Original Date: 08/14/2000
Revision Date: 01/18/2007
Best Practice : Product Recognition Technology Project
Many orthopaedic parts within a family are almost identical in size and shape. To resolve misidentification during batch runs, Howmedica Osteonics developed the Product Recognition Technology Project which consists of three elements: the Product Verification System; the 2-D Data Matrix Marking System; and the Elimination of Manual Data Entry. Each element’s goal is to reduce the number of product identity errors.
Howmedica Osteonics has adopted a cell-oriented manufacturing philosophy in which a semi-autonomous group is responsible for a specific part family. Many of the parts within a family are almost identical in size and shape. Because a typical product mix consists of many small batches and numerous manual operations, a great potential exists for product identity errors. To resolve this situation, the company developed the Product Recognition Technology (PRT) Project which consists of three elements: the Product Verification System; the 2-D Data Matrix Marking System; and the Elimination of Manual Data Entry. Each element’s goal is to reduce the number of product identity errors.
The Product Verification System is a compact, computer-controlled package consisting of a weigh scale and a 3-D, multicamera-based vision system. The operator begins the process by scanning the work order router sheet that accompanies each part. The computer then retrieves product dimensional and fixture/placement information on the part from a Microsoft Access database. Next, the operator positions the part on the weigh scale based on prompts from the computer, and presses the verify button on the touch-screen monitor. The 3-D vision system verifies that the part matches the information on the router sheet by comparing key features from the part’s image with corresponding features of the image stored in the database. The computer then displays a Go/No Go message and keeps track of the number of parts which have been measured. Although this method works for part identification, the measurement accuracy is generally insufficient to replace a separate dimensional inspection operation. As technology improves, Howmedica Osteonics will re-examine this method for dimensional inspection. The Product Verification System is easy to use, has gained operator acceptance, and has demonstrated its ability to catch errors prior to final packaging of the product.
The 2-D Data Matrix Marking System uses a rectangular matrix of tiny squares to encode product information on parts more efficiently than standard barcodes. In addition, standard barcode information obtained from a part’s router sheet can also be converted to the 2-D data matrix format, via software, to generate the machine-control code to produce the matrix. Up to 25 characters can be permanently encoded in a 1/8-inch-square matrix on the part’s surface. Data matrix codes can be applied with a laser marker (the most robust method), machined on the part’s surface, or cast directly into the part. The data matrix is applied early in the manufacturing process to minimize the possibility of part identity errors, and is read by handheld or fixed Charge-Coupled Device (CCD) scanners. The data-encoding scheme relies on light and dark areas within the matrix. Therefore, a special LytePype™ attachment, tailored to the geometry of each part, must be used with the CCD scanner to ensure appropriate lighting conditions. Once the data matrix is applied to the part, it can be used at any subsequent stage of manufacturing and possibly for field tracking in the future. Although this technology may eliminate the potential for part misidentification, several barriers must be overcome before it can gain plant-wide acceptance. These barriers include slightly increased lead times, the need for more robust scanning devices, determining an appropriate marking location for all part families, and identifying the earliest operation at which a code can be applied and still withstand all subsequent operations without damage.
The Elimination of Manual Data Entry of router information addresses two potential part identification problems. First, the operator can inadvertently apply the incorrect part number and/or lot identification to a part during the laser marking process. This problem is being minimized by using a barcode scan from the shop floor router as the only input to mark the part number and lot identification on the product. Second, the operator can inadvertently machine a part of the wrong size or type by entering incorrect information at the machining center’s Computer Numerical Control (CNC) console. This problem is being minimized by using a barcode scan from the shop floor router as the only input to activate the required CNC program for the manufacture of the product from raw stock. The barcode reader is interfaced with the serial communications port (RS-232) of the machine tool controller as well as the PC that contains stored programs for entire part families. Scanned information is passed to the computer and parsed to generate an appropriate four-digit part number, which is used in a look-up table to retrieve the correct CNC program. Up to 16 different machine tools can be controlled by a single PC. The process can also be used to pass statistical process control and shop floor control information from the controller back to the PC. The software also has the capability for data matrix encoding.
Howmedica Osteonics has implemented the three elements of the PRT Project in one or more manufacturing product cells, and will gradually be phasing them into other cells as the technology becomes more mature.
For more information see the
Point of Contact for this survey.