CT Scanning Moves from Lab to Production Line

In addition to raw computing power, new microfocus source designs and intuitive software make this technology ready for the production line.

Related Suppliers

If your quality control operation could benefit from X-ray vision, micro computed tomography (CT) scanning deserves a look. 

Or, perhaps, a second or third look. After all, the capability to “see inside” parts—to form detailed, cross-sectional images from multiple X-rays taken from different angles—is not new. What is new is that this technology is no longer too slow for high-production manufacturing operations. That goes for not only conducting the scan, but also selecting the proper scan settings, reconstructing the image and interpreting the data. This was the message of a recent presentation from Andrew Ramsey, X-ray CT specialist at Nikon Metrology, who showed a real-time video demonstration of a small, complex component undergoing the entire micro CT scanning process in less than one minute (video below). 

Such speed largely can be attributed to general advances in computing, Mr. Ramsey says. Faster CPUs, for example, enable capturing more X-ray frames more quickly than ever before, and higher-capacity hard drives with improved read/write speeds can store an unprecedented amount of data. Even the video game industry has inadvertently contributed with the development of high-end, programmable graphics cards that can divide the task of reconstructing scan data with the primary CPU. “During a game, the computer has to take a three-dimensional scene and generate 2D views for the players in real time,” Mr. Ramsey says. “This massively parallel processing is, essentially, designed to do CT reconstruction.”  

Courtesy of Nikon Metrology, this video demonstrates the speed capabilities of modern industrial CT scanning systems.

Yet, Nikon is also leveraging a development that has nothing to do with computing: microfocus X-ray sources capable of higher flux, a measure of the number of X-rays projected per second. The lower the flux, Mr. Ramsey explains, the longer the scanned object must rotate within the X-ray beam to produce a more detailed cross-sectional image—that is, an image with higher contrast between areas that absorb X-rays and areas that allow them to pass through and impact the detector. At maximum flux, the only options for increasing data-collection speed are lowering the resolution or limiting the number of X-ray projections used to compile the scan.

The primary limitation to X-ray flux is heat, and that goes for both the tiny, microfocus sources used in industrial CT scanning and the larger X-ray tubes in systems at, say, the dentist’s office or airport security. In both cases, a heated cathode projects a beam of electrons toward an anode target (Nikon’s microfocus sources use tungsten targets). Upon impact, more than 99 percent of the electron beam’s energy converts to heat. The remainder converts to X-ray photons, which are then projected toward the object to be scanned. The higher the power, the greater the density of the electron beam, and the greater the flux. At a certain point, however, this risks melting the target.

In a microfocus source, this limitation is particularly pronounced. To achieve the level of detail required for industrial CT scanning, these tiny sources concentrate the X-rays—and thus, the initial electron beam—into a far smaller spot size, one measured in microns rather than millimeters. The more concentrated the electron beam, the greater the target’s vulnerability to heat. “Tungsten melts at around 1 watt per micron,” Mr. Ramsey says. “That translates to 1 gigawatt per square meter, and I certainly wouldn’t want to be standing in that square meter.”

To better control heat and maximize flux, Nikon’s new anode targets don’t remain stationary during operation, as previous designs did. Instead, the electron beam is focused on the chamfered edge of a disc-shaped beam target that spins as fast as 6,000 rpm. A rotating disc, Mr. Ramsey explains, dissipates heat better than a stationary target because it provides more surface area for the beam to contact. That enables cranking the power as high as 5 watts per micron to achieve higher flux, or, as Mr. Ramsey puts it, “to get a lot more X-rays per second into the same, small spot size.”

Whatever the target design, and whatever the capabilities of the computers collecting and reconstructing the data into a cross-sectional image, setting up a scan in the first place can still take time. Likewise for analyzing and interpreting the data. Yet, Mr. Ramsey says software interfaces have come a long way toward streamlining both processes. Nikon’s Inspect-X system, for example, automatically stores CT scan settings like voltage, current, manipulator position and so forth in templates that can be re-used on similar applications. Similarly, data-analysis macros can enable employees without deep knowledge of the process to easily perform tasks like extracting regions of interest from a template or applying geometric dimensioning and tolerancing (GD&T) to determine pass/fail status, among other examples.

He adds that companies employing software developers can go even further with simplifying complex CT scanning tasks. That’s thanks to Inter-Process Communication (IPC), a published, programmable interface to Inspect-X that enables creating an individualized “front end” for the software with simple programs in Visual Basic, C++ or C#. IPC also allows interaction with third-party hardware and software, a capability that enables integrating external databases, robotic arms or other peripherals.

Mr. Ramsey emphasizes that speed hasn’t been the only advance in CT scanning during the past few decades. For instance, 2003 marked Nikon’s first use of amorphous silicon flat panel X-ray detectors with 3,200 by 2,300 pixels, which enabled far more detailed images than the 1,000- by 1,000-pixel CCD cameras used in prior systems. Yet, time-pressed manufacturers could rarely take advantage of this full capability, Mr. Ramsey says. Instead, most opted to scan at lower resolution, to use high resolution only for specific areas of interest, to have an outside party perform CT scanning, or to employ an entirely different technology for quality control. Now, the full capability of this technology is within reach. “With one smooth process from mounting the sample to getting the analysis report, we can confidently say that micro CT scanning is ready for the production line,” he says.

Related Content

Use Statistical Learning to Catch Surface Defects

Mectron’s Qualifier (Q-5000) inspection system uses its radial fire array for dimensional inspection of a variety of parts.