Synchronous Rectifier Loss⚠ unverified

Understanding the Parameters

• $R_{ds}$ — channel resistance when gated on. Falls with die size / process; rises with temperature and with lower $V_{gs}$. The live calculator labels $R_{ds}$ as dimensionless; it is ohm.

• $I_{out}$ — treated as constant DC load current. For significant inductor ripple, prefer $I_{rms}^2 R_{ds}$ over the SR on-interval (see Conduction Loss).

• $D$ — on-time fraction of the synchronous device. In a buck, the low-side SR is on for roughly $1 - D_{high}$; here $D$ is whatever fraction you assign to the SR conduction window.

• $P$ — heat in the SR MOSFET package.

Derivation (Approaching a Proof)

While the SR is on, $p(t) \approx i(t)^2 R_{ds}$. With $i(t) \approx I_{out}$ constant and the device on for fraction $D$ of each period,

$$P = \langle i^2 R_{ds} \rangle = I_{out}^2\,R_{ds}\,D.$$

This is Conduction Loss ($I_{rms}^2 R_{on}$) with $I_{rms}^2 \approx I_{out}^2 D$ for a rectangular on-window at current $I_{out}$. Compared with Diode Forward Loss $V_f I_{avg}$, sync rectification wins when $I_{out} R_{ds} \ll V_f$ (typical below ~5–12 V outputs at moderate current).

Not included: gate-drive power for the SR, body-diode conduction during dead time (Dead Time Loss), reverse recovery, or switching loss if the SR is hard-commutated.

History

Synchronous rectification became mainstream as CPU and POL voltages fell (sub-2 V rails) where a 0.4 V Schottky drop was an unacceptable fraction of $V_{out}$. Discrete SR controllers and integrated DrMOS stages now standard in computing and telecom power.

Related Concepts: Conduction Loss, Diode Forward Loss, Dead Time Loss, Buck Converter Duty Cycle, Efficiency Total

Notes: Registry calculator synchronous-rectifier-loss (unverified). DC-current approximation. Registry unit bug: $R_{ds}$ labelled dimensionless — should be ohm.