Is there a bug in 2011 SP5.0 for the energy norm error in a nonlinear study?  I get ERR = 1 everywhere, and ERR is not unitless.