Gallium-nitride switches shrank the primary side of USB-PD chargers, but the output stage still turns on a silicon Schottky: at 5–20 V and a few amps, the forward drop is the loss, the PDFN56 package is the heat path, and synchronous rectification is the only serious competitor. This guide walks the PPS load profile, the package, and the comparison.
The GaN Question: Where Silicon Still Fits
GaN switches at higher frequencies with lower switching loss, which shrank the transformer and the charger itself. The marketing story is about the primary side; the engineering story is that the output rectifier never moved. The secondary side still rectifies the transformer output at 5–20 V, and at those voltages the conduction loss of the diode—not its switching speed—is the dominant term.
The result is a division of labor: GaN on the primary for the switching, silicon Schottky on the secondary for the conduction. The output rectifier is not the part GaN replaced; it is the part that still does the low-voltage, high-current job better than a wide-bandgap device would.
The GaN story also raised the switching frequency, which made the output stage’s rectifier loop shorter and the transformer smaller—but the diode’s conduction loss at 5–20 V did not change with the frequency. The frequency win happened on the primary; the secondary still lives in the conduction world, which is why the selection method for the output rectifier is the same as it was before GaN.
The comparison also explains why wide-bandgap is not on the secondary: a GaN or SiC rectifier at 5–20 V would pay a higher forward drop for blocking capability the output never needs, and the switching advantage is irrelevant to a diode whose job is conduction. The material choice follows the voltage and the current, not the marketing.
Output Stage in a USB-PD PPS Design
Programmable power supply (PPS) adds another wrinkle: the output voltage is adjustable in small steps, and the load profile sweeps across the 5–20 V range. The output rectifier must perform across that whole range, not at one point, and its leakage at the highest reverse voltage belongs in the budget even when the load is at the lowest voltage.
The PPS profile also raises the light-load question: the charger spends time in standby and at light load, where the diode’s fixed drop is a big fraction of a small output. The low-VF part helps exactly there, which is why the efficiency story of a USB-PD charger is written by the output stage as much as by the primary.
The PPS voltage sweep also tests the leakage budget at the top of the range: at 20 V the reverse bias is highest, and the leakage at the hot junction belongs in the standby-power number. The datasheet curve is read at the working reverse voltage and temperature, not scaled from the cold figure.
The PPS current steps add a dynamic test: the output jumps between load steps, and the rectifier sees the step’s surge and the thermal transient together. The design validates the step response at the worst transition, because the PPS spec tests exactly that behavior.
PDFN56 for Thermal Density in Slim Chargers
The slim charger has no room for a heatsink, and the output rectifier’s heat must leave through the board. The PDFN56 exposed pad turns the PCB into the heatsink: a copper pad matching the pad, thermal vias into inner planes, and controlled solder volume make the package perform to its thermal data. The footprint rules from the PDFN design guide are the charger’s thermal design.
The same footprint serves the whole current range: the 5 A AMBRP5100 for light outputs and the 10 A AMBRP10H100 for the 100 W class, so a product family scales the current class without changing the layout.
The slim charger’s board is a thermal puzzle: the primary side’s GaN switch and the secondary’s rectifier share the same small PCB, and the rectifier’s heat must not drive the switch’s temperature up. The board layout separates the two thermal zones, and the rectifier’s pad and vias keep its heat on its side of the board.
The thermal puzzle is solved with the same tools as the electrical one: copper area, via grids, and controlled solder volume are the rectifier’s thermal design, and the case-temperature measurement at the sealed condition is the proof. The slim charger does not get a heatsink; it gets a well-designed board.
Loss Comparison: Silicon Schottky vs SR at 5–20 V Outputs
| Item | Silicon Schottky | Synchronous rectification |
|---|---|---|
| Loss mechanism | IF × VF at the load current | I² × RDS(on), duty-weighted |
| Typical drop at 5 A | 0.5–0.8 V class | Millivolts to tens of millivolts |
| Control burden | None | Gate drive, dead-time, controller |
| Failure mode | Passive, predictable | Drive fault possible |
| Best fit | Moderate targets, simplicity | Aggressive efficiency targets |
The table is the comparison in one view: synchronous rectification wins the loss column but pays a control burden, and the boundary is the efficiency target and the current. The full comparison lives in the synchronous rectification article; the charger takeaway is that the silicon Schottky remains the default until the target justifies the complexity.
The comparison also has a reliability angle: a diode cannot fail from a gate-drive fault, and its failure mode is predictable, while a synchronous MOSFET adds a control dependency that the charger’s firmware must manage. The complexity budget and the reliability budget are part of the same decision.
Component Selection Summary: AMBRP10H100 and AMBRP5100
For the USB-PD output band, the two PDFN56 parts cover the range: AMBRP10H100 (10 A / 100 V) for 5–10 A outputs and AMBRP5100 (5 A / 100 V) for lighter designs, both with the same footprint. The 100 V class covers the 20 V output rail with room for transients, and the low forward drop wins the conduction story across the PPS range. The AMBRP10H100 product page and the AMBRP5100 page carry the datasheets, and the board design follows the PDFN rules.
The selection summary is also the ordering rule: pick the current class for the output band, confirm the leakage at the top of the PPS range, and let the efficiency target decide whether the synchronous option is worth its complexity. The PDFN footprint keeps the layout stable across the family.
The two parts also cover the mechanical story: the same reel format, the same pad, and the same assembly process, so a product family scales the current class without a second qualification.
The selection also names the ordering rule clearly: the 100 V class covers the 20 V rail with margin across the whole PPS sweep, and the footprint keeps the layout stable whether the design lands on the 5 A or the 10 A part. The output stage’s part number is the last decision, not the first.
Design note. The loss-comparison table uses class-typical Schottky drops and the general synchronous-rectification loss model; the exact numbers come from the selected parts at the working current and temperature. The GaN-versus-silicon division of labor describes the architecture, not a claim about any specific charger.
Frequently Asked Questions
Does GaN replace the output rectifier?
No. GaN replaced the primary-side switches; the output stage still rectifies at 5–20 V, where conduction loss—not switching speed—dominates, and a silicon Schottky wins that job.
Why does PPS change the rectifier selection?
The output sweeps across 5–20 V, so the rectifier must perform across the range, and its leakage at the highest reverse voltage belongs in the budget at every output voltage.
How does the heat leave a slim charger?
Through the PDFN56 exposed pad into board copper and thermal vias. The footprint design is the charger’s thermal design.
When should I use synchronous rectification instead?
When the efficiency target exceeds what the diode delivers and the controller, gate-drive, and dead-time cost is justified. At moderate targets the silicon Schottky wins on simplicity.
Which part fits a 65 W output?
The 10 A / 100 V AMBRP10H100 for 5–10 A outputs, or the 5 A AMBRP5100 for lighter designs, both in the same PDFN56 footprint.
Conclusion
In the GaN era, the USB-PD output stage is still a silicon Schottky story: the low-voltage, high-current conduction job belongs to a low-VF part, the PDFN56 package carries the heat, and synchronous rectification is the only competitor worth the complexity. Pick the current class, follow the footprint rules, and let the efficiency target decide the rest.
Review the AMBRP10H100 and AMBRP5100 product pages on the Good-Ark site, and contact Good-Ark with your output range, current, and efficiency target for a rectifier recommendation.