It was found by experimentation and subsequent breaking of cracked specimens that, in the case of "tight cracks", different signal amplitudes were obtained from specimens containing the same size of cracks, and actual crack area was not proportional to the signal amplitude observed.

These two observations on the unreliability of assessment of tight cracks led us to believe that there should be a minimum amount of tightness of cracks above which the area - amplitude relationship is valid and the tight cracks can be reliably detected and assessed.

With this in view, a series of experiments were designed to find out the validity of this assumption and to estimate the minimum gap between the crack faces for reliable assessment of tight surface cracks.

The experiments consisted of generation of different sized tight surface cracks in maraging steel tensile specimens, monitoring the ultrasonic shear wave response of the cracks under increasing tensile loading till failure of the specimens and establishing from the response data, the relationship of signal amplitude versus tensile stress.

Results showed that the ultrasonic response from the tight cracks increases with the stress, reaches a maximum value and then remains constant till the failure of the specimen. This observation identified a low level stress value at which the ultrasonic response reaches the maximum amplitude. It was concluded that the increase in ultrasonic response till this maximum value was due to the release of tightness of the crack. It is found that at this stress level or beyond, the scatter in crack detectability due to the amount of tightness vanishes. An attempt was made to find the gap between the crack faces at this stress level with the help of fracture mechanics theories.

Results showed that transmission of shear waves at 4 MHz frequency through the crack faces ceased to exist at a gap size of the order of 0.0002 mm between the crack faces. The amount of ultrasonic shear wave transmission through the crack faces increases with the reduction of this gap (or the increase in the amount of tightness of the crack) and gives rise to varying ultrasonic response (unreliability in the detection and assessment). If the gap width is more than this threshold value, there is no transmission through the crack faces and there is complete reflection of shear waves resulting in the validity of area-amplitude relationship. It was also found that under this low stress (corresponding to the maximum amplitude) the detection of smaller size cracks is possible i. e. application of stress increases the sensitivity of crack detection.