Hello, when Setrmt sets the RMT sizes to 0, I would suspect that the atoms come unrealistically close to each other. Therefore, I think that there might be an error in your struct-File (this can easily happen when using cif2struct, but I would always put the blame on the cif-file, not the cif2struct-routine). Firstly, check your structure and run nn: What interatomic distances do you get? Are they realistic? If this is ok, set the RMT sizes per hand to some reasonable value. As David wrote already, something around 2 should be all right. If the structure permits you to do so, make the RMTs even larger, as this will save computing time.
If everything fails, supply us with your struct-file (and possibly the cif-file, too). Greetings Michael