Abstract
Wire harness assembly remains a critical yet labor-intensive component of electronics manufacturing. While automation via Wire Harness Assembly Machines (WHAM) offers a solution, traditional control architectures based on PCI buses or standard Programmable Logic Controllers (PLCs) often lack the flexibility and scalability required for modern production lines. This paper proposes a specialized application layer for the Controller Area Network (CAN) bus specifically tailored for WHAM. By implementing a custom 11-bit identifier structure and a mathematical processing efficiency model, the research demonstrates how a multi-master CAN architecture can meet the high-speed, real-time demands of complex assembly while maintaining robust error traceability.
Introduction
The assembly of wire harnesses is a complex task involving the precise routing, bundling, and termination of electrical wires. Automating this process requires Wire Harness Assembly Machines (WHAM) to coordinate multiple sensors and executors with high precision. Historically, these machines have relied on centralized control systems, such as PCI-based motion controllers. However, these systems are often rigid, making it difficult to add new nodes or adapt the hardware to different assembly requirements.
To address these limitations, there is a growing need for a communication system that is multi-master, real-time, and flexible. Such a system must ensure data moves fast enough to synchronize high-speed mechanical parts while remaining easy to reconfigure when adding new sensors or executors. This paper explores the application of the Controller Area Network (CAN) bus as the backbone for WHAM control systems.
Comparative Analysis of Field Buses
Before selecting CAN, the researchers evaluated several prominent field bus protocols, including MODBUS PLUS and PROFIBUS-PA:
- MODBUS PLUS: While widely used, it is often proprietary and lacks the flexibility needed for high-speed multi-node coordination in a specialized assembly environment.
- PROFIBUS-PA: This protocol typically operates on a master-slave architecture. The reliance on token rotation and a single master can create communication bottlenecks, particularly when multiple nodes need to report status or errors simultaneously.
- CAN Bus: CAN emerged as the superior choice due to its international standardization (ISO 11898/11519), low implementation cost, and inherent multi-master capabilities. Its non-destructive bitwise arbitration allows for efficient priority management, which is essential for real-time industrial control.
Custom Application Layer Design
The standard CAN protocol defines only the Physical and Data Link layers. To make it functional for WHAM, the authors developed a custom application layer by segmenting the 11-bit identifier (ID) field. This segmentation allows for "hardwired" priority and efficient message routing:
- Operation Modes (3 bits): These bits define the state of the message, such as Normal operation, Error reporting, Customized commands, or Self-check routines.
- Error Identification (4 bits): This segment allows the system to immediately identify the nature of a fault, facilitating rapid troubleshooting.
- Node Distinction (4 bits): This identifies the specific hardware node (e.g., a specific wire feeder or crimping tool). The current design supports up to 10 distinct functional positions.
By placing error codes in high-priority bit positions, the system ensures that emergency stop or fault messages take precedence over routine operational data, allowing the machine to halt instantly if a failure is detected.
Processing Efficiency Model
To validate the feasibility of the CAN bus in a high-speed environment, the authors developed a mathematical model to calculate the minimum processing time, denoted as $T(x,y)$. This model accounts for nominal bit times, frame lengths (Data vs. Error frames), and potential retransmission times due to collisions or noise.
The primary constraint for the WHAM is the requirement to process a single wire harness within a 1-second window. Given the physical dimensions of a standard WHAM unit (approximately 1.2m x 0.8m x 1.5m), the total bus length is minimal. This small physical footprint allows the network to operate at its maximum theoretical speed of 1 Mbps without the signal degradation typically seen over longer distances. The model confirms that even with the overhead of the custom application layer, the total message transfer time remains well below the 1-second threshold.
Results and Discussion
The proposed architecture offers several key advantages over traditional setups:
- Functional Redundancy: Unlike PROFIBUS-PA, the CAN multi-master approach prevents the "bottleneck" effect. If one control node is occupied, other nodes can still communicate across the bus, providing a form of functional redundancy.
- Enhanced Quality Control: The system includes logic to monitor node health. The proposed bit-assignment method allows the system to automatically halt the machine if a specific node reports more than ten errors, improving overall quality control and traceability.
- Scalability: The 11-bit ID structure provides a clear framework for adding new sensors or actuators without requiring a complete overhaul of the control logic or hardware transplantation.
Limitations and Future Work
While the mathematical model and protocol definition provide a strong theoretical foundation, the research currently lacks empirical data from a physical WHAM prototype. Real-world factors such as hardware jitter and electromagnetic interference (EMI) in a factory environment may introduce variables not fully captured by the fixed parameters (such as the prescaler pe) used in the current model.
Future research will focus on implementing this CAN configuration in a physical assembly unit to identify specific "time waste phases" in the assembly cycle. This data will be used to further refine the processing efficiency model and optimize assembly algorithms for even higher throughput.
Conclusion
By moving away from rigid PCI-based controllers to a flexible, multi-master CAN architecture, this paper provides a roadmap for more efficient Wire Harness Assembly Machines. The custom application layer ensures that real-time constraints and error-handling priorities are met, while the processing efficiency model confirms the viability of the approach within the physical and temporal constraints of modern manufacturing.
Summary updated