The performance of an isolated DC-DC converter is degraded by a persistent DC bias in the transformer current. Existing DC bias elimination methods estimate the bias either by directly sampling the terminal voltage ripple waveform, which is susceptible to switching noise and requires precise measurements, or by using additional sensors in the ac network, which add to the complexity and cost of implementation. This paper proposes a DC bias elimination strategy for isolated DC-DC converters based on a band-pass filtering of the terminal voltage ripple. The relation between the DC bias and the amplitude of the fundamental (switching-frequency) component of the ripple is analyzed. Based on this, a DC bias elimination technique is proposed, which suppresses switching noise in the ripple and can be implemented using analog circuitry and low-resolution sensors. The steady-state and transient performance of the proposed method is validated in simulations and experiments on a ∼ 150 W dual active bridge (DAB) converter.
