54 could lose gate control, turn on the BJT and eventually destroy the device. The configuration of an IGBT has evolved and improved with trenchgate field-stop (TGFS) technology. Implanting a back emitter and fieldstop in the IGBT allows for better control of the dynamic behaviour. This allows a thinner die, and thus a higher cell density. This results in improved performance with lower conduction and switching loss, greatly increased robustness and significantly reduced thermal resistance. Most of the loss in an IGBT is from switching loss, which is far less than a MOSFET. To address the limitations of the traditional MOSFET, the superjunction MOSFET structure has the trenches in the electron layer. This means that when voltage is applied the depletion layers expand horizontally and merge to form a depletion layer equal to the depth of the trenches, forming the superjunction. For example, the 600 V CoolMOS S7A superjunction MOSFET from Infineon is suitable for slow-switching applications, such as HV eFuse, HV eDisconnect and on-board chargers. Meanwhile, the 650 V CFD7A adds a fast body diode to make it suitable for hard-switching and resonant-switching topologies, on-board chargers, HV-LV DC-DC converters and auxiliary power supplies. Then there is the 800 V C3A family, which can be used for auxiliary automotive power supplies. The automotive p-channel MOSFET offers products in 30 V, 40 V, 55 V and 150 V, with the lowest RDS(on) at 40 V and the highest current capabilities. The automotive n-channel MOSFET is available in a wide variety of packages, starting from 3 mm x 3 mm up to a size of 10 mm x 15 mm. A halfbridge package for 40 V motor-drive applications fits into the newest topside-cooled 10 mm x 15 mm for highest power-density applications. However, for full-bridge designs, the routing with a dual MOSFET is complicated and requires a large printed circuit-board area. Additionally, the current rating in the dual packages is limited to 20 A. Using the integrated half-bridge, the PCB area is sizeably reduced, as it provides enhanced routing for bridge applications with a wide RDS(on) range from 3.0 mΩ to 7.0 mΩ, and an increased current rating of 60 A. Gallium nitride “There is a big macro shift with the power and cost of lithium batteries coming down,” says Alex Lidow, chief executive of GaN chip designer EPC and one of the original pioneers of the technology. “When that happens that goes right into the sweet spot of GaN with 26-48 V charging, 80 V and 96 V, and that is where GaN blows away the power MOSFET. “The second thing is the motor drive and GaN has made huge changes there. Brushless DC motors have been the highest efficiency for battery systems at 20 kHz and down because of the MOSFET. It has this diode recovery that takes a while, and the motor has a dead time requirement of 500 nanoseconds (ns) for the MOSFET. The higher the frequency, the less inductance you need in the motor for the same amount of power, and that’s weight, so the motors are lighter.” The dead time from the MOSFET is a counter force in the motor and takes away 5-6% of the power delivery to the motor. Then, as the motor drive shrinks, the electrolytic capacitors can be replaced with ceramic capacitors. The electrolytic capacitors traditionally are large and temperature-sensitive, and they are the first thing that fails. “In some cases, the GaN motors are a quarter of the size of the MOSFET, and we routinely see them at half the size as a result of GaN,” says Lidow. These are also audibly quieter and have lower EMI. “So far, the biggest successes have been at 100 kHz; all-ceramic capacitors, smooth motor drive with a smooth sinusoidal waveform – that’s the sweet spot. There’s a lot of work being done on motors for higher frequencies, and you see air-core motors at 250 kHz or 500 Hz. Where I think we will see those is in things that fly. A few of the eVTOL [electric vehicle take-off and landing] platforms have eight motors, each 200 A; the bus voltages are 96 V and that’s a good spot for GaN, so it’s not out of range,” he adds. Lidow points to a laboratory in Torino, Italy, which is developing a large motor for eVTOL designs. “If it flies it’s going to be GaN,” he says. EPC developed enhanced-mode GaN HEMT transistors, which start with a thin layer of aluminium nitride (AlN), grown to isolate the device structure from the substrate. On top of this, a layer of highly resistive GaN is grown. This layer provides a foundation on which to build the GaN transistor. Aluminium gallium nitride (AlGaN) is applied to the GaN, and this layer produces a physical strain, which attracts electrons to the interface. March/April 2024 | E-Mobility Engineering Integrating power devices into a PCB (Image courtesy of Schweizer)
RkJQdWJsaXNoZXIy MjI2Mzk4