Origin for Acoustic Emission Testing

"Origin is so broad, I can apply it in any situation. It is one of my go-to tools..."

Gregory Daspit, a researcher at Southern Research Institute, has been using Origin in his research since 2006. He has a wide range of tools and software at his disposal; however he finds that most of these cater to specific areas of research analysis unlike Origin, which he notes:

“Is so broad, I can apply it in any situation. It is one of my go-to tools”.

Mr. Daspit works on Non Destructive Testing (NDT) for evaluating material characterization and precision testing of materials and instruments. His recent addition to the Origin graph gallery was based on a project involving the use of modal acoustic emission (MAE) as an NDT for predicting the onset of damage during the manufacture of circuit boards. His goal was to use Origin to create a graph that would visually display the effects of spherical bending tests. Just like a tree branch makes a cracking sound when breaking, this test relies on the sound waves produced as material cracks under stress to predict the pattern of failure. Mr. Daspit hoped to show his client as well as other researchers and technicians who have been working with these sorts of tests for years, when and in what pattern the cracks develop between the package and circuit board in a static graph. The circuit boards were placed on eight pads for support (only four are shown in Figure 1), while a central pin applied bending pressure to the board. Four piezo-electric sensors picked up the waves which were then analyzed by third party software that output text files as voltage versus time, and energy & location per event. These files were imported into Origin for further analysis. Mr. Daspit set the circuit board on a Cartesian coordinate system with (0,0) in the center. This allowed him to split the circuit board into four regions or quadrants. Much of the analysis was done using formulas in the Set Column Values dialog; Mr. Daspit set query conditions on the formulae to pull out the information specific to each quadrant by setting < and > conditions. He also made use of the Stack and Unstack Column feature for splitting data columns.

The processed data was then used to create a graph depicting the top view of the circuit board. Mr. Daspit also added the positioning of the piezo-electric sensors as well as the bending pads using Objects. The corners of the package (i.e. chip) were denoted as P1, P2, P3 and P4, and the cracking was denoted as energy events. The higher the energy, the larger the event and the greater the amount of energy released during the crack propagation. This information was plotted as scatter plots on the actual location of occurrence since the board was set on a Cartesian coordinate. The colors denoted the time of the event occurrence and the symbol size was set to reflect how great the event was. This way the plot could be used to gain a spatially visual understanding of the magnitude, time of occurrence and specific location of the damage.

Figure 1, shows the crack wave propagated from the top left corner to the bottom right corner, as the symbols go from green to red. The most intense damage was sustained along those edges where the largest symbols are located. Near the end of the test, the damage was localized in the center of the board and characterized by smaller energy events as denoted by the smaller red symbols in the center, likely showing damage directly at the loading location. Electrical open is a term used to mark a point when the damage is sufficient to break the built-in daisychain circuit and is used to determine damage onset. The graph of the test also allowed the viewer to gauge the time and likely damage location that electrical open occurs- in this case around 82.1 seconds halfway between the top, left and right sides of the board. As Mr. Daspit noted “This is one of those plots that to me provides an overall sense quickly and yet shows more and more behavior of the system the longer it is pondered.”

The testing yielded ten to fifteen samples. Mr. Daspit was able to use the many duplication techniques in Origin to reproduce the graph for all the samples without manually creating each graph from scratch. He felt he was successful in helping his client understand the process and thus improve understanding of damage between a chip package and the circuit board over time under stress. The plot also helped Daspit and his colleagues be more selective about their data acquisition. The research and plot was also part of a paper presented at the 63rd IEEE ECTC at Las Vegas, NV on May 28-31, 2013. Origin’s publication quality graphs were exported in the required format for the paper and the useful Push to PowerPoint feature was used for presentations. The graph opj along with the data is accessible in a zip format on our Graph Gallery page.