The main problem I encountered when modeling as a truss member is the fact that I also need to model the sag of the cables. The problem is that when I model the sag and discretize the continuous parabolic shape of the cable into a series of straight elements and make them axial only members, there is a loss of stiffness at the joints. This is easily understood since I have two members framing into one joint at different angles and I say that they can only take load along their centerline. The forces do not line up, hence there is no stiffness there as the computer sees it. Because the stiffness of the cables are very small when compared to the supporting poles, I have modeled as a fixed connection at all points in order to model. This leads to some moments in the cable, but nothing major, especially when compared to the poles.