drop to second order from the larger first order tension fluctuations that are characteristic of slack 

 cables. The influence of strut rigidity upon the measured in-air (structural) damping is discussed in 

 further detail later in this Appendix. 



The measured in-air damping is generally referred to as the structural damping because it is 

 assumed that external energy losses to the air or supports are negligible when compared to the cable's 

 internal losses. The validity of this assumption depends upon the magnitude of the measured damping 

 and upon the compliance of the cable supports. Estimates of the fluid losses in air can be used to check 

 that part of the assumption, but losses to the supports are difficult to determine or even to estimate. In 

 the present experiments, test runs with the same cable for two strut rigidities made it possible to iden- 

 tify changes in the damping due to the very different support conditions. This task was complicated by 

 the nonlinear behavior of the in-air damping, and to discount this nonlinearity the log decrements for 

 the Uniline and DAS cables are presented in Figs. C3 and C4 for approximately the same initial ampli- 

 tude (0.1" or 2.5 mm). In both cases the data generally.correspond to taut cable conditions, although at 

 the lowest tensions the DAS cable behavior is in the range of slack cable effects (see Appendix B). 

 From the results in Fig. C3 it appears that the Uniline cable is not influenced by variations in the rigi- 

 dity of the struts. The DAS cable behavior is significantly dependent on the rigidity and, as shown in 

 Fig. C4, the measured damping is lower when the struts are more flexible. The relative importance of 

 strut rigidity in each case is consistent with the natural frequency measurements in Figs. CI and C2. 



Any influence of the initial amplitude could not be detected in one or even several cycles of the 

 vibration decay. The procedure adopted was to take 150 to 450 consecutive vibration cycles and to 

 compute the log decrement of the damping over various segments of the record. The lengths of the 

 record and therefore the lengths of the segments were dictated by the resolution of the measuring sys- 

 tem and the rate of the vibration decay. 



The structural log decrements obtained in this manner appeared to be a simple linear function of 

 cable amplitude, but it is quite possible that this linear behavior is due to the limited range of 



141 



