High- and low-tech ways to ensure parts are correctly placed.

A core element of Lean manufacturing philosophy is the requirement to understand where inefficiencies and waste exist, in order to eliminate them. Even in a factory that thoroughly embraces Lean manufacturing and trains its personnel on all aspects of their jobs, there will be inefficiencies and waste related to operator error, supplier quality issues and process variation. SigmaTron International’s facility in Suzhou, China uses a two-part strategy to address this. At a systems level, it developed a proprietary Manufacturing Execution System (MES) system known as Tango to enhance shop floor control. At a granular level, the team evaluates production operations for areas where tooling or fixtures can minimize process variability and the potential for operator error. This month, we look at some of the common areas where this defect mitigation strategy can be most effective.

SigmaTron’s IT department takes a distributed approach to continuous improvement in its systems by letting teams at individual manufacturing facilities identify specific gaps in shared systems and develop appropriate software tools. These solutions are then tested at the facility that identified the need and later transferred across all facilities. Following development and testing in Suzhou, Tango was deployed to other facilities beginning in 2014.

Tango uses starts in incoming inspection. When the bill of materials (BoM) is originally entered, all special instructions are entered as well. This ensures handling and storage instructions are always associated with each part in the system. As material is received, the system also creates an internal log number that lists purchase order number, date code and lot code and outputs a barcoded Tango “card.” This supports not only device history recordkeeping during manufacturing, but also a first in, first out (FIFO) system during kit preparation. When material is stocked in the warehouse, handlers scan the Tango card and enter the location in the system.

From a defect mitigation standpoint, this ensures material is stored consistent with its handling instructions. It also establishes its priority in the FIFO system, plus ensures product shelf-life information is associated with its record. Finally, it ensures material can be easily found when needed.

There is also an efficiency benefit. While special handling instructions may dictate that some material must be stored under tighter environmental or security controls, the vast majority of material can be stored anywhere in the environmentally-controlled warehouse. The location feature in Tango allows warehouse personnel to load pallets on the nearest empty shelf. This facilitates efficient space utilization and faster material handling. Additionally, the data stored support customer traceability requirements.

On the production floor, Tango aids reel/feeder verification. When production operators load a reel, they scan both reel and feeder. The system verifies the reel is correct for that feeder number and that the feeder is loaded into the correct slot in the feeder table. Operators scan the product’s barcode at each production operation, and Tango tracks the sequence and also records which operators are involved in the process. Test results are also tracked.

This contributes to defect mitigation in two ways: It eliminates the possibility of incorrect component placement by validating the correct components are loaded in the correct slot in the placement machine, and it validates product is following the correct sequence of production activities by only registering when scanned if the prior step was completed.

This tracking activity continues to support traceability requirements and device history recordkeeping. More important, the team has created a warehouse for data acquired via Tango, enabling data mining of trends. In the event of a customer-returned material authorization (RMA), the team can go back to the data and quickly determine:

• The historical quality trending on that part number compared with that unit’s lot.
• The historical defect part per million (DPPM) trends on the components that are involved in the unit failure.
• Whether the defect is isolated in the returned unit or there is a possible component defect that may impact a larger number of units.
• If the issue appears to be related to component quality, the number of units impacted and their current location.

The ability to study this level of detail makes it much easier to identify the root cause of defects and quickly initiate corrective action.

The manufacturing engineering team also focuses on low-tech solutions by developing fixturing to minimize operator error and process variation in manual or semi-automated assembly operations. For example, a recent industrial product box-build required production operators to mate a connector on the printed circuit board assembly (PCBA) with a power cable. Because the connection involved powering a unit, a misaligned connection could create a safety hazard. While in theory operators could be thoroughly trained to assemble the part correctly, in reality, even diligent operators working at speed make occasional errors. A poka-yoke exercise was done on that assembly step, and a low-cost gauge developed to ensure the power connector was plugged properly without misaligning the pins.

In another situation, a poka-yoke fixture was created for an assembly operation that involved use of a large number of small screws. Operators could drop screws into the fixture, and they would fall into small holes, giving the operator easy access to the exact number of screws needed to assemble the unit. This sped up that stage of the assembly process and eliminated attrition occurring when operators attempted to count the correct number of screws on the workbench.

An approach that combines real-time data collection and warehousing with strong manufacturing engineering focus on minimizing process variation helps lower defect potential, while providing a path to learning and improving the process when defects do occur. This two-pronged approach improves throughput while supporting customer requirements for device history recordkeeping and traceability.

Hom-Ming Chang is vice president, China operations at SigmaTron International (sigmatronintl.com); homming.chang@sigmatronintl.com.

• Prev
• Next