Advancing Energy’s Journey Through Semiconductor Innovation
Higher power density is critical to improving digital connectivity without compromising our electronic devices, energy sustainability and reliability.
With the world consuming energy at an unprecedented rate, the demand for more data and more connected devices will not diminish in the near future. It’s an energy issue: how can you get more with less power? Data centers, electric vehicles, consumer electronics, medical equipment and other industrial applications demand larger power supplies, but the systems must also reduce their size, weight, environmental impact and cost.
Increased power density – defined as the ability to pack more power delivery capacity into a smaller volume – is critical to improving our electronic devices, energy sustainability and digital connectivity, without compromising reliability. The best way to efficiently address power density challenges is through innovations in semiconductor technology. Semiconductors have never had a great opportunity to change the way energy flows.
Addressing power-efficiency needs in server PSUs
Gallium nitride (GaN) is gaining popularity due to its ability to deliver higher power density and better performance than traditional silicon metal oxide semiconductor field effect transistors (MOSFETs) and insulated gate bipolar transistors (IGBTs). Especially in server and telecommunication power applications, gain reduces inductor size and improves efficiency in power-factor correction (PFC) and DC/DC converter stages. In recent years, server power-supply unit (PSU) density has increased from 40 W/in 3 to 90 W/in 3. Increasing user demand for more power has avoided increasing PSU size. Power-circuit efficiency is an important factor for managing power dissipation in current PSU volumes and demands have increased from 93% to >96%. In fact, the European Union has set an efficiency target of >96.5% by 2023 for all data center AC power supplies.
Server and data center PSUs are beginning to adopt GaN and Silicon Carbide (SIC) FETs to improve the efficiency of their systems. Although there is some overlap in the power levels delivered by gain and seek, the gain is basic characteristics that make it a good and durable fit for high power density applications. While gain-based architectures can perform well at very high switching frequencies, inductor and magnetic component sizes may be significantly reduced in the future as well. Separately, these applications have a higher track of further increasing power density with gain than other competing technologies.
Complex power topologies enable higher efficiency and power density
To limit waste heat in AC/DC PSUs and achieve high power levels in increasingly small form factors, designers have introduced advanced architectures such as bridgeless PFC circuits. High-frequency switches in a bridgeless PFC topology require little or no reverse-recovery charge to reduce switching losses; Therefore, a wide bandgap device (SiC or GaN) is required in a bridgeless PFC topology. For the discrete DC/DC stage of the PSU, designers are moving from hard-switching converters to soft-switching converters to reduce magnet size and increase efficiency. In this case, SiC or GaN offer better switching coefficients than silicon devices, leading to better devices at as high frequencies as possible. As shown in the figure below, switching from a traditional bridged PFC rectifier to a bridgeless PFC rectifier reduces the total number of diodes and components, helping to reduce efficiency and size.
Unlocking the full potential of GaN
R-integrated gain devices are unique in providing overcurrent and overtemperature protection, undervoltage lockout, and voltage and current sensing that cannot be achieved with discrete power switches. With integrated drivers, Texas Instruments Gain FETs have switching speeds up to 150 V/ns. This switching speed, combined with a low inductance package, reduces losses, enables clean switching and reduces ringing. This enables engineers to achieve switching frequencies in excess of 500 kHz, resulting in up to 60% lower magnetization and increased performance, reducing system costs. Combined with gain, real-time control microcontrollers (MCU) such as the TI C2000 family offer advantages such as maximizing gain-based power solutions and complex, time-critical processing; accuracy control; and software and peripheral scalability. In addition, they can fully unlock the potential of gain-based power solutions by supporting different power design topologies and higher switching frequencies to maximize the power efficiency of the MCU design.
As we look to the future, the world’s appetite for energy will continue to grow; Technology has brought great changes and improvements. The impact of this development is increasing global demand for innovation and performance excellence in applications such as data centers and electric vehicles. Anticipating consumer needs, adapting to industry trends and quickly adopting new technologies are critical for efficient and secure energy flows.