used silicone grease as the coupling agent. The difference between the 

 wet and dry measurements depended on the smoothness of the concrete surface. 

 The smoother the surface the closer they matched. Consequently, to reduce 

 the error between the wet and dry measurements, silicone grease was also 

 used underwater to improve acoustic coupling. 



The signal detection threshold of the V-Meter also causes erroneous 

 transit time data to be recorded on the digital readout of the instrument. 

 This happens when the amplitude of the first peak of the received signal 

 is below the threshold voltage triggering level of the V-Meter. This 

 effect is illustrated in Figure 16, taken from Reference 12, which also 

 discusses this problem. When the instrument detects a following peak, 

 this causes an apparent transit time increase of one-half wavelength or 

 more. For example, if the sound velocity were 12,000 ft/sec in the con- 

 crete under test and the frequency being used is 54 kHz, the wavelength 

 of the transmitted signal is about 2.7 inches. An error of one-half 

 wavelength under these conditions, over a path length of 1 foot, results 

 in an 11% error in measured sound velocity. A plot of the half -wavelength 

 detection error as a function of path length and pulse velocity at 54 kHz 

 is shown in Figure 17. This error is inversely proportional to the path 

 length and the ultrasonic test frequency. 



Indirect transmission is more prone to errors associated with the 

 detection threshold and the degree of acoustic coupling than direct trans- 

 mission because of the much lower signal strengths. Two actual signal 

 waveforms shown in Figure 18 for indirect transmission further illus- 

 trate the problem. Both signals were transmitted through the same test 

 block, over the same path length, but the coupling for the right waveform 

 was much better than the left waveform as indicated by the received signal 

 amplitude. The digital indication obtained from the V-Meter is shown on 

 each waveform and indicates the detected peak. The difference in measured 

 transit time was 29 microseconds, an error of approximately 20%. There- 

 fore, during all acoustic measurements, silicone grease was used to improve 

 acoustic coupling and the received signal was recorded on an oscilloscope 

 to verify the digital readout from the V-Meter. 



Sound velocity measurements were taken on five concrete test blocks 

 (10x11x24 inches), both dry and submerged in water. Direct transmission 

 was used and the data are tabulated in Table 6 for comparison. Blocks 1, 

 2 and 3 would be rated as "good" concrete while blocks 4 and 5 would be 

 rated 'questionable' according to Table 5. The average sound velocity 

 measured when the blocks were dry was slightly higher than the average 

 sound velocity when the blocks were submerged. The standard deviation 

 of the dry measurements is slightly lower than for the measurements taken 

 underwater. The trend of slightly higher averages, coupled with lower 

 standard deviations for the dry measurements compared to the same measure- 

 ments taken underwater was attributed to better acoustic coupling. 

 However, as expected, there are no significant differences in measured 

 sound velocity between the wet and dry measurements. 



Laboratory tests were conducted using indirect transmission to eval- 

 uate the ability of ultrasonics to detect cracks in concrete that are 

 around 1/32 of an inch wide and of varing depth. Several test specimens, 

 each with a different depth crack, were made for the evaluation. The 

 depths of the simulated cracks varied from 0.5 to 2.25 inches deep and 

 the measured direct sound velocity through each specimen averaged about 



13 



