I found that one can fairly simple produce small simulation sample using only BFC with geant maker. You may probably need private build of gstar if you need something which is not in the default version. My macro and condor submitter for that can be found in the CVS: offline/users/mnaglis/Sim. I did simulation of the phi->KK with 100% BR to this channel and everything looks fine. BFC options are taken from: http://drupal.star.bnl.gov/STAR/starsimrequests/2010/apr/05/beam-energy-scan-request-auau-collisions-energy-77-gev-39-gev-and-115-gev

Later I found that all TOF variables are equal to -999. This is because I didn't have a chain option which activates simulation for TOF. Also I could not generate phi meson with zero width with gstar. I have an idea how to generate and decay phi with external generator and then read list of dougter particles in my run_sim.C macro, but this is not ready yet.