Toolbox for 3D acoustic imaging of manufactured electronic circuits

A signal processing toolbox has been realized to enhance the acoustic signal processing of manufactured electronic systems. The toolbox contains a range of signal processing algorithms to help better differentiate the 3D scanned acoustic signals acquired from manufactured parts. The processing has been used to assess the through-life reliability of systems by non-destructively sampling systems during extended reliability tests. The algorithms are based on a variety of novel signal processing methods such as wavelet, time frequency domain imaging, dictionaries and Fast Fourier Transform (FFT) for comparison. The paper will explore how the various algorithms can be tuned to test samples using the designed toolbox. The toolbox has a graphical user interface (GUI) designed in MATLAB. It can control the processing and display of 3D image data “cubes” acquired from an array of A-Scans of a sample obtained by an acoustic microscope. In this case a Sonoscan Gen6 C-SAM acoustic microscope has been used, but if the data format is known the toolbox can be applied to any 3D cube of data. The data sizes can be large, so processing times can be of the order of minutes for 1k × 1k × 1k array. Recently the execution times have been speeded up considerably by employing parallel algorithms and multicore processing. Some routines now only take seconds which is a very significant speed-up.