For more than 50 years, a New York manufacturer of jewelry castings and settings used employees to manually inspect its die-struck parts. Although reasonably accurate, it is a time-consuming and tedious process. The company manually confirmed if the dimple (or “pinprick”) on a 1/4-in.-sq gold earring back was near the center.
If the dimple is incorrectly positioned by more than about 5/1000 in. there could be problems in the next manufacturing step, which consists of automated crimping. If the dimple is missing altogether, the forming punch breaks.
The SolutionBecause of the time consuming nature of the inspection process, the company chose to automate parts of its manufacturing process and implemented a machine vision system from Faber Associates, an application solutions provider. The foundation of the machine vision system design is a vision appliance from DALSA Corp. The automated system ensures that the jewelry parts are in order.
The system has a vibratory bowl that feeds parts down a vibratory track, single file, to be inspected. The system mechanics ensure that the parts are right side up and in the same orientation. After an earring back is inspected, and gets a passing grade from the vision subsystem, it continues down the track until it falls into a collection bowl containing usable parts, which will later be fed to the automatic crimping machine. If a part fails, either because the dimple is missing or is located in the wrong position, a pneumatic-actuated pusher is triggered that sends the part into a bowl of rejects.
DALSA provided the dimple-analysis system based on a VA40 vision appliance mounted to the upper right of the inspection station. The appliance’s embedded computer contains all of the hardware and software required to capture, process and display images. The VA40 also can handle communications, networking, user interface and industrial I/O, and it runs the software that analyzes images to determine whether parts are good or bad.
DALSA’s Sherlock software, a general-purpose image-processing package optimized for automated inspection applications, displays live images as the earring backs are processed along the vibratory track, provides a visual and numeric indication of the dimple offset, and indicates the average dimple position for the inspection run. An optical fiber displays a bright red circle on the track from a red LED to serve as a trigger signal for image capture -- all in time to view, analyze and judge the images before the parts reach the reject mechanism.
BenefitsBenefits of New System Prior to installing the system, the workers manually went though thousand and thousands of parts, piece by piece. Although they were able to detect an offset of the dimple from the center, it took many human work hours. With the system in place, all that’s required is that they take a look at the equipment every so often to make sure nothing is jammed up from a mechanical standpoint. The real magic of the new system, what makes it simple, is its very robust handling system and using the right lens and lighting.< br>
605 McMurray Road
Waterloo, Ontario, Canada N2V 2E9< br> Tel: 514-333-1301