Eddy-Current Inspection for Production Applications
It's not just for cracks any more
By Jim Destefani
Traditional applications for eddy-current inspection involve a technician using a hand-held instrument to look for microscopic cracks in heat exchanger tubing, turbine blades, aircraft skins, and fasteners.
But, the current emphasis by automakers and other OEMs on zero defects is forcing suppliers of fasteners and other small parts to rely on 100% inspection to assure dimensional accuracy, heat treatment, and other properties. On the production line, a key is integrating high-speed eddy-current inspection into the manufacturing and material handling systems.
Eddy-current testing involves subjecting metal parts to a changing electromagnetic field created by passing AC current through a probe or coil. When the coil is placed near a conductive test piece, the electromagnetic field causes eddy currents to flow in the test piece.
As the eddy currents flow, they generate their own magnetic field. This field interacts with the magnetic field generated by the probe or coil, changing its impedance. Eddy-current test systems quantify and display the impedance changes, allowing operators to infer information about the properties and condition of the test piece.
George Nygaard, VP and founder of General Inspection Inc. (Davisburg, MI), says eddy-current testing is used in a comparative mode for such applications. "Known 'good' parts and known 'bad' parts are compared, and their respective signatures as displayed on the eddy-current tester are noted," he explains. "Typically, a 'gate' is placed around the good part signal, and parts that do not enter the gate are caused to reject."
According to Nygaard, vastly reduced defect levels experienced by many users of eddy-current technology are proof that the technology works. "Many manufacturers report lowering defect levels from a few thousand ppm to near zero by employing 100% inspection/sorting machines," he says.
The heart of any eddy-current test system is the probe, which must be designed for the specific application. "The key to the success of any eddy-current test system is the design and performance of the probe," says Martin Bryant of Uson LP (Houston TX). "It is through the probe that the alternating current is applied and the responding signal is measured."
According to John Hansen of GE Inspection Technologies (St. Albans, Hertfordshire, UK), users can select from several types of probes. These include:
- Absolute probes, which normally consist of a single coil and can be used to detect cracks as well as more gradual variations such as variations in metallurgy, heat treatment, and shape.
- Differential probes, which use two balanced coils and respond only to sharp changes such as cracks.
- Reflection probes, which use separate driver and sensor coils and provide a wider frequency range than other coil arrangements.
- Shielded probes, which use magnetic shields to focus the magnetic field and provide sensitivity to small cracks with low influence from edges, geometry changes, and other conditions.
- Unshielded probes, which are less expensive than shielded types and provide a wider scan area but are affected by edges and nearby discontinuities.
William Keely, VP of NDT Technologies Inc. (Holly, MI), says absolute probes are typically the least expensive, and they can discern differences in chemistry and hardness as well as thread presence and condition. "But this system detects all four eddy-current parameters from each sample, and consequently the associated part population bell curves are unduly wide," he says. "As the size of the defect to be detected decreases, more and more good parts must be rejected from the inspected sample population to insure that all bad parts are also rejected."
Differential probe systems effectively reject unwanted elements from the eddy-current signature, allowing more sensitivity to geometry differences. "Lot-to-lot variations in base material chemistry, hardness, and temperature are effectively rejected by the system and do not affect the sort," Keely says. "This allows one or two missing threads to be discerned from fully threaded holes."
Production Applications. Donald N. Bugden, VP, Magnetic Analysis Corp. (MAC; Mt. Vernon, NY), says eddy-current can be integrated into a production operation relatively easily. "Because it is a noncontact method and requires no coupling medium [unlike most ultrasonic devices], eddy-current testing is widely used to inspect both ferrous and nonferrous mill products such as bar, wire, and tubulars," he explains. "It's also becoming more common to build custom machines to verify the integrity of critical portions of a part. A good example is piston rods for McPherson struts, which are inspected using spinning probe coils."
Keely says detection of both internal and external threads is one of the most common production applications for eddy-current testing. "Thread detection using eddy-current inspection has become sophisticated to the point where not only the presence of threads can be sensed, but also the number of threads in the hole can be accurately determined and, with properly engineered systems, the portion of the last thread helix that is present in the hole can be determined," he says.
According to Keely, other production applications of eddy-current inspection include crack detection; runout or displacement measurement; material properties including composition, heat treat condition, and hardness; and weld seam inspection in tubing.
"In finished surfaces, differential probes can reliably find cracks less than 0.001" (0.03 mm) wide," he says. Eddy-current inspection can also detect very small cracks in holes, as well as pre-cure cracks in powder metallurgy (P/M) components.
Eddy-current profiles of a correctly threaded transmission part (left) and a part with half threads. Green lines are ±3σ limits, yellow is the good part profile, and the red line overlays the yellow line where the part falls outside the control limits.
Displacement or runout measurement makes use of stand-off, a characteristic of eddy-current inspection that in some cases can cause problems. "Stand-off is the separation between the sensor and the part," Keely explains. "In most eddy-current inspections, stand-off is considered as 'noise' that must be minimized in the sensed reading. But, by precisely sensing stand-off, displacements in the micron range can be measured in real time."
For checking material properties, General Inspection's Nygaard says magnetic permeability and electrical conductivity are the key parameters. "Permeability is the variable usually used to detect and sort improperly heat-treated ferromagnetic parts," he explains. "Typically, parts pass through doughnut-shaped eddy-current coils from a conveyor or down a V-slide, and an automatic gate separates the properly heat treated parts from those that are defective."
Electrical conductivity is used to detect metallurgical changes in nonferrous parts. "The aircraft industry uses eddy-current instruments that read out in percent IACS [International Annealed Copper Standard], whereby pure copper is considered 100% conductive and the conductivities of other nonferrous metals are compared to known standards," Nygaard says.
He also points out that, regardless of application, it's important to minimize variables other than that being tested. "For example, temperature affects conductivity so it should be kept constant or corrected for during testing. Also, size and mass differences may affect heat treat sorting," he says.
An example of use of eddy-current testing in a production environment is provided by NDT Technologies, which supplied a system to profile threaded holes in transmission output shafts. The 12-mm diam holes had a thread pitch of 1 mm. A pilot hole and chamfer appeared as part of the hole profile.
Eddy-current profiles were acquired by fully inserting the probe into the hole, then withdrawing it at a rate of about 2 in./sec (51 mm/sec). The system recorded probe position with respect to the part datum using a linear potentiometer mounted in the probe fixture.
Profiles of ten known good parts were "learned" by the system; the computer then calculated the ideal profile and ±3σ limits of the family of profiles. The limits are displayed along with the profile of the part under test. If the test part's profile falls within these limits, it passes; otherwise, it fails.
"This type of profiling inspection produces data that can be stored for further analysis using statistical process control or other trend analysis programs. It can also provide a permanent historical record of machining processes," Keely says.
Contact + Eddy Current = Precise Bore Measurement
One of the latest developments in eddy-current technology is a system from NDT Technologies that combines contact gaging and eddy-current testing.
The patent-pending Ball Runner system can measure hole diameter, ovality, cylindricity, taper, and surface finish with a single probe. According to the company, measurement takes place as fast as the probe can be inserted into and then withdrawn from the hole.
The Ball Runner probe allows accurate measurement of any ID greater than about 0.2" (5 mm) to within ±5 µm. This is accomplished by inserting the Ball Runner probe into the hole and then withdrawing it while recording the position of the probe with respect to an established part datum. During probe movement, both the "contact" eddy-current signature of the part extracted by the probe and the position of the probe within the hole are inputted to a computer. Software then plots probe position versus part signature, generating a profile that represents the inside diameter of the hole on the computer screen.
The system was developed when NDT Technologies was asked to determine the ID of a center hole in electric motor rotors, which were produced from a stack of laminates arranged so that a center hole in each of the laminates would align to create the center hole in the stack. The customer wanted to measure the hole ID within ±0.00075" (19.5 µm). Several attempts had been made by other companies, using various inspection technologies, with no success.
Print requirements were for the center hole in production parts to be from 0.4970 to 0.4985" (12.62 - 12.66 mm) diam. Measurement was complicated by the fact that the ends of the stack were staked, thereby decreasing hole size of a few of the outer laminates by about 0.0009" (22.9 µm). This made it difficult to measure the actual ID below the ends of the stack in a production environment.
To provide an absolute gage reference for measurement of the rotor bores, NDT Technologies located two "gage rings" that bracketed the part ID dimensions so that the probe would first pass through them as it entered the part. The ID profile of these gage rings was therefore stored in the computer memory along with the ID profile of the part. Profiles generated by the system show both the two gage ring profiles and the part ID profile.
When NDT Technologies used the system to perform measurements on rotor laminate stacks, technicians discovered that none of the parts tested was within the required tolerance.
This article was first published in the July 2005 edition of Manufacturing Engineering magazine.