DEAD-TIME ELIMINATION FOR VOLTAGE SOURCE INVERTERS

inShare

A novel dead-time elimination method is presented in this paper for voltage source inverters. This method is based on decomposing of a generic phase-leg into two basic switching cells, which are configured with a controllable switch in series with an uncontrollable diode. Therefore, dead-time is not needed. In comparison to using expensive current sensors, this method pre­cisely determines the load current direction by detecting which anti-parallel diode conducts in a phase-leg. A low-cost diode-con­duction detector is developed to measure the operating state of the anti-parallel diode. In comparison with complicated compen­sators, this method features simple logic and flexible implementa­tion. This method significantly reduces the output distortion and regains the output RMS value. The principle of the proposed dead-time elimination method is described in detail. Simulation and ex­perimental results are given to demonstrate the validity and fea­tures of this new method.

1. Introduction

To avoid shoot-though in voltage source inverters (VSI), dead-time, a small interval during which both the upper and lower switches in a phase-leg are off, is introduced into the standard pulse width modulation (PWM) control of VSIs. How­ever, such a blanking time can cause problems such as output waveform distortion and fundamental voltage loss in VSIs, es­pecially when the output voltage is low.

To overcome dead-time effects, most solutions focus on dead-time compensation by introducing complicated PWM compensators and expensive current detection hardware. In practice, the dead-time varies with the gate drive path propa­gation delay, device characteristics and output current, as well as temperature, which makes the compensation less effective, especially at low output current, low frequency, and zero current crossing. Several switching strategies for PWM power converters have been proposed to minimize the dead-time effect. A dead-time minimization algorithm was also discussed earlier to improve the inverter output performance. A phase-leg configu­ration topology proposed prevented shoot through. However, an additional diode in series in the phase-leg increases complexity and causes more loss in the inverter. Also, this phase-leg configuration is not suitable for high-power inverters because the upper device gate turn-off voltage is reversely clamped by a diode turn on voltage. Such a low voltage, usually less than 2 V, is not enough to ensure that a device is in its off-state during the activation of its complement device.