Machine Vision is Important Even to the Lowly Fastener

One of the earliest application-specific implementations of a machine vision-based gauging system that I was involved with back in the late-70's while at EMR Photoelectric/ Schlumberger was dedicated to inspecting threaded fasteners manufactured for aerospace applications.  It employed a backlit arrangement to measure head, distance to first thread, min/max thread diameters and pitch. Both material handling and the properties of vidicons presented significant challenges, not to mention the limitations of the PDP-8.  To paraphrase the old Virginia Slims ad - Machine vision, you've come a long way baby!

Non-contact, machine vision-based systems offer distinct advantages to contact-based gauging: eliminate concerns for wear and damage to gage surfaces as well as eliminate reproducibility errors introduced by variations in 'touch' by a single operator or between operators. The fastener industry has been using one or another form of machine vision for many years as a means of assuring the integrity of a fastener. At the very least, these systems monitor the fastener to assure that all the features are present: head, thread, length, etc and sort production anomalies from good product. Some of the more sophisticated versions also verify pitch and other geometric properties. Most of these can be characterized as bulk sorters as opposed to online process monitoring systems. Bulk sorters typically use a vibratory bowl feeder to singulate and feed the individual fastener past the vision sensor at rates up to 500 to 1200 or so per minute.

Systems are also available that operate as a laboratory instrument to sample check a batch. Conventional laser gauging systems and machine vision-based optical coordinate measuring machines are also often used in the laboratory to spot check critical diameters. The off-line systems lend themselves to measuring die features to assure properties before the fasteners themselves are cut.

While there are systems available that perform a 100% dimensional inspection on fasteners, most widely used in the fastener industry are systems that can be best described as 'sorters.' The challenge is to eliminate a 'foreign' fastener (one that is not similar to the rest in a batch) as well as to verify that minimum conditions exist (such as head, threads, and appropriate length).  In this case, sorters act more like 'go-no go' gauges rather than base sort decision on absolute dimensional measurements on each fastener. Some systems combine machine vision with eddy current or other sensor approach to verify base material is correct for the batch or that every fastener has been properly heat treated or is not otherwise flawed. In addition, where fastener integrity is of the utmost concern due to the safety issues associated with use (aerospace, medical devices, etc.) machine vision-based systems to inspect the wire used to manufacture the fastener for surface defects that could effect the structural integrity of the fastener.

The use of sorters in this industry has been widespread. Sorters are especially well suited for finding foreign products, which are sorted based on gross size variations and/or materials (using non-optical sensors) and are designed to sort 'foreign' product based on 'go-no go' measurement techniques and not validate dimensions.

Machine vision-based sorters can be based on conventional video-based or line scan-based, single or multi-camera approaches or laser-scanning/profiling-based approaches or combinations of these approaches or systems based on light-blocking principles using photocell detector schemes.  Sorters are not always physically located at the fastener manufacturing facility itself. Those companies that maintain their own distribution centers will generally install sorters at the distribution point rather than in the manufacturing facility.

Machine vision systems have proven to be most valuable to address fasteners made to the most demanding specifications or sold to the most quality demanding customers or sold for applications where safety is a concern. Today as more and more assembly processes are being automated, the need for defect-free fasteners is even more critical than before.

In addition to machine vision-based products targeted at specific market segments (aerospace, automotive, appliances, medical devices, electronics, eyewear, etc.) the fastener industry can be characterized as plant sites that produce

High mix/low volume
Medium mix/low volume
Medium mix/medium volumes
Medium mix/high volumes
High mix/high volumes
Low mix/low volumes
Low mix/high volumes

The offline systems are well suited for high mix/low volume scenarios, while the low mix/high volume scenarios are well suited for online or near online-based, bulk sorting systems. In the case of fasteners for the automotive industry batch sizes run in the 50,000 to 500,000 range with 250,000 being typical. At a typical rate of 250 parts/minute, a lot might take from 3 - 4 hours to 25 - 35 hours to complete. In the aerospace industry the lot sizes are typically smaller (5000 - 50,000) but production rates are generally slower typically on the order of 100 parts per minute or less.

Most systems are being used to perform a final sort before shipment to a customer. True online systems deployed immediately following the header operation (where a separate operation from thread forming) or thread forming operation present serious challenges for machine vision.  One major challenge is the production rates: most thread formers produce fasteners at a rate of 100 to 500 per minute. Other issues which dictate where in the production process one finds machine vision systems include:

Certified System Integrator Program

Set Yourself at the Forefront of the Global Vision Market

Vision system integrators certified by A3 are acknowledged globally throughout the industry as an elite group of accomplished, highly skilled and trusted professionals. You’ll be able to leverage your certification to enhance your competitiveness and expand your opportunities.

GET CERTIFIED

• Unwanted materials on the surface of the threads (oil, particulate, etc.) that are acceptable and thus result in false measurement data could influence dimensional properties measured on line.
• Checking for thread damage requires inspecting while turning the part and not just looking a single silhouette view.
• There are subsequent operations such as heat treating and forming that are performed in bulk process steps that could lead to thread damage and changes in dimensional properties. Since these are bulk operations it is also possible that a 'foreign' fastener can get into a specific lot. A sorter can be used to make some measurements and at the same time separate out 'foreign' product based on material as well as size/shape.
• Given the absence of dedicated lines today, systems have to be able to handle multiple part models. This challenge relates as much to part handling compatibility as to dimensional measurement compatibility. The material handling would have to be able to handle a range of part sizes (both diameters and lengths) as well as materials that could be produced on the same thread roller.
• If tooling is set up properly process failure is not abrupt. Consequently even a low sample inspection hourly is adequate to avoid scrap production. A final inspection is sufficient to verify correct fastener is being shipped.

There are about 30 separate geometrical features and dimensional characteristics in the design and fabrication of screw threads. There are a number of dimensional standards for the fastener industry that reflect how measurements should be made: ANSI/ASM B1, Mil-S-8879 and FED-STD-H28. The most rigorous standard inspects 11 major thread characteristics.

Properties measured include:

a) Head diameter
b) Head height
c) Shank Length
d) Shank diameter
e) Major/minor thread diameter
f) Pitch diameter
g) Head and tip properties
h) Slope
I) Flank angles
j) Root radii of curvature
k) Concentricity
l) Thread damage
m) Pitch cylinder taper
n) Missing thread

Some of these measurements are less important: since they typically do not vary given the setup is correct; for example, flank angles and root radii of curvature. Other parameters that might be measured include the distance of the first thread to the back of the head, countersink slope, thread runout, lead error on thread, grip length/length of unthreaded diameter, head angle, radius between head and shank and between shank and thread, shoulder-to-shoulder dimensions on bolts based on multiple extrusions, grooves. The challenge is that there could be many part specific/part application specific features. In the case of systems offered making dimensional measurements, inspection data can be recorded and archived as well as reduced to standard SPC type reports.

The tolerances are not the same for all dimensions of the fastener and vary by application/industry. In the aerospace industry the tighter tolerances are generally +/- 0.0005' and on threads a total tolerance of 0.0002' and/or +/-0.0001' and angle measurements +/- 1 degree. In the automotive markets the tighter tolerances are typically: diameter a total tolerance of 0.002' and/or +/-0.001'; length tolerances generally more like 0.010' to 0.020'; angle measurements +/- 1 degree.

The automotive industry is pushing in general for tighter tolerances and even the fastener industry is beginning to sense this pressure. There is also a pressure to reduce the variances around the nominal of a tolerance especially where mating parts are involved.

A complementary requirement in the fastener industry involves sorting nuts. Sorting might be based on presence/absence of locking crimp, locking ring, threads, bore chips as well as dimensional checks: bore diameter, thread pitch, hex size, flange, thickness, height, etc. While machine vision and lasers are being deployed for these applications, an alternative is probes based on capacitive sensing.

Because of the increased emphasis on quality driven by manufacturers seeking a competitive edge from consumers, and the increased use of automated assembly tactics to improve productivity seeking a competitive business edge, the need for non-contact gauging and machine vision/laser-based optical gauging is expected to increase in fastener manufacturing. In automated assembly applications performing 100% sorting eliminates the potential jamming of high-speed fastener insertion equipment as well as the potential failure of the assembly itself due to a fastener without threads, for example.

What capabilities one expects from such systems must be well understood before a decision is made to purchase one. The market has many suppliers of such systems, but all systems are not equal in terms of parametric performance. The ideal is to match the performance of the system to the requirements. Today with cost-effective telecentric optics, digital, high-resolution cameras and machine vision systems with robust sub-pixel measurement techniques, systems are available that are well suited to handle the metrology requirements of the fastener industry for online, offline and bulk handling.