slickdeals,
Would it be needed to do the static analysis first and, after that, to scale the spectrum cases to reach 85% of the static base shear force? If, from analysis, you get >90% horizontal vibrating mass accelerated, then you will get the base shear force = >90%*mass*acceleration. This is more than 85%. So, the problem reduces to requesting sufficient number of eigenmodes to be found by the program, so the influence of the higher modes is considered in analysis. The number of modes I usually request is [3*sqrt(number of floors)]. But, to be overconservative, one can request [3*(number of floors)], which means [(translation+translation+rotation)*(number of lumped masses at story levels)].