Development of a Current-Source Inverter

Project goal

The goal of the project is to develop a current-source inverter (CSI) based on bidirectional GaN switches.

Today, voltage-source inverters (VSIs) are predominantly used in industry. They can be built using only six conventional power semiconductors, such as IGBTs, and are therefore cost-effective. Although VSIs meet the requirements of most drive control systems, they have the disadvantage of generating very steep voltage edges at the output. These damage motor windings and bearings and can cause EMC problems when there are long cables between the inverter and the motor. Filters are typically used in such cases, but they are bulky and increase power loss.

In a CSI, the DC link capacitor is replaced by a choke that carries a constant current. As a result, the motor is no longer subjected to steep voltage edges but only to a chopped current. Due to the inductive behavior of the drive, only a few current and voltage harmonics are generated, which can be smoothed out using comparatively small capacitors.

The power semiconductors in a CSI must be capable of blocking in both directions. For this reason, a CSI has traditionally required twelve power semiconductors, which increases both cost and power dissipation.

Bidirectional GaN switches have been available on the market for a few months now and meet this requirement with a single device. They also have low conduction and switching losses, making them very well suited for use in a CSI.

Project duration

  • November 2026 through March 2027 (5 months): Bachelor’s thesis – Simulation of the overall system
  • March 2027 through August 2028 (estimated) (18 months): Master of Applied Research – Design, construction, and testing of the hardware

Project staff