In an AC-DC adapter, the output rectifier’s forward drop is the last big lever left for efficiency, and every 0.1 V at the output current is a watt of heat in a sealed case. This guide follows the loss through a 65 W USB-PD output stage, compares the Schottky with synchronous rectification, and lands on a component choice.
Efficiency Regulations Keep Raising the Bar
Efficiency standards for external power supplies have tightened in successive generations, pushing adapters toward higher average and light-load efficiency. The trend is regulatory and commercial at once: a more efficient adapter ships with a smaller case, passes the next standard generation, and costs less to run. The article does not promise compliance with any specific rule—the requirement set changes by region and year—but the design direction is universal: every milliwatt of loss is visible in the efficiency number and in the case temperature.
The output rectifier is the visible lever because its loss is a fixed drop times the load current, present at every load. The design implication is practical: the drop is measured at the load current and the hot condition, and the efficiency number on the datasheet is an average across the load range, so the part is chosen for the points the standard weighs rather than for a single condition.
The light-load story is the second half of the trend: at 10% load the output current is small, but the efficiency number at light load is measured and published, and the rectifier’s drop still costs a fixed fraction of the small output. The drop that looked acceptable at full load can be the difference at light load, which is why the low-VF part earns its place across the whole load range.
The light-load number also explains why the output stage determines the adapter’s average-efficiency story: the drop is a fixed cost on a small output, so the low-VF part helps exactly where the standards weigh hardest.
Where a 0.3 V VF Difference Goes
The output stage carries the full load current, so the forward drop multiplies directly into watts. A 65 W adapter at 20 V delivers 3.25 A; a 0.3 V difference in forward drop is about 1 W of loss—roughly 1.5% of the output power, sitting in a sealed plastic case. At light load the same drop is a smaller wattage but a bigger fraction of the output, which is why light-load efficiency targets push the drop lower still.
The loss is dissipated in the rectifier and must leave through the package and board; a low-VF part shrinks both the efficiency gap and the thermal load at once. The physics of the drop belongs to the forward voltage guide; the adapter takeaway is that the drop is the design variable. The comparison between parts is therefore read at the same current and temperature: a part that looks better at 10 A may not be better at the adapter’s 3.25 A, and the VF at the working condition is the only fair number.
Output Rectifier Options: Schottky vs Synchronous Rectification
Two options compete at the output. A low-VF Schottky rectifier is simple, passive, and reliable; synchronous rectification replaces the diode with a MOSFET driven by the controller, cutting the drop to an RDS(on) product at the cost of a driver, a controller channel, and dead-time management.
The boundary is the efficiency target and the output voltage: at 20 V and 3.25 A, the Schottky’s 0.5–0.8 V drop costs roughly 1.6–2.6 W, and synchronous rectification can recover most of it—but only if the added circuit loss and cost are justified by the target. At lower currents or looser targets, the Schottky wins on simplicity and cost; at aggressive efficiency targets, the comparison is covered in the synchronous rectification article. The diode selection here still establishes the baseline.
The reliability angle favors the passive part: a diode has no gate drive, no dead-time window, and no controller dependency, so it cannot fail from a drive fault. The synchronous MOSFET’s efficiency advantage comes with a control burden, and the product’s complexity budget is part of the same decision. The gate-drive and controller cost also scales with the number of synchronous channels; a single-output adapter pays one channel, a multi-output design pays several, and the comparison changes with the architecture.
Design Example: 65 W USB-PD Adapter Output Stage
| Item | Value |
|---|---|
| Output | 20 V / 3.25 A (65 W) |
| Rectifier candidate | AMBRP10H100, 10 A / 100 V Schottky, PDFN56 |
| VF reference | 0.76 V typical at 10 A and 25 °C (conservative upper bound for this current) |
| First-pass loss | 3.25 A × 0.76 V ≈ 2.5 W (conservative) |
| Heat path | PDFN56 exposed pad into board copper and vias |
The 10 A rating leaves margin over the 3.25 A output, the 100 V class covers the output rail with room for transients, and the PDFN56 pad turns the board into the heatsink. The first-pass loss uses the 10 A datasheet point as a conservative bound; the final budget reads VF at the working current and adds the leakage at the hot junction, then confirms the case temperature on the prototype. For a lighter output, the AMBRP5100 (5 A / 100 V, same footprint) covers the 1–3 A band in the same PDFN56 layout.
The board design completes the example: a copper pad matching the exposed pad, a thermal via grid into inner planes, and stencil apertures that control solder volume. The PDFN rules turn the package’s 3 °C/W junction-to-case resistance into a usable system number, and the case-temperature measurement on the first build confirms it. The thermal result depends on the enclosure: a plastic adapter case with no airflow relies entirely on the board and the case surface, so the pad and via design carry the whole thermal story, and the prototype measurement at the sealed condition is the number that matters.
Component Recommendations: AMBRP10H100 and AMBRP5100
The two PDFN56 parts cover the adapter output band: AMBRP10H100 for 5–10 A outputs and AMBRP5100 for 1–5 A, both 100 V class with the same footprint so a design can scale the current class without a layout change. The board design follows the PDFN rules—pad matched to the exposed pad, thermal vias into inner planes, stencil apertures controlling solder volume—so the package performs to its thermal data.
The output stage’s switching node also deserves a layout note: the rectifier loop from the transformer secondary through the diode to the output capacitor is kept small and symmetric, because the loop inductance rings at every switching edge and shows up in the output noise. The layout guidance follows the rectifier layout guide; the adapter takeaway is that the low-VF part and the tight loop work together.
Design note. The 65 W example uses the AMBRP10H100’s 10 A datasheet VF as a conservative upper bound for the 3.25 A output; the final loss budget reads VF at the working current and junction temperature, adds the leakage term at the working reverse voltage, and confirms with a case-temperature measurement on the prototype. The 1.5% efficiency figure is the arithmetic of 1 W on 65 W, not a compliance claim.
Frequently Asked Questions
Why is the output rectifier the efficiency lever in an adapter?
Because it carries the full load current and its loss is a fixed drop times that current, present at every load. Every 0.1 V at 3.25 A is about 0.3 W in the case.
When should I use synchronous rectification instead?
When the efficiency target exceeds what the diode can deliver and the added controller, driver, and dead-time cost is justified. At moderate targets the Schottky wins on simplicity and cost.
What rectifier fits a 65 W USB-PD output?
A 20 V / 3.25 A output is served by a 10 A / 100 V Schottky like the AMBRP10H100 in PDFN56, with margin for the rail and transients; the 5 A AMBRP5100 covers lighter outputs in the same footprint.
How is the loss dissipated?
Through the exposed pad into board copper and thermal vias. The PDFN footprint rules—pad match, via grid, stencil apertures—decide whether the package performs to its thermal data.
Does the leakage term matter at 100 V?
At the working reverse voltage and hot junction it belongs in the budget; read the datasheet curve at the operating condition rather than scaling the 25 °C number.
Conclusion
The adapter output is a forward-drop race: the drop multiplies by the load current into watts, and the watts land in a sealed case. Choose the lowest-drop class that fits, size the PDFN board for the heat path, and let the efficiency target decide whether synchronous rectification is worth its complexity. The race has a floor, though: below a certain current the fixed overhead of the synchronous controller exceeds the diode loss it saves, which is why the passive part stays in low-current products.
The adapter takeaway is the same at every power level: define the efficiency target, measure the drop at the working current, and let the passive-versus-synchronous comparison follow the arithmetic rather than the fashion.
The output stage is where the adapter’s efficiency story is written, and the rectifier is the pen.
Review the AMBRP10H100 and AMBRP5100 product pages on the Good-Ark site, and contact Good-Ark with your output voltage, current, and efficiency target for a rectifier recommendation.