If possible, please provide us more information about your calculations, e.g., what are the exact numbers for the high bias or small bias voltages mentioned by you? what kind of system is under your study?
Actually, if the applied bias voltage is too high, the use of Landauer-Buttiker approach could be not valid. In this forum there are several threads on the convergence problem of the calculations under high bias voltage, so you could check them by searching "high bias".
Since the calculations under small bias voltages can be converged, you could start from the converged case and gradually increase the bias voltage.