I am working towards a transmission built out of library components, and am having issues with the clutch;  I am connecting a rotational inertia through a clutch to a signal-driven ideal torque source, setting the clutch control signal (fn) to close the clutch at (t =1 sec), and simulating for 5 seconds.  The inertial component begins accelerating at (t =0) despite the fact that there should be no transmitted torque. Fn is zero until (t =1), and the torque (as I understand it) is a function of fn.

Is this a problem anybody else has encountered?  I think the torque is also related to the mode, and there might be a problem with the initialization values, or I might have just selected parameter values that are unrealistic.