Access Official Technical PDF

Share Summary:LinkedInEmail

A radiation-hardened GaN converter can help address a persistent challenge in space power electronics: delivering efficient, compact conversion while operating reliably in a radiation environment. A demonstrated 500 W isolated design illustrates how radiation-tolerant gallium nitride (GaN) field-effect transistors (FETs) can be applied in a phase-shifted full-bridge (PSFB) converter with synchronous rectification. The converter operates from a nominal 100 V DC bus and regulates its output to 20 V.

The design is relevant beyond spacecraft. It highlights engineering trade-offs that also matter in other high-reliability power systems: component qualification, switching stress, magnetic utilization, gate-drive timing and the need to distinguish a promising prototype from a fully qualified product.

Why radiation changes converter design

Space power electronics must contend with single-event effects (SEE), total ionizing dose (TID) and displacement damage dose (DDD). These mechanisms can damage semiconductor devices, shift electrical characteristics or produce transient signals that interfere with control and protection circuits. In a switching converter, a false gate command or an unintended conduction path can have serious consequences.

Conventional silicon power MOSFETs can be made radiation tolerant, but the design changes and qualification demands create trade-offs in size and electrical performance. Radiation-hardened GaN FETs offer a potential alternative. Their device structure and demonstrated radiation tolerance make them promising candidates for space power stages, while their switching performance can support higher-efficiency converter designs. Device selection still needs to be based on the relevant radiation environment and qualification requirements; the use of GaN alone does not establish system-level radiation qualification.

Why choose a phase-shifted full bridge?

Converter topology determines how effectively the transformer is used, how much stress the switches experience and how complicated the control becomes. Single-ended designs such as flyback and forward converters are established options, but they use the transformer core over one quadrant of its magnetic operating range. Double-ended topologies, including the full bridge, drive the transformer in both directions and are better suited to higher power levels.

A full bridge uses four primary-side switches, which makes gate timing and isolation more involved than in simpler topologies. It also requires careful consideration of fault conditions, including the possibility of simultaneous conduction in switches in the same bridge leg. These concerns become particularly important in radiation environments, where transient effects can affect control signals.

The PSFB approach controls power transfer by adjusting the phase relationship between the bridge legs. This enables zero-voltage switching (ZVS) under suitable operating conditions: a switch can turn on when the voltage across it has been brought close to zero. Reducing hard-switching transitions can lower switching loss and device stress. The design work compared a phase-shifted converter with a conventional hard-switched full bridge, making ZVS and phase-shift modulation central considerations rather than simply changing the switching devices.

Coordinating the rectifier, control and magnetics

On the transformer secondary, the converter uses a current doubler rectifier with synchronous GaN FETs. A current doubler uses two output inductors and alternates current between the rectifier paths. Synchronous rectification uses controlled transistors in place of passive rectifier devices. This arrangement requires accurate coordination between primary-side switching and secondary-side gate signals to ensure the intended current paths are established.

That coordination places practical demands on the gate-drive and control circuits. The design addressed isolated gate-drive requirements, signal conditioning, dead-time control and the generation of phase-shifted switching signals. Dead time—the interval that prevents opposing switches from conducting simultaneously—must be managed carefully: insufficient dead time risks shoot-through, while timing choices also affect switching transitions and ZVS operation.

Current sensing is another important control function. The implementation considered sensing methods for peak-current-mode control, including the challenges of capturing the switching current signal and delivering a usable signal to the controller. In space-oriented designs, these circuits must be assessed not only for normal operation but also for their response to radiation-induced transients and component variation.

Magnetic components also influence overall performance. The converter evaluation compared transformer winding constructions and different resonant-inductor implementations. This matters because ZVS depends on the interaction of circuit energy and parasitic or added inductance, while winding and core choices affect magnetic losses and physical design. The work’s comparison reinforces a practical point: semiconductor selection does not determine converter efficiency on its own. Transformer, inductor, rectifier, switching frequency and control behavior all contribute.

What the 500 W design demonstrates—and what it does not

The prototype delivered a nominal 500 W from a 100 V DC input bus to a regulated 20 V output. Its design process included power-stage calculations, loss analysis, PCB layout considerations and experimental comparisons between the PSFB and hard-switched full-bridge approaches. It also examined paralleling GaN FETs and evaluated magnetic configurations.

These details make the work useful as an implementation reference for engineers considering GaN in isolated power conversion. They show that the transition involves more than substituting a GaN device for a silicon MOSFET: the gate driver, layout, current sensing, switching strategy and magnetics must be engineered as a system.

However, the work did not address the failure mode and effects analysis (FMEA) or the protection circuitry required for complete radiation qualification of the PSFB topology. A laboratory prototype and a radiation-hardened transistor are not, by themselves, evidence that a complete converter meets a mission’s qualification requirements.

Practical implications for power-system designers

For engineers and system integrators evaluating a radiation-hardened GaN converter, the main takeaway is to treat topology, devices and control as a connected design problem. A PSFB can combine efficient transformer use with phase-shift modulation and opportunities for ZVS, but realizing those benefits depends on switching timing, resonant behavior and careful hardware implementation.

Before adopting this approach, teams should define the radiation and reliability requirements, assess the device and peripheral-circuit qualification evidence, and validate gate-drive behavior, current sensing, magnetic losses and fault responses across the intended operating conditions. GaN can expand the options for space power conversion, but system-level analysis and qualification remain essential to turning that potential into a dependable converter.

Access Official Technical PDF

About The Author