Optimised input filter design and low-loss switching techniques for a practical matrix converter

The matrix converter permits frequency conversion in a single-stage process. The perceived disadvantage of the matrix converter is that conduction losses are high. However, semisoft current commutation and optimal sequence switching can be used to minimise commutation losses so that at high switching frequencies the total losses in the matrix converter can be less than those in a conventional rectifier-inverter combination. The viability of the matrix converter depends to a large extent on the size and cost of the input filter components required to meet international power quality standards. In the paper filter designs are examined and guidelines established. Practical tests have been carried out on a 3.5kW converter to validate computer models. It is concluded that the matrix converter is viable if the right combination of semiconductor switching techniques and input filter design are employed.