In Faddeev calculations of 3 H we study the dependence of the binding energy on the three nucleon force. We adopt the 2π-exchange Tucson-Melbourne three-nucleon force and investigate phenomenologically the dependence on the strength of the individual three-body force operators (a-, b-, c-and dterms). While the a-term provides a tiny contribution the b-and d-terms are important to gain the experimental binding energy. We find two solutions for the c-term, one around the value used in the Tucson-Melbourne model and a new one close to zero, which supports the recent recommendation of chiral perturbation theory that the short-range c-term should be dropped.