Which is better, a Schottky or a fast recovery diode? The answer is decided by gates, not by a topology slogan. A Schottky stores no PN-junction minority carriers, so it has essentially no PN-style reverse recovery and can beat an FRD even in hard-switched stages when its voltage, leakage, capacitance, surge, and thermal gates close; the FRD becomes the alternative when one gate fails. This article compares the two families across the voltage range with that gate-based procedure.
The Gate-Based Comparison
Family comparisons fail when they start from the device instead of the requirements. Six gates decide between a Schottky and an FRD, and they should be checked in order:
- Reverse voltage. The maximum repetitive peak must fit the Schottky’s class with 20–30% margin; above its ceiling the FRD or SiC is the only candidate.
- Leakage at temperature. Reverse leakage at the operating junction temperature and reverse voltage must stay inside the loss and stability budget.
- Junction capacitance and displacement current. The Schottky’s turn-off is dominated by junction capacitance; at high dv/dt the displacement current can ring against layout inductance—measured, not assumed.
- Forward drop. Conduction loss is IF × VF; the Schottky’s advantage lives here.
- Surge capability. IFSM is a non-repetitive survival margin checked against the actual inrush waveform.
- Cost and supply. System cost and availability close the list.
The topology and waveform shape how each gate is evaluated, but they do not replace the gates. The same Schottky can be the right answer at 40 °C ambient and the wrong one at 80 °C, because leakage grows with the temperature the part actually reaches—the gate list keeps the working condition in the decision.
The gate order also prevents the most common selection error—choosing by the VF column first. VF is the visible number, but voltage, leakage, and displacement current can eliminate a family before conduction loss is ever compared.
Recovery Physics, Stated Correctly
The physics separates the families cleanly. A Schottky conducts through a metal-semiconductor barrier with majority carriers only, so it stores no minority carriers in the junction; it therefore has essentially no PN-style reverse recovery, and its turn-off is dominated by junction capacitance and the resulting displacement current. A fast recovery diode is a PN device whose stored minority carriers must clear at every turn-off, producing a finite recovery charge (Qrr) and a recovery-time window (trr). The full waveform anatomy belongs to the reverse recovery guide; the decision-relevant fact is this: the FRD is the family that carries a recovery charge, not the Schottky.
The consequence corrects a common rule of thumb. Hard switching does not inherently favor the FRD—if anything, the recovery charge the FRD must clear is exactly the burden the Schottky does not carry. The Schottky’s real hard-switching questions are whether its displacement current rings the node at the actual dv/dt, whether its leakage holds at the hot junction, and whether its voltage class fits. Those are gate questions, not topology verdicts.
The distinction matters for the datasheet too: an FRD’s recovery numbers are minority-carrier parameters that grow with temperature and recovered charge, while a Schottky’s trr-type figures reflect junction capacitance under the measurement conditions. Compare the two families’ recovery columns only after reading what each column actually measures.
When the FRD is used in a hard-switched stage, its recovery loss is a real cost: Psw ≈ Qrr × VR × f sizes the commutation-path event. Worked example: a 200 V hard-switched stage at 100 kHz with 200 nC of recovered charge contributes about 4 W of switching loss to the path—the split between the diode and the switching device depends on the circuit waveform. The FRD’s advantage is its controlled recovery characteristic, which limits ringing when the Schottky’s displacement current does not.
The Gates Across the Voltage Range
| Voltage band | Typical starting point | What closes the choice |
|---|---|---|
| ≤100 V outputs | Schottky | Leakage at the hot junction |
| 100–250 V | Both families | Reverse-peak margin, leakage, capacitance ringing, surge; FRD if a gate fails |
| 250–600 V | FRD (SiC entering) | Schottky voltage ceiling; recovery and cost |
| >600 V | FRD or SiC | Voltage class, temperature, frequency |
| Line frequency | Standard silicon / bridge | Voltage and surge |
| PV bypass | Low-VF Schottky | Junction-box heat-flow route |
The bands are starting points based on where the gates typically close, not rules. A 150 V hard-switched stage can still favor a Schottky if its displacement current, leakage, surge, and thermal gates all close; a 48 V stage can still favor an FRD if a Schottky’s leakage fails the sealed-box stability budget.
How Topology and Waveform Enter
Topology shapes the gates instead of replacing them. In a hard-switched stage, the FRD pays a recovery charge at every edge while the Schottky’s junction capacitance produces displacement current—the comparison is measured at the real dv/dt, not read from a topology label. In a soft-switched stage, the resonant tank removes most of the recovery transient, so conduction loss dominates and the Schottky’s lower forward drop wins directly. At line frequency, recovery is irrelevant and voltage and surge decide.
Walkthrough: a 48 V adapter output at 8 A hard-switched at 100 kHz. The Schottky conducts at roughly 0.45 V—about 3.6 W—and its displacement-current signature is verified at the node; the FRD at roughly 1.0 V costs 8 W before recovery is added. The gates close for the Schottky, so it wins even in hard switching. The same procedure applies to the AMBRP10H100 class at 100 V and the 250 V class at the top of the silicon range.
The dv/dt and di/dt of the actual waveform set how large the displacement current is and how much the FRD’s recovery interacts with layout inductance; a waveform measurement with a current probe and a differential voltage probe on the switching node turns the gate discussion into numbers.
A Selection Walkthrough
Three scenarios show the gate order. First, a 48 V, 8 A adapter output: the reverse-peak margin closes, leakage at the hot junction is inside the budget, the displacement current is benign at the measured dv/dt, and the surge and thermal margins hold—the Schottky wins. Second, a 400 V hard-switched industrial stage: the voltage gate alone removes the silicon Schottky from the room, so the comparison is FRD versus SiC. Third, a sealed 200 V stage at 80 °C ambient: the Schottky’s leakage term climbs with temperature, and if the stability check fails, the FRD is the alternative despite its recovery cost. Each outcome comes from the gates, not from the voltage class name.
Engineering note. The gate-based method above reflects how recovery loss enters a hard-switched budget (Psw ≈ Qrr × VR × f sizes the commutation-path event, with the diode-switch split depending on the waveform) and how conduction loss dominates where recovery is removed (PD = IF × VF). The AMBRP10H100 and AMBR40250S references are datasheet-published examples on the Schottky side; any FRD comparison column must come from the specific candidate’s datasheet. Run both candidates with maximum values at the operating temperature, and confirm the displacement-current signature and thermal result on a prototype.
Frequently Asked Questions
Is a Schottky always better than a fast recovery diode?
No. When the gates close—voltage, leakage, capacitance, surge, thermal—the Schottky often wins, including in hard switching, because it carries no PN-style recovery charge. When a gate fails, the FRD is the alternative.
Does hard switching favor the FRD?
Not inherently. The FRD is the family that carries a recovery charge; the Schottky’s hard-switching limits are displacement-current ringing, leakage at temperature, and voltage class—each measured at the working condition rather than assumed from a topology label.
When does recovery loss dominate?
For the FRD in hard-switched, high-frequency, high-voltage stages: the energy scales as Qrr × VR × f, so a 200 V stage at 100 kHz pays recovery loss at every edge. At line frequency the term is negligible and voltage and surge decide.
What decides the 100–300 V band?
The six gates at the working condition: reverse-peak margin, leakage at temperature, junction capacitance and displacement current, forward drop, surge, and cost. Topology shapes each gate but does not replace them; run the loss and ringing checks on the actual waveform.
When should I choose SiC instead?
Above the silicon Schottky voltage class, or where high-temperature leakage matters more than cost. SiC Schottky combines the no-recovery behavior with a high voltage class at a premium; the comparison is its own decision.
Conclusion
Schottky versus FRD is a gate question: reverse voltage, leakage at temperature, junction capacitance and displacement current, forward drop, surge, and cost—with the waveform shaping each check. The Schottky wins when its gates close, even in hard switching; the FRD is the fallback when a gate fails. Run the checks with maximum values at the working condition and validate on a prototype.
Compare families side by side in the fast recovery rectifier diodes and Schottky rectifier diodes categories on the Good-Ark site, and contact Good-Ark with your reverse peak, leakage, and thermal data for a device recommendation.