Despite being available for many years, Eddy Current Array (ECA) is still not a well-known technique. Many users have experience with conventional eddy current testing from which they’ve learned the smaller the coil, the smaller the detectable flaw. This remains true with ECA, but other parameters are introduced when bundling multiple single coils into an array. The power of an ECA probe for detecting small flaws is largely thanks to the probe’s resolution and uniformity. But why and how exactly?
The coils most commonly used in ECA probes are what we call “pancake coils”, because they are actually small discs of wound copper wire. Considering their circular shape, it would be false to think the sensitivity zone corresponds to their full diameter. If there is a flaw on the edge of the coil, it could be missed or the signal amplitude might be such that an inspector or analyst could easily dismiss it—leaving behind a potential source of failure of the asset. Ideally, we aim to go no lower than 85%, depending on the application and target defect, of the maximum amplitude wherever the flaw is located. In the image below, the vertical dashed lines show the limits of the actual functional zone of sensitivity of the coil. To work around this, conventional eddy current systems were raster scanning the surface with a wide overlap to avoid a flaw falling in the low sensitivity zone. This method was proven to be effective, but it also increased inspection time considerably, and deployment was more complex and less efficient.
The shape of the curve depends on different parameters of the coils (injection frequency, coil diameter, etc.), but also on the topology used (topologies explained here). The image above depicts a single coil in impedance absolute mode, but some topologies have an intrinsic wider signal curve.
To get around this drop in signal at the coils’ edge, ECA probes are designed to have individual signal curves meet high enough to keep an optimal signal stability, wherever the flaw passes.
It is clear that the second row of coils reduce the maximum signal drop compared to a single-row probe.
It is important to note that Figure 3 above depicts the results for a probe working in impedance absolute mode (each channel composed of a single coil). For transmit-receive probes, doubling the resolution as shown above would imply to use four rows of coils as the topology requires two rows to have axial and transversal channels.
The number of rows may be adjusted until the desired level of sensitivity and uniformity is obtained. Though, it is important to understand that this probe design alteration increases the number of coils, thus the number of channels.
ECA probe signals must be uniform when a defect is in a different location underneath, especially narrow defects, or when sizing defects. Uniformity depends on the spatial response to defects of individual elements, the distance between coils, the number of coil rows, and the topology as some topologies are known to offer higher signal uniformity than others.
The smallest detectable flaw is linked to the coil size but also to the resolution of the array. Resolution also impacts the image resolution obtainable to create the raw C-scan image. For a given size of coils, the smaller the resolution value, the better results we get.
Resolution is not the same as uniformity but is directly related to it; an increase in number of rows also has a positive impact on the resolution. Greater resolution improves probability of detection. Figure 5 below demonstrates the effects of signal amplitude on a small flaw when using a probe with higher resolution.
As depicted on the left of the figure above, a low-resolution probe will offer a low signal of a flaw that is smaller than the probe resolution, considerably increasing the risk of not being able to identify and locate it. On the right of the figure, we realize that the Probability of Detection (PoD) of the same small flaw is considerably increased. Even if slightly shifted left or right, the flaw will be in an area of the probe where the signal amplitude is close to the maximum possible amplitude.
As a rule of thumb, the smallest detectable flaw for a specific ECA probe will be approximately 0.5 to 0.3 times the channel width. Resolution, coil size, drive frequency, topology, surface finish, etc. will have an impact on this value.
The resolution of an ECA probe depends on the coil dimensions and their spacing. This spacing and, by extension, the resolution, must therefore be defined commensurate with the smallest defects to be detected and be finer than target defects.
So, what is the optimal coil size, coil spacing and number of rows for an ECA probe? There is no absolute answer to this question. It is mainly driven by the minimum flaw that needs to be detected by the system. Use the information above to determine which one of our ECA standard probes is best suited for your application, or email us if customization is required to perfectly suit your application!