Defective parts are extremely unwelcome in the automotive world, and nobody knows this better than parts suppliers such as The Meister Group, a Belgian industrial outfit that largely caters to the automobile market. With factories in Belgium, France, and the Czech Republic, Meister specializes in the mass production of cut steel parts. The challenge facing Meister, whose production units use specialty precision lathes, is that it must manufacture complex parts in seconds and guarantee parts conformity on delivery.

The Meister factory in Haute-Savoie, France, manufactures electric valve parts for automobile equipment manufacturers specializing in ABS braking systems. Nearly 24 multi-spindle lathes produce 120,000 parts each day, with an annual production of 35 to 40 million parts.

In a sector where the smallest incident on an assembly line creates exhaustive investigations and leads to complicated and costly procedures for the subcontractor, zero faults is a necessity.

However, given the manufacturing techniques used and demands of mass production, this goal cannot be attained using only machines. A checking and sorting system is needed to remove defective parts, such as missing or loose components, metal shavings, and damage from vibration or knocks.

Previously, sight checks were conducted by human operators, which limited the number of defective parts to around one in 1,000. This defect count was unsatisfactory.

Meister’s technicians began considering automated vision checks, as they already had experience with industrial vision for a dimension-checking application. A seminar organized by the Alpsitec company, a systems integrator, provided technicians with additional information on the capabilities of In-Sight vision systems from Cognex Corp., Natick, Mass.

Alpsitec was asked to demonstrate that the Cognex cameras were capable of “seeing” faults. Following this first test on Meister’s production line, a prototype was created and assessed for a month. Meister then ordered and installed two test systems at the end of the production line, which were used to perform a final inspection of each part just prior to packaging.

Here, the parts are put into their packaging by a robot. Once packaging is complete, the robot picks up the packaged part and places it on the test surface. Then, the robot takes hold of an In-Sight 1000 vision system linked to a lighting system and passes it along the packaging mesh and over the parts. The system sends information from the inspection to the robot’s control center. The robot then takes any defective parts and deposits them into a chute, based on the nature of the fault found, where they are then transported to a hopper.

One of the test benches is outfitted with two In-Sight 1000 systems and operates at a rate of 6,000 parts per hour. The other system comprises a single sensor and functions at a rate of 4,000 parts per hour. During the first few months of the operation, both systems worked as dual sorters. The bulk of updating the application consisted of identifying faults recognizable by the checking systems and “teaching” these faults to the vision sensors. This procedure is essential for optimizing the efficiency of the inspection system.

The rate of faulty parts delivered to customers has rapidly dropped to 40 per every million packaged. Powerful processor algorithms and detailed analysis could further reduce faults, to below 20 ppm.
Jean-Marc Sermet, Technical Director of Meister France, has supervised this project from start to finish and is pleased. Alpsitec has also trained a technician who has taken charge of setting the vision systems’ parameters.

Meister is now able to internally input data for new defects to be identified and is able to modify these parameters in relation to the 15 different types of parts to be inspected. The greatest reward has been considerable return on investment in less than six months. The significant improvement in quality has greatly strengthened customer relations, and Meister is currently considering additional vision applications on its lines. For more information, visit Cognex.