Fig. 5 shows C-scan images taken from the as welded strip for welds made at two different currents. Both images have a distinctive outer circle and substantial areas of 0% signals (coloured white). The circle and the area of 0% signal for the weld made at 30 kA current are bigger than that at 22 kA. The circular feature in each C-scan image is consistent with depicting a spot weld nugget, as in planar view this would be expected to be roughly circular and higher current is accompanied with a bigger weld. However, the significant areas of zero signals meant that the images could not be unambiguously interpreted. Zero signals indicated that a proportion of the reflected signal, which should have reached the detector, had been lost. It was suspected that loss of signal was due to sound waves being refracted away from the detector. According to the Snell's Law, when sound waves travel at an angle from one medium to another with different travelling speeds, refraction occurs. In the situation described here, the sound waves travel from the scanner's transducer through the inbuilt water tank coupled to the aluminium surface by gel. As sound in water travels at an average speed of 1482 m/s; whilst in aluminium is of 6420 m/s, the spherical nature of the interface refracted the sound waves away from the detector. To compensate for some loss of signal it is possible to increase the gain of the system, but this can result in more noise being introduced; and in this instance did not enable the true weld nugget diameter to be resolved. It is therefore suggested that, whilst the images of the as welded condition relate to the spot welded areas, the information in the images are influenced by the surface indentations and do not depict the true nugget diameters. To prevent signal loss and improve the possibility of imaging the true weld nugget, removal of the indentation to provide a flat surface for the scanner was considered.