The truth about matrix converters

Eureka’s article on Yaskawa Electric’s new matrix converter drive in April (page 49) caused quite a stir amongst other drives manufacturers. Eureka therefore decided to ask the heads of R&D at these companies for their views on the technical advantages of matrix converter drives and whether manufacturing drives based on the technology is economical. Dr Bill Drury, executive VP technology at Control Techniques, was the first to challenge Yaskawa’s ‘breakthrough’. He commented: “In the early 1990s, we undertook a three-year project with the Universities of Bristol and Birmingham, with DTI support, to consider the relative merits of matrix converters and other novel converter topologies, compared with the conventional circuit configuration. “As part of this, we built and tested a 5.5kW drive using IGBT switches, soft switching techniques and optimised input filter designs. The results of this study were interesting. First, there is no need for large reactive energy storage parts as there is no DC link. Also, you get inherent four-quadrant operation without the need for extra components. Control can readily ensure that the supply currents drawn are near-sinusoidal, except for switching frequency harmonics. Finally, most control algorithms will allow the input displacement factor to be set to unity, as in the conventional system, or in some cases at a leading value. “However, the output voltage is limited to 87% of the input voltage. Techniques exist for overcoming this limitation, but at the cost of added complexity and deterioration in the supply current waveform quality. Also, conduction losses in a matrix converter are inherently higher than that for an equivalent rectifier-inverter circuit. But, by adopting soft switching techniques, losses can be reduced to around those of the conventional topology,” he explained. He also pointed out that an input filter is needed to eliminate the high switching frequency harmonics from the supply and provide compliance with EMC regulations. And, while regenerative braking is inherently possible, the supply, he said, must be receptive to the braking energy. “It is not possible, as is the case in a conventional rectifier-inverter circuit, to include controlled braking with the energy being discharged into a braking resistance. Such a scheme will provide braking torque in conditions where the supply is lost or unable to accept the energy such as a generator supply.” Real cost challenges He also advised that care needs to be taken in the consideration of short circuit capability. While a basic circuit diagram of a matrix converter shows just nine switches, the practical realisation of a single switch is of two IGBT’s plus two diodes. “This presents a real cost challenge,” he warned. Drury concluded that Control Techniques “saw no opportunity to be able to offer any improved performance with matrix converters over the conventional rectifier-inverter topology”. He said that there is no inherent capability to change the applied voltage vector any quicker and since the three-year project, the company’s view hasn’t changed. But he did admit that the matrix converter does have some merit: “The limited output voltage was, and is, the most significant drawback, when considering its use as a drive for general application. This would not of course be an issue for a custom solution such as an integrated motor drive, where the motor winding voltage could be selected. The absence of a DC link removes the need for significant reactive energy storage components, however, the holy grail of an all-static solution isn’t achievable since an input filter is essential.” ABB, another major drives player, broadly agreed with Drury. In a joint statement to Eureka, Matti Kauhanen and Roelof Timmer, R&D managers at ABB Finland commented: “Today, matrix converters do not provide a cost effective alternative for state-of-the-art IGBT PWM inverters [two-level VSIs]. The firm is actively involved in research projects and has concluded that matrix converters need to be further developed before they can compete with state-of-the-art VSI technology, if ever.” ABB’s statement concluded: “The matrix converter does not have any significant performance advantage over IGBT PWM inverters, but instead, shows a number of disadvantages, is less reliable, more complex and more costly. Therefore, ABB believes that IGBT PWM inverters will dominate for years to come.”