[phenixbb] occupancy refinement when alternate conformers clash

Hi, I am refining an RNA structure where on set of base pairs has alternate conformers, and the B alternates from the paired bases clash. That is, the structure should be viewed as A alternate in one base paired with B alternate in the other base. The refinement blew both B alts out of the density to prevent their clash. How can I fix this? Thanks for your help, Kendall Nettles