How to Specify Vision Systems when Automating Measurement and Inspection Processes

✕

Today automation is increasingly being incorporated into manufacturing processes including quality control (QC). And for good reason! Automation helps speed up productivity, reduce or eliminate manual errors, and can alleviate some of the burden due to the lack of skilled operators.

Today the shortage of skilled operators is still prevalent, putting added stress on performing the quality control needed to meet tight delivery deadlines. Vision system technology has been developed to directly target these throughput issues with automation, which can allow users to leverage powerful system technology to complete faster, more accurate quality inspections in a fraction of the training time once required for quality control personnel.

Perhaps you already have a good idea that you want to procure a vision system but are wondering how to go about specifying the system when also automating measurement and inspection processes and operations. The following guidelines can help to determine the right solution.

Examine Application Requirements

Vision systems that offer automatic pattern recognition can be extremely productive, especially for a high volume, low mix of parts. Pattern recognition is where inspection routines are created for workpieces based upon the master template, complete with tolerances. Precise stage motion and lighting control can be built in as programmable steps to ensure maximum repeatability, regardless of operator skill. Once the program is created, an operator can place a sample on the stage and initiate the routine. When the routine is initiated, the system will automatically run the inspection, taking all relevant data points and providing accurate results in seconds.

What makes technology such as pattern recognition truly powerful is its independence for success without relying on the operator’s skillset in quality control and inspection. On a fixed stage system, for example, automated pattern recognition can be pre-set, allowing virtually anyone on the shop floor to walk up, place a part on the stage, and get results in seconds. If the part in question already exists in the system as an inspection routine, then simply placing the part within the field-of-view of the system is all that is required to initiate the routine. In certain situations, where quality checks are being completed on large batches of the same part, a robot or cobot can even be utilized to pick and place parts for in-process checks, removing the operator from the process entirely.

Tremendous amounts of data can be collected quickly and easily from each part inspection in real-time and documented with date, time, lot number, job number etc. This removes potential operator error and provides comprehensive data for statistical process control (SPC) and traceability. With automatic pattern recognition, 100% inspection is possible, and inspection speed and output can also be dramatically increased due to automating the measurement process.

Users can also leverage features such as integrated stage motion to act as a force multiplier, exponentially increasing the measurement throughput of a single operator. On larger field-of-view systems, for example, more features of a single part can be viewed at one time, allowing for rapid data collecion. Multiples of the same part can then be placed on the workstage, and the inspection routine can be run and applied to all pieces on the stage. What once took minutes, or much longer, now takes seconds. Custom fixturing can even be applied to the system, ensuring that all parts are uniformly postioned, if required.

Applications

In a real-world example, Alpha Grainger Manufacturing Inc. (Franklin, MA) produces a high volume of intricate, small turned parts and more for a wide range of industries. The ability to capture part dimensions quickly and accurately is critical, while reducing labor cost and time. “Part features may need to be checked to ten-thousandths of an inch,” said Mr. Grainger, “Parts can also have more than 20 features, such as on an automotive part, so our measurement solution needs to be fast, reliable and as automated as possible.”

That’s when the company found Starrett Digital Video Measurement Systems, and today Alpha Grainger has 14 Starrett systems such as the HDV400. The HDV400 combines the best features of a horizontal optical comparator and a vision metrology system by using a uniquely mounted 5MP digital video camera, coupled with a choice of quick change telecentric lenses or 6.5:1 zoom for micron-level resolution. Using MetLogix™ M3 touch screen software, the HDV can import DXF CAD files over a network and make automatic 2D “go/no-go” comparisons to an engineering design by using video edge detection. Mr. Grainger said that, “On average, inspections are being done every 20-25 minutes so every bit of time we save using a semi-automated inspection process versus a manual one, increases our productivity.”

Video edge detection, which detects transitions between light and dark fields, further helps guarantee accuracy by eliminating operator subjectivity. Using a user-friendly process, known as walkup metrology, means that once a part profile is created, any operator can simply initiate the program and instantly obtain the same results with confidence. For example, Roush Yates Engines (RYE) of Mooresville, NC, and its manufacturing division, Roush Yates Manufacturing Solutions (RYMS), chose the Starrett AV450 Automatic Vision System for quality control. RYE is the exclusive engine builder of Ford Performance for the top levels of NASCAR including the NASCAR Cup and Xfinity Series. In addition, RYMS is a precision CNC contract manufacturing solutions organization with expertise in multi-axis machining of metal and plastic parts with complex geometries. RYMS’ in-house production includes machined parts for the Ford FR9 V8 NASCAR and Xfinity Engine for Ford Performance programs, and contract CNC manufacturing for other motorsports, aerospace, defense, space, and medical markets. They also supply parts for the new NASCAR “Next Gen” Car to Ford, Chevy and Toyota teams.

The Starrett AV450 multi-sensor system can be used to manually measure a single feature on one part or to automatically measure multiple features on many parts in a program. A combination of probing methods may be combined in a single measurement sequence. Integrated MetLogix™ M3 software has unique features and touchscreen functionality that even enable minimally trained personnel to effectively utilize the system.

Functions such as palletized inspection routines, pattern recognition and autofocus increase throughput while reducing operator subjectivity. M3 software allows for making electronic graphic overlays using a DXF of the models, allowing the measurement of more complex 2D geometry, such as arcs of non-uniform radii, ellipses, etc. Powerful “Profile Fit” functions enable fast analysis of parts and non-prismatic features that are out of conformity.

Additional system features, such as LED quadrant darkfield illumination, enable even the most difficult-to-discern edge detail to stand out on parts that are hard to see and detect.

A variety of measurement applications ranging from simple flat gaskets to pushrods to spring retainers with 3D geometry can be measured on the same AV450 system. This flexibility is due to the versatility of probing, feature imaging, and edge detection tools. Since the AV450 system offers automated system measurement routines, inspection speed and throughput can be dramatically increased. Repetitive parts data is cached, and the program can automatically call-up the stored data when the same part needs inspection again.

Vision systems are incredibly powerful tools for modern manufacturing facilities. Their speed and accuracy, combined with powerful automation capability, provide an increased level of inspection confidence, regardless of which operator is using the system. This capability is a massive advantage to anyone seeking to streamline their manufacturing process and decrease lead times.