The first optical comparators, known as “shadowgraphs,” entered the factory when customers’ needs for identical parts became critical.
The shadowgraph presented a magnified shadow of a part that allowed easy comparison to an accurate profile of the part, usually in the form of an overlay.
The operator would use a micrometer head to move the X- and Y-axes to approximate the horizontal and vertical deviation between the overlay and the part and, making note of any discrepancies, the operator went back to his machine, reworked the part and repeated the process until the part was within tolerance.
It was essentially a simple go/ nogo pass/fail test.
Today, the optical comparator is a reliable factory stalwart that is easy to understand and use, whether on the shop floor or in the inspection room.
But they are a far cry from the basic shadowgraph.
The transition from purely observational optical measurement to direct digital analyses began about 20 years ago.
Modern comparators are equipped with better optics, readout devices, software interfaces, geometric measuring capabilities and screens that range in size from 12 inches to 40 inches.
Light path options are offered in horizontal or vertical orientations, determined by the application.
Today, many optical comparators are configured with optical edge detection that greatly aids the user. They are simple, repeatable and enable all users to gain the same results.
The evolution of the optical comparator has accelerated in recent years with increasingly powerful Digital Readouts (DRO) and PC based readout systems.
There is even a new hybrid that combines a comparator with a video adaptor to create a videobased measuring system with the superior light transmission of a video zoom lens and magnification beyond the range of comparators.
In keeping with the optical comparator tradition, this hybrid is simple and cost-effective.
Stepping Up to Vision
While the comparator is a versatile measuring and inspection tool, it has limitations when it comes to requirements for high surface illumination and magnification.
For these applications, the manual vision system is the next logical step up from a comparator.
A manual vision system provides superior surface illumination and the flexibility of a zoom lens with magnification ranging from 10x to 240x.
At one time, the manual vision system used a simple crosshair, CRT monitor and geometric readout. Those that offered edge detection required the use of complex computer systems.
However, today, video edge detection is a common feature and operators that are familiar with comparators find the accompanying electronics and readout intuitive and easy to learn.
Popular features include the ability to capture images and archive them, add text and upload data to a flash memory device.
In fact, the latest manual vision systems provide most of the capabilities found on computer-based systems without the longer learning curve and higher costs.
Users who are familiar with comparators frequently find manual vision systems to be easy and surprisingly affordable step up in capability.
Automated Inspection Speeds Production
Growing production volume inevitably will require increased throughput and repeatability from the inspection process.
With large batches, inspection often involves repetitive measurement routines that lend themselves to automation.
At the same time, increasingly tight tolerances may be incom-patible with manual inspection and operator subjectivity.
When inspection can no longer keep pace or accurately inspect complicated parts, automated non-contact inspection processes may be required.
For manufacturers that have large batches of components, the inspection process also can be streamlined by arranging the work on pallets.
The solution is to use fixtures to organize parts.
By creating a fixture set-up in rows and columns, the offset between parts will be a known and constant size in both the X- and Y-axes.
The offset information is used to program a measuring routine appropriate for the fixtured parts.
For some applications, though, the fixture could be as simple as a rightangled plate against which parts are placed, one after the other.
The Need to See More
As parts become smaller and more intricate, the demands on inspection systems become greater.
Higher magnification and better illumination underscore a basic, yet enduring principle for vision systems: If you cannot see it, you cannot measure it.
Lighting, for example, is essential for obtaining optimum contrast.
Quadrant lighting provides individual control over the angle and direction of the light source. Throughthe- lens lighting is frequently necessary for height measurements and for seeing down blind holes.
Automated vision systems use video edge detection to automatically establish points.
Some systems now are able to take as many as 300 data points with the click of a mouse, to measure a feature such as a circle or a line.
In 2-D scanning mode, the software can take as many as 5,000 data points around the shape. The more data on the part, the more precise the comparison to the reference file is.
For example, an operator using a traditional overlay to check parts now can import a 2-D CAD file, create a tolerance and using a video tool scan around the part. With this method, the operator now can collect nearly 5000 data points.
Then, the system automatically fits the points to the CAD profile and tells the operator whether the part has passed or failed.
This process is a little longer than the traditional methods, but the operator is able to obtain infinitely more data for the comparison.
What’s more, the data points can be stored for future reference or inspections —something that cannot be done with the overlay method.
For parts that have tight tolerances, accuracy also may require the use of an automated vision system. Some of these systems are have a working accuracy of less than 4 microns.
For example, medical instrument manufacturers are finding the inspection of small intricate parts such as surgical blades and bone drills have become so complex that they require the magnification and accuracy of a vision system.
Optical comparators still are used for many shop floor inspections while vision systems are providing measurement repeatability and increased productivity in the QC department.
Non-contact measurement continues to evolve
Successfully measuring complex parts requires the most suitable scanning method and the mathematical power to extract and compare key dimensional features.
Today’s advanced vision systems are integrating video and laser scanning with touch-probe inspection.
This creates a multi-sensing vision system capable of switching between methods, depending on the needs of the part.
With these new systems the phrase we used earlier: “If you cannot see it, you cannot measure it,” is no longer a concern because if the vision cannot see it then you can probably measure the part using either a touch probe or a laser.
The decision to go with a multisensing inspection system warrants careful analysis.
Some parts are best suited to physical measurement using touch probe inspection.
Other, more intricate parts with 2-D features, for example, are better suited for a vision system because of the system’s ability to precisely find the part’s edge and take a large number of data points.
Then there are situations in which assemblies can include both 2-D and 3-D features that require the capabilities of both a touch probe and a vision system.
This is where a multi-sensor inspection system becomes essential.
Multi-sensor inspection systems enable the operator to use the touch probe when it is most appropriate, and switch to the vision system as necessary, treating the entire inspection in a single routine.
For accurate results, fixturing needs to be considered when using automated multi-sensor measuring systems.
Proper fixturing is critical for producing reliable measurements.
For example, different fixturing is required when making contact with the part via touch probe versus when you are using vision or laser options that do not require contact with the part.
Totally automated inspection is becoming reality today.
For example, some vision systems can interface with pallet-loading robots. The vision system coordinates positioning the work stage, loading the work piece, measuring the part and calling on the robot to present the next part for inspection.
In the future, it may be possible to interface inspection with a pallet loading robot and the machining center so that systems can be used to measure, then visualize modifications and communicate the new offsets and dimensions back to the machining center.
Today’s visual inspection goals will continue to be removing subjectivity from the production process, while improving accuracy, repeatability and throughput.
Non-contact measurement has evolved into a resource with many options and designed to suit a wide array of applications.
While the future relevance and use of optical comparators once was the subject of debate, shop-floor friendly comparators still have their place in the production area.
Where illumination and magnification have become issues, the transition to a manual vision system is relatively easy and economical. And, where throughput, speed, volume and repeatability issues loom, automated inspection systems are an ideal solution.
This article was prepared by The L.S. Starrett Company (www.starrett.com) of Athol, Mass. • Edited by Bruce Vernyi, Editor-in-Chief
