Crossing Automation
Larry Wise
VP of Engineering
Design for Modularity
The concept of modularity in software has come a long way since the pioneering work in 1970s; in mechanical design, modularity is often an incidental outcome. Similar to modularity in software design, the concept must be embedded during the early phases of the system architecture design. Mechanical functionality must be redistributed such that the failure points are isolated to individual modules, and this concept is at the core of Crossing Automation's approach to wafer-level automation. Modular systems allow module replacement without impacting overall system performance, facilitate preventive maintenance planning, reduce the time to repair and improve system availability and uptime.
Design Rules of Modular Automation Systems
The primary objective of modular design is to configure a complex system as a set of independent and distinct sub sets or modules. Modular designs create options that are not available for non-modular or interdependent systems. One option is the ability to re-configure a system for varying customer needs. Functional partitioning allows modular designs to segregate complexity into smaller and more manageable chunks; this reduces the uncertainty in the overall system performance.
We have found the following rules to be helpful in developing modular designs:
- Simplify functionality and isolate failure points
- Re-group failures and implement robust design
- Design modules for independent test and characterization
- Modules are designed for manufacturability
- Failure detection and swap-out field replacement
Module characterization to compensate for manufacturing variability High performance equipment, such as wafer automation systems are sensitive to variability introduced by manufacturing and assembly processes. Robotic mechanisms and variation in the end-use of the equipment, as examples, require an understanding of the optimal values of the system control parameters, and modular design permits each module to be characterized independently.
Data acquisition and analysis to enhance predictive maintenance implementations
Isolating failure points and determining the optimal set of control parameters provides the ability to measure drift in performance to the module level. Since measurement points are isolated and non-interacting, it is possible to pin-point deviations from normal performance and correlate the deviations to actual hardware to eliminate guesswork. The ease of field replacement, combined with real-time diagnosis, allows for preventive maintenance and results in high overall equipment utilization.