Reply To: Response Spectrum Sign Loss

[dlooman]

I don't see how that could work.  The sign of the node and element results could be plus or minus depending on the mode.  You could create your own combination method in which you algebraically sum the model results without doing any squaring.