Model Steel plate shear walls

I am trying steel plate shear walls. In 2D version however I had to model it using a simplified method called strip model that models the steel plates with ten diagonal tension-only members in opposite direction to the loading. I used spring elements to model tension only members by supplying zero capacity as their compression capacity. the results are reasonable but when I modeled the frames in both direction (i.e almond shapes formed inside the frame) and applied the same monotonic load the results was different. The question is if the springs don't work in compression then why after adding strips in other direction the results were different? By the way I used bi-linear with slackness rule as the hysteresis model for springs. I am not very sure about the values that I've chosen for the hysteresis rule.