations. Monitoring parameters such as process line, temperature, current, and voltage, and part specifications, enables manufacturers to identify the root causes of defects and take corrective actions to maintain product quality standards. There may also be qualitative variables that contribute to defects.

Alarms management: Alarms play a critical role in traceability and quality, highlighting issues that require immediate attention. By analyzing alarm data, manufacturers can implement preventive measures to avoid recurring problems and maintain consistent product quality. Effective alarm management ensures immediate responses to potential issues, minimizing downtime and disruption to production operations.

Energy Management 
Predictive maintenance electrical (PME) energy consumption is a challenge in manufacturing. Energy management systems monitor energy usage through various types of equipment and processes using sophisticated models to predict and optimize energy consumption, thereby contributing to overall operational efficiency and cost savings. By managing energy consumption effectively, manufacturers can reduce their environmental impact and achieve sustainability goals.

Remote Support
Remote support capabilities enable manufacturers to monitor and manage equipment performance and facilitate quick diagnostics and interventions without onsite visits. It also minimizes downtime, reduces maintenance costs, and ensures continuous production operations. Secure communication channels ensure effective and safe access to equipment, enhancing operational efficiency and maintenance processes.

Predictive Maintenance
Predictive maintenance is a highly valuable capability of data-driven industrial digital solutions. It uses machine learning models to analyze historical and real-time data to predict when equipment is likely to fail. By performing maintenance just-in-time, manufacturers can avoid downtime losses and prevent unexpected failures, ultimately extending equipment life. Predictive maintenance leads to preventive maintenance, increasing equipment reliability and operational efficiency.

User Interaction
Industrial digital solutions must be user-friendly, providing operators with intuitive interfaces to interact with monitoring systems, enabling them to respond quickly to issues. Operators can access real-time data and alerts through smart devices to correct shop-floor deviations. Process engineers require analytical interfaces to streamline the analysis and optimization of manufacturing processes. These interfaces should enable access to detailed process data, disclose inefficiencies, and support improvements through extensive training. Maintenance technicians benefit from digital solutions that simplify the scheduling and execution of tasks. Intuitive interfaces can help them quickly diagnose equipment issues, track maintenance history, and manage spare parts inventory, ultimately reducing downtime and improving operational efficiency.

Best Practices
Management may be reluctant to adopt a fully digitalized/automated system because of the upfront costs and the need for an automation-educated workforce. In my 50-plus years of experience, I have found it easier to upskill current employees than to hire automation consultants. Adding a series of sensors to your manufacturing floor, updating your equipment, and implementing data management systems can be costly. However, with the right implementation plan, you can repay most of these costs over time with reduced operating costs, labor costs, and improved yields.

Another hesitation in making this move is the increased chance of a cybersecurity attack. You can manage these risks with carefully planned firewalls, VPNs, security scanners, and Linux-based controllers designed with cybersecurity in mind.

References

1. "Happy’s Tech Talk #32: Three Simple Ways to Manage and Control Wet Processes,” by Happy Holden, PCB007 Magazine, Sept. 3, 2024.

Happy Holden has worked in printed circuit technology since 1970 with Hewlett-Packard, NanYa Westwood, Merix, Foxconn, and Gentex. He is currently a contributing technical editor with I-Connect007, and the author of Automation and Advanced Procedures in PCB Fabrication, and 24 Essential Skills for Engineers.

This column originally appeared in the January 2026 issue of I-Connect007 Magazine.