Can Rfree go up during B-value refinement?

[Raman]

  Can the B-factors ever be over-refined? In other words, theoretically
  
  is it possible for the R-free to increase during B-factor refinement

  in X-PLOR ?  Erroneous B-values could increase R-free, but the

  question is, would the R-free increase if 'nstep' is increased to
  
  say, 100?  


  chumma at purdue.edu