The stored charge in a diode junction is a time bomb that releases every time the diode turns off. In an SMPS switching at high frequency, that release happens thousands of times per second, and each release sends a current overshoot back into the inductor — a spike that heats, stresses neighbors, and radiates EMI if it is not designed out. Reverse recovery is the physics behind that overshoot, and the waveform it leaves on a scope tells you which diode family was chosen. This article explains what reverse recovery is, contrasts soft and abrupt recovery families, shows the circuit effects of the overshoot, lays out the design fixes, and closes with what to look for when watching recovery on a scope.
The Stored-Charge Time Bomb: What Reverse Recovery Is
The name “reverse recovery” describes the delayed turn-off of a diode, and the mechanism is stored charge. While a diode conducts forward, the junction is populated with charge carriers; when the circuit reverses the voltage, the diode cannot simply snap off — it must first sweep that stored charge out of the junction. The sweep takes time, draws a reverse current, and defines the diode’s recovery behavior.
The measured quantity is the reverse recovery time, trr, the interval from the zero crossing to the point where the diode has recovered its reverse blocking. More important than the nameplate time is the recovery charge — the stored charge that must be swept — because it is the charge that becomes current when the diode turns off. A circuit that switches a diode off stores energy in the junction and releases it as a reverse-current pulse; the bigger the stored charge, the bigger the pulse. The fast recovery diode guide and the diode forward voltage article develop the physics and the datasheet terms.
The “time bomb” framing is literal in one respect: the release is unavoidable. Every turn-off of a conventional rectifier produces a reverse-recovery event, and the design question is not whether the charge releases but how fast, how hard, and where the current goes. The waveform of that release is the subject of the next section.
Soft vs Abrupt Recovery: Two Families, Two Waveforms
The recovery event does not look the same for every diode, and the waveform it leaves is the fingerprint of the family. The two classic shapes are soft and abrupt recovery, and they differ in how the reverse current decays.
An abrupt-recovery diode turns off hard: the reverse current rises quickly to its peak and then snaps to zero, with a sharp derivative at the end. The sharp snap is electrically violent — it drives the voltage overshoot in the surrounding circuit and is the shape most likely to produce oscillatory ringing. A soft-recovery diode turns off with a tail: the reverse current decays gradually, spreading the stored-charge release over time and reducing the sharp derivative. The soft tail trades a slightly longer recovery interval for a gentler circuit signature. The fast recovery diode guide and the standard vs fast recovery article contrast the two families and their waveforms.
The family choice is a trade between speed and gentleness. Where the circuit can tolerate the sharp release and the frequency demands a short recovery, the abrupt fast part wins; where the ringing and EMI of the sharp derivative hurt the neighbors, the soft part is worth its longer tail. Watching the scope is the way to see which family is in the board, and the final section returns to that read.
Circuit Effects: Overshoot, EMI, and Stressed Neighbors
The recovery event does not stay inside the diode; it pours into the circuit, and the effected neighbors are the reason the physics matters to the design. Three effects dominate.
The first is the voltage overshoot. When the reverse-recovery current snaps off into the circuit’s inductance, it drives a voltage overshoot across the switch and the rectifier — an overshoot that can stress the very parts the design thought it had protected. The second is EMI: the recovery transient is a broadband event, radiating noise that couples into nearby wiring and electronics, and at high switching frequency the accumulated EMI is a compliance and performance issue. The third is the stressed neighbors: the switch, the snubber, and the magnetic components all absorb parts of the recovery energy, and a hard recovery with a big overshoot pushes them closer to their limits. The ripple and noise diagnosis guide and the EMI filter article trace the recovery transient’s EMI into the system, and the power supply repair guide shows the neighbor stress in real boards.
The design consequence is that recovery cannot be read from the datasheet alone; it has to be read from the circuit. The same diode in a hard-switching, high-loop-inductance circuit produces a larger overshoot than in a gentle, low-inductance one, because the recovery transient’s impact scales with the circuit it releases into.
The overshoot is worth quantifying because the numbers make the risk concrete. In a flyback whose output rectifier turns off against a loop inductance L, the recovery current snap drives a voltage overshoot roughly proportional to the rate of current change times the inductance. A soft-recovery part that spreads the same total charge over three times the interval cuts that dv/dt by roughly the same factor, and the overshoot with it. Doubling the switching frequency doubles the number of events per second, so the same stored charge releases twice as often into the same neighbors — which is why a converter moved from 100 kHz to 400 kHz without re-examining the rectifier’s recovery behavior often reveals the problem the previous frequency had hidden. The flyback snubber guide develops the same numbers from the leakage-energy side, and the gate drive design article shows the dv/dt role of the switching driver.
The neighbor stress also explains a field pattern that misleads repair benches: an SMPS that intermittently fails at the switch or the capacitor is often really a recovery problem at the rectifier. The parts that fail are the stressed neighbors — the switch that bore the overshoot, the capacitor that absorbed the ringing — while the rectifier that released the charge reads nominal on the bench. The diagnosis habit from the power supply repair guide is to check the recovery waveform before replacing the visibly failed neighbor, because the visible failure is the symptom and the stored-charge release is the cause.
Design Fixes: Snubber, Driver, and Part Choice
The design fixes for reverse recovery attack the release from three angles: absorb it, slow it, or choose a part that releases less. Each fix has its place.
The snubber is the absorber: an RC or RCD element placed across the rectifier or the switch absorbs the peak of the recovery transient, damping the overshoot and the ringing. It is the classic fix, and it is the subject of the flyback snubber design. The driver is the slower: a gate driver with controlled switching edges slows the current changes the circuit asks for, reducing the dv/dt that the recovery transient magnifies. The part choice is the release reducer: a fast-recovery or soft-recovery diode releases less charge and less harshly at the source. The flyback snubber guide and the gate drive design article cover the snubber and driver fixes, and the fast recovery selection covers the part choice.
The order matters: prefer the part that releases less, add the snubber where the part cannot, and tune the driver last because it affects the whole switching edge. A design that throws a snubber at a problem the part choice could have solved pays twice; one that chooses the soft part and then verifies the circuit has spent the budget where it works.
Watching Recovery on a Scope: What to Look For
The verification of a recovery design is a scope, and the waveform tells the story if you know what to look for. The observation is the last section because it is the read that closes every design decision in the article.
The setup is a current probe on the rectifier and a voltage channel on the switch. Watch the turn-off: the current dips below zero as the recovery event sweeps the stored charge, peaks at the reverse current maximum, and then returns. The three things to read are the overshoot, the tail, and the ringing. The overshoot is the depth of the reverse-current excursion relative to the operating current; the tail is how fast the current returns to the zero line, with the soft part showing a visible tail and the abrupt part snapping; the ringing is any oscillation that follows the snap, its amplitude and its settling time. The fast recovery diode guide and the ripple diagnosis guide describe the scope read in the context of real SMPS waveforms.
The verdict comes from the three reads together. A small overshoot, a controlled tail, and no ringing mean the recovery is designed out; a deep overshoot with ringing means the snubber or the part choice needs work. The scope is the honest judge, and the same waveform that confuses a first read is the fingerprint that confirms the family and the fix.
Reverse recovery is the current overshoot you have to design out, and the design is the whole chain of this article. The stored charge releases at every turn-off, the waveform tells the family, the overshoot and EMI stress the neighbors, and the fixes — part choice, snubber, driver — attack the release at its source. Watch the scope at the operating frequency, read the overshoot and the tail, and the SMPS output rectifier stops being a source of hidden spikes. The power MOSFET category and the Schottky category supply the parts whose recovery behavior this design method selects.