[Wien] Optimize and Geometry minimization

[yxl at email.jlu.edu.cn]

Dear ,
   I encounter some questions about Optimize and Geometry minimization,my optimize procedure is (1) loop over different coa's at constant volume
   Since what i'm calculating is impurity system,so I should do geometry minimization.In optimize.job I use  min -I -j "runsp_lapw -orb -so -cc 0.001 -i 40 ", In this procedure ,"RMT is too big ,mini step is restricted" is often appeared in case_cao__-5.5.  But when I reduce Rmt ghost bands appeared. And i know s-o coupling is considered only within the atomic spheres (About this I don't fully understand, would you like to tell me?),so the Rmt should be big .
this is a contradiction,I expect your reply,Thank you !
                                               hongxia