Eddy Current Testing: Near surface flaw detection

Sept. 1, 2000

Eddy Current Testing

Near surface flaw detection

By Jim Cox

September 2000

As far back as the late 1700's, scientific investigation into electromagnetism was conducted by a number of creative people. They didn't have very sophisticated tools to work with (some wire, a magnet, a compass, and perhaps a crude battery), and yet they were able to derive the theoretical and mathematical relationships that define the eddy current testing process as it is still used today. The basic rules of how an eddy current test works are well established. If we can create a changing magnetic field near an electrically conductive material (metal) then we will induce (electromagnetic induction) a current (eddy current) to flow in that metal. The relationship between the energy that we use to create the test (primary forces) and the resultant current and magnetic field in the material (secondary forces) allows us to begin to measure certain characteristics of that piece of metal. If the metal is completely homogenous (no cracks, no corrosion), then the induced current will flow unobstructed in the metal. If the metal is somehow degraded or damaged, the eddy currents can't flow in the same way. This change in their flow pattern is detectable and can be displayed by the eddy current test system. In today's world of electronics, our coils are energized with an alternating current (AC) to create a varying magnetic field around the test probe. Coils are usually designed and optimized for specific tasks. Eddy current testing (ECT) systems allow us to do real-time analysis by watching a trace or dot move on an X-Y screen. We get direct feedback about the coils' energy and how it is affected by the test sample.

Material and Test Parameters
One of the historical arguments against using eddy current testing is that it is too sensitive. The signal changes that we might detect during our inspection could be created by some combination of material changes. The possible variables include the material's electrical conductivity, or its magnetic permeability, and/or the geometrical factors that are encountered while performing the examination. With that many possible responses, how do we figure out what really changed?
Conductivity (s) is an electrical property. It determines how well electrons will move through a material. Metals are normally classified as conductors. Theoretically, we could do an eddy current test on any piece of metal.
Resistivity (r) is another term that we can use to describe the electrical properties of a material.
Permeability (m) is a magnetic property.
Not all metals have significant levels of permeability. The permeability value of a metal determines how it will alter a magnetic field moving through it. In metals with high permeability values (carbon steels) the eddy current test process is only capable of detecting material changes that occur right on the surface immediately in front of the probe. Fortunately, the largest percentage of aircraft structures and surfaces are made from non-ferromagnetic materials (i.e., aluminum). This means that the magnetic field created by the coil moves through aluminum easier than if we were trying to inspect a piece of steel with the same test conditions. In aluminum we can detect changes that are much deeper in the material. The normal limit for this test process is about .2 to .3 inches into the nonferromagnetic materials.
The third major test variable is geometry. In surface scanning operations we typically define our geometry response as Lift-Off. This will occur any time that the coil detects a "change in the spacing" between the coil and the material. This type of signal would be created by the operator if the probe were rocked from side to side by just a few degrees. Geometry responses are considered to be a limiting factor in our ability to detect true material changes. Anything that we can do to limit these geometry responses during the test will improve the quality of the test.

Coils options Our choice of coil size and type will determine how well we can link our coil's energy into a test piece. Most of the coils that we use in aerospace component inspection would be addressed as "Probe" coils. They may contain one or more coils. Probe coils are typically moved across the surface of a material in the area of interest instead of just sitting in one position and taking a measurement. This scanning process can be done manually or it can be motorized or automated.
If the flaws that we are looking for are assumed to be near surface, then we will probably stay with the simple "pancake" or pencil probes shown in Figure 1. These typically have smaller diameters that would allow good sensitivity to short cracks. Their design parameters also dictate that they will have fairly high frequency ranges (50 to 500 kHz). This should put most of the coils' energy right at the coil/material interface. This will improve the sensitivity of the test. • If we expect to have a problem with the geometry of the part (curved surfaces, changes in surface geometry, etc.) then maybe we should select a "spring-loaded" coil. These would normally be used on flat plates but the spring-loaded coil approach should help limit lift-off responses in other surface geometry situations also. • If we expect to have difficulty in reaching the area of interest (through a hand-hole, on the backside of a bulkhead, in a tight radius) or encounter restrictions in the test area (bolt heads, seams, etc.) then we may need to try using a small diameter "pencil probe." Mechanically, based on how well we could control the position of the probe with respect to the surface, this would be less stable. We would then have to try to compensate for any geometry-related signals by using specific scanning techniques or applying electronic tools. Signal Displays / Calibration Eddy current test information can be displayed in many different ways. There are still a lot of meter-based eddy current systems in use but in most modern systems we would see a dot displayed on the screen of a "X-Y" system. A meter-based system only displays half of the information being generated as the probe responds to material changes. A trained technician using an X-Y system has a better chance of detecting small flaws while not being overly conservative.
Once we have decided which of the surface probe designs we need to accomplish this test the calibration process is going to be basically the same. We need to start with a sample (calibration block) that we know has regions that are undamaged, but which also contains representative flaws. This "cal" block will be used to establish our system settings of: frequency, gain, phase, and any display features that might enhance our detection potential. A good thing to check on any system before starting the calibration process is to make sure that all special system features (mixers, filters, alarms, etc.) are off. Another key issue is to verify that the X:Y screen display sensitivity settings are in a nominal mode (1:1). Some systems allow the vertical and horizontal (V/H) screen sensitivity settings to be adjusted independently. This is an excellent tool for signal enhancement later, but it can be confusing if this feature is being used during the initial calibration process. Check the probe to verify that it is in good condition. If this is a small pencil probe remember that over time the wear surface on the face of the coil can be removed just with normal use. You may need to add a protective layer of material to the face of the probe to protect the coil from damage. This might be a piece of Teflon® tape or a layer of something like nail polish or other hard material that can be easily applied.

1. If the system allows for multiple modes of coil drive operation (absolute, differential, driver/pick-up) then determine which is correct for this particular surface probe.
2. Select a frequency that is within the specified range of the coil.
3. Set the gain control to a setting that is in the low to medium range of its scale.
4. Place the probe on a good (non-flawed area) of the "cal" block.
5. Perform the electrical "balance" or "Null" function provided on the system.
At this time you should see a dot on the screen.
6. Rock the probe slightly from side to side to create a "lift-off" response. You do not have to completely remove the probe from contact with the surface. Note the direction of the lift-off response.
7. Rotate the "phase" adjustment on the system so that the lift-off response leaves the screen directly to the left. This would be the nine o'clock position on a clock. In ECT, we call that screen position "zero degrees" or we say that lift-off is now "on the horizon." Ninety degrees is at the 12 o'clock position and we continue through the range until we reach 360 degrees back at the 9 o'clock point.
8. Keeping the probe in a vertical position, move it slowly over the range of discontinuities on the standard. These will probably be EDM notches that represent cracks. The screen response should be a rapid movement of the dot into the upper left-hand quadrant of the screen.
Once you can verify that you have flaw detection capability then the only issues that remain are how to optimize the signal variation between flaw and non-flaw conditions. You may need to adjust gain to increase signal amplitude from the smaller EDM notches. You might want to alter the V/H ratio to improve the displayed signal-to-noise ratio. Keep in mind that filters can be used in some applications to improve flaw sensitivity but they should never be used with any manually controlled scanning process.

Signal Analysis The system should now be ready to look at our test samples that may contain flaws. Scanning of the critical regions where damage might typically occur needs to be performed carefully to assure complete coverage of those critical regions. If the probe is moved on the surface and the dot doesn't move (very much) that tells us that nothing is really changing in the test environment. If the dot does move, we have to try to decide why.
We are always going to have some geometry response as we move the probe across the test specimen. We always try to limit our visual sensitivity to geometry by putting that response at a known position on the screen, normally "horizontal." Permeability changes are not a major factor in most aircraft examinations. That only leaves material conductivity as a variable. Cracks should exhibit a change along what is referred to as the conductivity line. Once we understand this relationship it is a fairly straightforward process to decide what is a true flaw and what isn't. Discontinuities generally cause the effective conductivity of the test object to be reduced. If we look at the arrangement of signals on the screen in Figure 2 this would mean that any flaw would be represented by a "positive" or upward response on the screen from the balance point (NULL). Any response on the horizontal would be related to lift-off (geometry).

ECT requires trained personnel to understand it and apply the techniques properly. This generic discussion has been limited to detection of near-surface cracking in non-ferromagnetic alloys. However, there are many applications of ECT. Weld inspection, near and far surface crack detection, sub-surface corrosion detection and sizing, material properties measurements (conductivity), hardness measurement, thickness measurements, etc., can all be accomplished when the correct style of probes and techniques are applied.
There are a lot of misconceptions about the limitations of eddy current testing. With the advances in ECT equipment and probe technology that have been achieved in the last few years, it is now possible to apply ECT to many test scenarios that previously could not be handled. Inspection procedures and codes that have been in place for a number of years probably do not address these newer options. The eddy current method, when combined with modern tools and techniques, may actually be faster, cheaper, and even more sensitive to the defined flaw conditions than those other NDE testing processes that are specified or authorized. Even if ECT is not addressed in your inspection criterion, you might want to give it a try.