Automation: Its name is software

An x-ray inspection system obviously needs hardware that can acquire and transmit images quickly and accurately for processing. But the explosion of computing power in inspection systems has dramatically increased the importance of software, both for preparing the inspection regimen at one end and for analyzing the generated images at the other. And as happened with conventional test equipment a generation ago, software is rapidly becoming the “great differentiator” in inspection applications.

Software used for developing inspection applications must make it easy for engineers to set up numerous standard inspection scenarios and then redefine the scenarios as needed. Users also need the ability to instruct an inspection activity to be more or less comprehensive without changing the inspection program.

Consider manufacturing activities related to preproduction and ramp-up to full volumes. The number of defects for a given process should fall with time. Therefore, after the initial ramp-up, the production line may not require a full-blown inspection step most of the time. A comprehensive test on a subset of units and a cursory test on the rest should suffice.

This scenario is similar to a sampling scheme, but it allows a manufacturer to look for specific failures in particular parts of every assembly without unnecessarily taxing production-line throughput. An available “on-off switch” can permit users to decide how much of an inspection step to execute depending on changing circumstances.

With those goals in mind, phoenix|x-ray in Germany, working with Contax in the UK, developed the Xe2 analysis software. The Xe2 software (X-ray image Evaluation Environ-ment) is a graphical development system that generates various measurement scenarios for analyzing 2-D images in real time.

During development of the inspection procedure, users can define new quantitative tasks, such as determining the mean gray-scale value of an image detail, as well as statistical analyses, such as the mean of a group of measurements. The x-ray images are taken with a 12-bit image intensifier or with a high-contrast (16-bit) digital detector. By providing a graphical representation of the results, the system can provide a more visual picture of the quality of the product and the overall efficiency of the manufacturing process than could the raw data alone.

Fig. 1 The red arrow indicates a solder bridge. The blue arrows highlight nodes with insufficient solder. Courtesy of phoenix|x-ray.

The phoenix|x-ray system includes some standard modules that address common failure mechanisms. A so-called quad flat-pack (QFP) module examines gull-wing solder joints on any type of device package, determining misregistration of the lead, the joint width, and the existence of the heel fillet and the toe fillet. It also measures the gray scale of the heel fillet and the solder joint itself, and it detects voids in the solder and solder bridges between joints on the board. Other standard modules inspect any type of solder joints on a quad flat-no-lead (QFN) package and classify plated-through holes.

Fig. 2 In this detail, the blue arrow shows a joint with no solder. The amber arrows highlight solder voids. Courtesy of phoenix|x-ray.

Also critical to the success of any software environment is the user interface. The interface must be easy to understand and must make it easy for users to set up tests and hone them as conditions change. Figure 3 shows a graphical display of the test setup.

Fig. 3 This graphical representation shows the QFP test setup. The module permits the classification of gull-wing solder joints according to their width and average gray scale. A bridge check is also performed. A check of the heel fillet is optional. Courtesy of phoenix|x-ray.