How to Choose the Right Ultra Fast Recovery Rectifier Diode?

The recovery number that decides your diode is not the one in the product summary. Take one real 600 V, 60 A ultrafast rectifier: its summary line quotes a 35 ns maximum reverse recovery time, yet the same datasheet measures 92 ns at 125 °C with 60 A forward current, 400 V reverse voltage and a 200 A/µs turn-off slope. Recovery is a function of the current and the di/dt you switch, so a selection made from the headline figure is a selection made at somebody else’s operating point.

Last updated: October 11, 2026

Key Takeaways

• Reverse recovery time (t_rr) is defined by its test conditions: forward current before turn-off and the rate of change of current, dI/dt. Eaton’s fast recovery diode application note (ELX1366) states the dependence directly.
• Recovery is not one number but three: t_rr, the reverse recovery charge Q_rr, and the peak reverse recovery current I_RM. The charge is what turns into loss; the peak current is what the switch beside it has to absorb.
• Soft recovery matters as much as fast recovery. A hard, abrupt turn-off raises dI/dt and radiates EMI; a soft characteristic trades a slightly longer tail for quieter switching.
• Fast recovery diodes exist for the voltage range Schottky diodes cannot cover — above roughly 100 V of blocking capability, the p-n fast recovery family takes over.
• Select in this order: reverse voltage against the worst-case node, thermal path against continuous current, then recovery behaviour against your switching frequency. Doing it the other way round is how parts get re-specified mid-project.

What Recovery Speed Actually Means

A diode does not stop conducting the instant the voltage reverses; it keeps passing current while the stored charge in the junction is removed, and that interval is the reverse recovery time. Everything the rest of the circuit experiences — loss, overshoot, noise — happens inside it.

The mechanism is worth one paragraph because it explains why the numbers move. In forward conduction the junction is flooded with charge carriers. To block, those carriers have to be swept out or recombined, and the removal takes a finite time during which reverse current flows. Eaton’s technical note describes the consequence precisely: t_rr is directly related to the forward current flowing before the reverse voltage is applied and to the rate of change of current at turn-off. Push more current through the diode, or switch it off faster, and the recovery interval — and the energy inside it — grows.

That is why three parameters travel together. The reverse recovery time says how long the event lasts. The recovery charge Q_rr is the area under the recovery current waveform and is the quantity that turns into heat. The peak reverse recovery current I_RM is what the companion switch or the loop inductance has to deal with. A diode can look fast on time and still be hard on the system if the peak current is high, because the peak is what rings.

Reading the t_rr Row Properly

Always read the test conditions next to the number, and then find the row that matches your circuit. Manufacturers legitimately quote more than one.

Where the figure appears	Test conditions	What it is useful for
Product summary	Small-signal, low-current bench conditions	Comparing families at a glance; never a design number
Dynamic characteristics table, small-signal row	e.g. IF = 1 A, VR = 30 V, dIF/dt = −50 A/µs	Practical for low-power converters running at moderate slope
Dynamic characteristics table, power row	e.g. Tj = 125 °C, IF = 60 A, VR = 400 V, dIF/dt = −200 A/µs	The row that matches a real high-power switching stage

In the STMicroelectronics STTH60RQ06 datasheet — a 600 V, 60 A ultrafast rectifier — the product summary quotes a 35 ns maximum t_rr, while the power-condition row at 125 °C, 60 A, 400 V and −200 A/µs reports 92 ns with a typical 660 nC Q_rr and a maximum 16 A I_RM. Both are correct. They are answers to different questions, and only one of them is about your converter.

The Selection Path, in Order

Work the four steps in sequence, because each one can eliminate a candidate before the next becomes interesting.

• Reverse voltage first. The rated repetitive peak reverse voltage has to cover the worst-case node voltage — DC link plus the overshoot your layout and stray inductance will actually produce — with derating margin on top.
• Thermal path second. Continuous current is limited by junction-to-case thermal resistance and the heatsink behind it, not by the current in the part number.
• Recovery behaviour third. Now compare t_rr, Q_rr and I_RM at your switching conditions, not at the summary conditions.
• Softness and packaging last. Check the recovery shape against the ringing budget of the loop, then confirm the package fits the mechanical and creepage design.

Two devices from the same vendor and the same voltage class show how the trade lands, and they are close enough that the choice is genuinely a decision rather than an obvious ranking.

Parameter (datasheet values)	STTH30RQ06	STTH60RQ06
Repetitive peak reverse voltage VRRM	600 V	600 V
Average forward current IF(AV)	30 A at TC = 103 °C, δ = 0.5	60 A at TC = 115 °C, δ = 0.5
Forward voltage VF (25 °C / 150 °C)	2.45 V typ at 15 A / 1.15 V typ, 1.45 V max at 15 A	2.45 V typ at 30 A / 1.15 V typ, 1.45 V max at 30 A
Reverse recovery trr	30 ns typ, 55 ns max (1 A, 30 V, −50 A/µs)	35 ns max (small signal); 65 ns max (1 A, 30 V, −50 A/µs)
IRM at 125 °C, rated current, 400 V, −200 A/µs	8 A typ, 11 A max	12 A typ, 16 A max
Junction-to-case thermal resistance Rth(j-c)	0.70 °C/W (0.90 °C/W for DO-247 LL)	0.38 °C/W
Surge current IFSM (10 ms sinusoidal)	200 A (TO-220AC, DO-247, DO-247 LL)	425 A (DO-247)
Leakage IR at VRRM, 25 °C / 150 °C	40 µA max / 800 µA max	80 µA max / 1600 µA max
Packages offered	TO-220AC, D²PAK, D²PAK HV, DO-247, DO-247 LL	DO-247, DO-247 LL

A Worked Loss Model from Published Coefficients

ST publishes the conduction-loss equation for these parts, which means the estimate does not have to be invented. For the STTH60RQ06 it is P = 1.05 × I_F(AV) + 0.013 × I_F RMS².

Assume a square-wave current with 30 A average and a 50 % duty cycle, which makes the RMS current 30 / √0.5 = 42.4 A.

• P = 1.05 × 30 + 0.013 × 42.4² = 31.5 + 23.4 = 54.9 W.
• Junction rise above case = 54.9 × 0.38 °C/W = 20.9 °C.
• At a 100 °C case the junction sits near 121 °C against a 175 °C limit — but the case temperature is set by the heatsink and the mounting interface, which is why the next number is measured on the bench rather than read from the datasheet.

Apply the same method to the 30 A part with its own coefficients (P = 1.05 × I_F(AV) + 0.026 × I_F RMS²) at 15 A average and a 50 % duty cycle: I_F RMS = 21.2 A, P = 15.75 + 11.7 = 27.4 W, and the rise above case is 19.2 °C. Both parts land near a 20 °C rise; the difference is that one of them is carrying twice the current while doing it. That is the whole point of a thermal path, and it is a model built from published coefficients and stated assumptions rather than a measurement.

Boundary Conditions That Cancel a Selection

Three datasheet rows routinely disqualify a part that looked correct on paper.

Leakage at temperature is the first. Both parts above are specified at tens of microamps at 25 °C and at hundreds to sixteen hundred microamps at 150 °C — a rise of roughly twenty times. In a high-impedance or high-temperature node that is a real current, not a rounding error. Surge is the second: the 425 A and 200 A non-repetitive figures are single 10 ms sinusoidal events, which is an inrush number rather than a repetitive rating, and treating it as continuous capability is a common source of early failure. Recovery at temperature is the third: t_rr and I_RM are quoted at 125 °C in the power rows because they grow with temperature, so a part validated on a cold bench has not been validated.

There is also a technology boundary worth stating plainly. Schottky diodes switch with essentially no minority-carrier recovery, but their blocking capability generally does not extend beyond about 100 V, which is why they belong on the low-voltage secondary side of a converter and fast recovery diodes take over on the high-voltage, high-frequency side. Eaton’s note puts the split in exactly those terms, and it maps directly onto where each device appears in a switch-mode supply: conventional rectifiers in the mains bridge, fast recovery diodes before the transformer, Schottky devices after it.

Where Ultrafast Rectifiers Fit

The application list follows from the frequency and the voltage, not from marketing categories. These parts appear in switch-mode power supplies, solar and PV inverter stages, DC-DC converters in IT and data-centre power, automotive lighting and power electronics, and UPS and energy-storage systems. They are also used for boost, bootstrap, bypass, freewheeling, snubbing and resonant paths, and in continuous-conduction-mode boost power-factor-correction stages, where the diode’s recovery current adds to the current the MOSFET has to turn on into and therefore worsens the switch’s turn-on loss.

The practical design consequences follow: keep the current loop small so the recovery event does not couple into the rest of the board, verify the softness of the recovery against your ringing budget, and confirm the package can carry the continuous current with the heatsink you have actually specified. Where a supplier can support that work, ask for the datasheet with its power-condition recovery rows, the thermal and derating curves, the package and outline drawing, and the qualification evidence that applies to your market — including the automotive-qualified equivalent if the design is going into a vehicle.

FAQ

Can an ultrafast recovery diode replace a Schottky diode?

Sometimes, but not as a like-for-like swap. Schottky diodes win on forward drop and have essentially no minority-carrier recovery; fast recovery diodes exist because a Schottky’s blocking capability generally stops around 100 V. Above that voltage the fast recovery device is the practical choice, and below it the forward-voltage penalty usually decides against it.

Why is the recovery time I measure longer than the datasheet figure?

Most often because you are measuring at a different operating point. t_rr depends on the forward current before turn-off and on the dI/dt of the switching edge, and it grows with junction temperature. Compare your conditions with the test conditions printed beside the number.

Does a lower t_rr always mean lower switching loss?

Not always. Recovery charge and peak recovery current matter as much as time: a short event with a high peak current can dissipate more and ring harder than a slightly longer, softer one. Evaluate t_rr, Q_rr and I_RM together.

How much reverse-voltage margin should I leave?

Enough to cover the worst-case node voltage including the overshoot your layout and loop inductance produce, not just the nominal DC link. The switching transient is usually the highest voltage the diode ever sees, which is precisely when it must block.

What should I request before approving a part?

The datasheet including its power-condition recovery rows and thermal derating curves, the junction-to-case thermal resistance for the intended mounting, the package and outline drawing, and the qualification or test evidence for your market.

Related Articles

• Fast recovery diodes and switching performance
• How to compare rectifier diode datasheets
• TO-220 fast recovery diode switching
• TO-247 fast recovery diode selection
• Talk to a Good-Ark engineer

Official Resources

• Good-Ark rectifier diode families
• Good-Ark technical documents and datasheets
• Good-Ark product application know-how

References

• Eaton, Understanding fast recovery diodes and their uses, Technical Note ELX1366 — source of the definition of t_rr and its dependence on forward current and dI/dt, the soft-versus-hard recovery and EMI discussion, the Schottky blocking-voltage limitation, and the use of each diode type in a switch-mode supply. eaton.com — ELX1366 application note (PDF)
• STMicroelectronics, STTH60RQ06 — 600 V, 60 A ultrafast high voltage rectifier, DS12515 Rev 3 — source of the 600 V/60 A ratings, 0.38 °C/W junction-to-case thermal resistance, 425 A surge rating, leakage at 25 °C and 150 °C, the 92 ns / 660 nC / 16 A recovery figures at 125 °C and the published conduction-loss equation. st.com — STTH60RQ06 datasheet (PDF)
• STMicroelectronics, STTH30RQ06 — 600 V, 30 A ultrafast high voltage rectifier, DS12164 Rev 5 — source of the 30 A variant’s ratings, 0.70 °C/W and 0.90 °C/W thermal resistances, 200 A surge rating, leakage figures and its conduction-loss equation. st.com — STTH30RQ06 datasheet (PDF)
• Vishay, Ultrafast Recovery — TO-247 — product-family overview used to confirm the commercial breadth of this device class. vishay.com — ultrafast recovery TO-247

Continue through this engineering topic

For the complete selection, design, verification, and troubleshooting path, explore the Rectifier Selection and Ratings.

• Practical guide: 1N4004G Rectifier Diode: Specifications, Polarity and Applications
• Practical guide: 1N4001G Rectifier Diode: Key Specs and Low-Voltage Circuit Checks
• Practical guide: 1N4002G Rectifier Diode: Specs, Polarity, Uses and Replacements
• Ask Good-Ark technical support about this application