The aerospace industry is undergoing a profound power architecture shift. For decades, radiation-hardened (rad-hard) silicon MOSFETs were the undisputed workhorses of spaceborne DC-DC converters. However, as modern satellite payloads demand higher efficiency, tighter density, and lower mass, silicon is hitting its physical limits.
Enter gallium nitride (GaN) high electron mobility transistors (HEMTs). Boasting a wider bandgap, superior electron mobility, and inherent robustness against certain radiation mechanisms, GaN is the key to unlocking next-generation space power densities. Yet, translating these raw material advantages into flight-ready, highly reliable hardware requires a deep understanding of its unique driving requirements and its interaction with rad-hard pulse width modulation (PWM) controllers.
The rad-hard landscape: Silicon vs. GaN
Space radiation hazards generally fall into two categories: Total ionizing dose (TID) and single event effects (SEE).
In silicon MOSFETs, TID causes a build-up of trapped holes in the thick gate oxide, leading to a severe negative shift in threshold voltage (Vth) and increased leakage current. On the other hand, GaN HEMTs lack a traditional gate oxide, utilizing a Schottky or p-GaN gate structure instead. Because there is no oxide to trap charges, rad-hard GaN devices exhibit exceptional inherent tolerance to TID, often surviving exposure well past 100 krad(Si) to 1 Mrad(Si) with minimal parameter shifts.
While GaN shines under TID, SEE is where the engineering nuances lie. Silicon MOSFETs are susceptible to single event burnout (SEB) and single event gate rupture (SEGR) due to heavy ion strikes creating parasitic bipolar conduction paths or destroying the gate oxide.
But GaN HEMTs don’t suffer from traditional SEB or SEGR because they are majority-carrier devices without parasitic bipolar structures. However, they are prone to single event transients (SETs) and catastrophic degradation at high drain-to-source voltages (VDS). Under heavy ion bombardment, localized high electric fields near the drain can cause high-current leakage paths. Consequently, a VDS derating of 30% to 50% of the maximum rated voltage is standard practice for spaceflight GaN applications.
Alex Lidow, CEO and co-founder of Efficient Power Conversion (EPC), notes the physical limitations of legacy materials: “Silicon has had a glorious 60-year run, but it has hit its theoretical performance wall. In space applications, where every gram of weight translates directly to launch cost, GaN isn’t just an alternative—it’s an absolute architectural necessity.”
Driving the delicate GaN gate
The primary challenge when designing a flight-ready GaN converter is managing the gate drive. Silicon MOSFETs typically feature a comfortable gate threshold of 2 V to 4 V and can tolerate gate voltages up to ±20 V. GaN HEMTs are far less forgiving:
- Low threshold voltage: GaN devices typically turn on at a mere 1.5 V to 2.0 V.
- Fragile gate rating: The absolute maximum gate-to-source voltage (VDS) is often capped at a tight -5 V to +6 V.
- Ultra-low gate charge (QG): GaN switches an order of magnitude faster than silicon. While this minimizes switching losses, it introduces severe dv/dt and di/dt sensitivities.
If the gate drive circuit experiences even minor parasitic inductance, the rapid dv/dt transition can couple back through the device’s Miller capacitance (CGD), generating a transient voltage spike on the gate. If this spike exceeds 1.5 V, it triggers a catastrophic spurious turn-on (shoot-through), potentially destroying the power stage. Conversely, if the gate driver overshoots beyond 6 V due to ringing, the gate permanently degrades.
Interfacing with heritage PWM controllers
Because dedicated rad-hard GaN-integrated drivers are still emerging, power engineers frequently pair discrete rad-hard GaN FETs with established, flight-proven rad-hard analog PWM controllers.
These heritage controllers were designed to drive the heavy, capacitive gates of silicon MOSFETs, delivering output swings from 0 V to 12 V or higher. Interfacing these high-voltage controllers with a delicate 5-V GaN gate requires a meticulously designed intermediate drive stage and layout discipline.
- Voltage clamping: Direct connection is catastrophic. Engineers must employ a high-speed level shifter or a dedicated, rad-hard gate driver buffer (for example, ISL71020M) that accepts standard PWM logic levels and provides a tightly regulated 5-V drive output.
- Asymmetric gate resistance (RG): The gate resistor network must be split into separate turn-on (Rgon) and turn-off (Rgoff) paths via a diode-resistor network. Rgon is optimized to purposefully slow down the turn-on dv/dt to suppress gate ringing. Rgoff is kept near 0 Ω to provide a low-impedance hold-down path, ensuring the gate remains firmly below the threshold voltage during rapid drain voltage transitions.
- Dead time management: GaN HEMTs lack a native body diode; so, they conduct in reverse through the channel when the gate is off. During this dead time, the reverse voltage drop (VSD) can be quite high (2 V to 3 V). If the PWM controller introduces excessive dead time, efficiency penalties from reverse conduction negate the switching advantages of GaN. Therefore, designers must utilize PWM controllers with highly precise programmable dead-time control or implement an external low-forward-drop Schottky diode in parallel.
In my years managing power electronics design, I have watched countless clean schematics fall apart under the oscilloscope simply because an engineer treated a fast wide-bandgap loop layout like a legacy 100-kHz silicon board.
When marrying a heritage 12-V PWM architecture to a 5-V gate, your layout must be an absolute work of art. Parasitic inductance can easily destroy the gate on the very first pulse if the loop area isn’t locked down.
Packaging innovation and real-world use cases
The unique performance GaN metrics are actively reshaping the size, weight, and power (SWaP) equation across various orbital profiles.
In satellite bus power, utilizing isolated GaN-based flyback or forward topologies allows engineers to push switching frequencies past 500 kHz—up from the standard 100 kHz legacy limit—directly translating to a 60% reduction in magnetics volume. In point-of-load (POL) converters, synchronous buck configurations supply core logic rails for spaceborne FPGAs and deep-space processing computers with near-zero reverse recovery losses. Furthermore, in space robotics and motor control, three-phase GaN inverter stages enable compact, motor-integrated electronics housings that bypass heavy shielded cabling.
To support these high-frequency applications, manufacturers have developed innovative, low-inductance packaging structures specifically engineered to eliminate the internal bond wires that cripple traditional high-reliability packages.
Vendor landscape and flight heritage
Navigating the space-qualified GaN marketplace requires examining components that meet the rigorous screening levels required for aerospace reliability. Below are three design case studies.
- EPC Space
The EPC Space family of devices commands a long track record in commercial-volume rad-hard discrete GaN deployment. Its enhancement-mode (eGaN) discrete FETs (such as 60-V EPC7014) and integrated power modules have accumulated an impressive lineage, with thousands of devices actively operating in orbit since January 2019. These components are heavily utilized in LEO small-sat constellations and GEO communication platforms, powering intermediate bus converters and high-speed pulsed laser drivers for autonomous LiDAR systems.
Figure 1 EPC Space’s hermetic packaging replaces traditional wire bonds with broad, low-profile bottom contact pads to eliminate loop inductance. Source: EPC
- Infineon Technologies
Infineon’s CoolGaN family represents the entry of traditional military-standard (MIL-PRF-19500) JANS-grade rigor into the wide-bandgap space ecosystem. Leveraging its heritage in spaceborne silicon MOSFETs, Infineon modernized GaN packaging by removing wire bonds entirely. It PowIR-SMD package delivers a 49% footprint reduction compared to legacy housings and reduces internal parasitic package inductance by 97%—dropping to a mere 0.1 nH. This nearly eliminates internal gate ringing, allowing clean interfacing with fast PWM controllers.
Figure 2 The PowIR-SMD technology minimizes internal parasitic inductance down to 0.1 nH via an advanced die-free construction. Source: Infineon
Regarding the validation of these advanced architectures, a high-reliability engineering expert at Infineon stated: “Removing internal wire bonds was the final frontier for GaN in space. By developing a die-free, surface-mount package like PowIR-SMD, we didn’t just solve thermal management—we completely neutralized the parasitic gate inductance that historically caused engineers to shy away from high-speed wide-bandgap switches in critical flight hardware.”
- Renesas
Renesas entered the space-qualified GaN arena by executing a brilliant ecosystem play: combining its decades-long legacy of rad-hard analog power management with high-reliability GaN FETs like ISL73024SEH. These devices are frequently designed alongside dedicated multi-phase synchronous PWM controllers—such as the ISL73847SEH—to form the fundamental backbone of spaceborne core power bricks.
Figure 3 The Renesas/Intersil radiation-hardened GaN family is engineered in rugged, space-qualified ceramic flatpacks optimized for severe thermal environments. Source: Renesas
Achieving mission success
While designing deep space profiles, we must ruthlessly enforce gate-clamping rules. Designing radiation-hardened GaN converters for space applications is ultimately an exercise in managing extremes. GaN offers game-changing thermal and volumetric efficiencies, but its unforgiving gate drive margins and extreme switching speeds mandate a departure from legacy silicon layout rules.
The accumulated flight data from the GaN industry pioneers has rewritten the rulebook for space power systems, dispelling early engineering anxiety through years of anomaly-free orbital operation. When carefully controlled by a robust, rad-hard PWM controller, these devices provide a mature, predictable, and remarkably rugged path toward achieving unparalleled power density in the cosmos.
Bharrat Mehta, a senior space scientist, is former deputy project director of Indian Space Research Organization (ISRO).
Related Content
- GaN Transistor for Space Missions
- Infineon unveils rad-hard GaN transistors
- Deep Space Challenges Rad-Hard IC Development
- Test and Measurement Priorities for GaN-Based Power Electronics
- GaN in Space: Unlocking Efficiency and Performance in Satellite Systems
The post Rad-hard GaN converters: Design insights for space hardware appeared first on EDN.