Three voltage columns on a rectifier datasheet answer three different questions: VRRM is the repetitive peak reverse voltage the part blocks, VRMS is the RMS AC level it can handle, and VDC is its DC blocking capability. The margin rule that ties them together—20–30% over the worst real peak—is the habit that prevents most voltage-related failures.
VRRM, VRMS, VDC: Three Columns, Three Questions
VRRM (maximum repetitive peak reverse voltage)
is the headline blocking number: the highest reverse voltage the part can withstand repeatedly, cycle after cycle. It is the column to check against the worst reverse peak the circuit produces. VRMS (maximum RMS voltage) describes the AC side: the RMS level of a sine input whose peak the part can block. VDC (maximum DC blocking voltage) states the DC voltage the part can hold off continuously.
The three are related through the peak factor of a sine wave: the peak of an RMS value is about 1.414× the RMS, so a part that handles 250 V DC blocking can also be rated for a lower RMS input whose peak stays inside that envelope. The columns are not interchangeable—checking VRRM against a DC bus voltage, or VRMS against a DC level, reads the wrong question.
The conversion matters in both directions: reading a bridge datasheet by its RMS column alone misses the peak the diodes actually block, and reading a DC column against a sinusoidal input overstates the headroom.
The AMBR40250S shows the relationship in practice: 250 V of DC blocking, 175 V of RMS capability, and 250 V of repetitive peak reverse voltage are the same physical device viewed from three angles. The RMS column is lower because the peak of an RMS sine is about 1.414× the RMS value, and the part must block that peak, not the RMS level.
Where Peak Voltage Comes From in Real Circuits
The reverse stress a rectifier actually sees rarely equals the nominal rail. Three sources dominate:
- Rectified mains peak. A 230 V RMS input produces about 325 V peaks across each bridge diode; a 120 V RMS input produces about 170 V. The diode must block the peak, not the RMS value.
- Flyback and switching spikes. In a switching converter, the leakage inductance of the transformer adds a spike on top of the reflected voltage at every turn-off. The spike is real stress and belongs in the margin.
- Inductive-load turn-off. A freewheeling diode across an inductive load sees the supply plus the induced voltage when the switch opens; the blocking requirement follows the whole excursion.
Every source is a peak, and every peak is what the margin rule is designed to absorb.
The three sources also compound: a flyback spike lands on top of a rectified mains peak, and an inductive kick adds to the bus. The design habit is to sum the terms in the worst switching state rather than checking the nominal rail, because the nominal rail is the one value the circuit never actually sees at the switching instant.
The 20–30% Margin Rule and When to Break It
The engineering habit is to select a voltage class 20–30% above the worst repetitive peak the circuit can produce. The margin absorbs tolerance spread, temperature effects, aging, and the difference between the simulated peak and the measured one. It is a design rule, not a datasheet guarantee—the datasheet states the rating, and the designer states the margin.
The rule has exceptions that widen it. Noisy rails, inductive loads, automotive transients, and lightning-coupled events argue for more than 30%. The exception that shrinks it is rare and usually wrong: a cost-driven selection at the rating edge leaves no room for the very transients the rating exists to survive. When in doubt, take the next class up and pay the small forward-drop and leakage cost, because a voltage failure is a field return while a slightly higher class is a line-item difference.
The margin also covers the difference between simulation and measurement: a layout with extra trace inductance, a snubber that is not yet tuned, or a transformer with a higher leakage spec all raise the real peak above the first-pass number. Selecting the class before the layout is finalized, then re-measuring after, is the honest sequence.
Voltage Class Tables for Common Applications
| Application | Typical reverse stress | Suggested class (with margin) |
|---|---|---|
| 5–24 V output rectification | 20–60 V | 45–100 V |
| 12 V automotive rail | 30–60 V with transients | 100 V |
| 48 V output rail | 70–100 V | 100–150 V |
| 230 V mains input bridge | ~325 V peak | 600 V-class bridge |
| OBC / server intermediate bus | 200–400 V peaks | 200–250 V Schottky or FRD/SiC above |
The table is a starting point, not a substitute for measuring the actual peak at the node. The classes shown are standard catalog values; the margin rule turns the application into a class.
The 600 V-class bridge row deserves emphasis: it is not because the application is 600 V, but because a 230 V RMS input produces 325 V peaks and the margin rule puts the class at the next standard rung. The class is a function of the peak plus margin, which is why the table lists stress, not nominal voltage.
Checking Blocking Voltage With Your Scope
Four steps confirm the selection on the bench:
- Probe the node with a differential or isolated probe at the diode’s cathode-to-anode terminals, not to chassis ground.
- Capture the worst condition—maximum input, maximum load, and the transient event the application can produce.
- Read the peak on the scope, including the overshoot, not the average.
- Compare with margin against VRRM, and repeat at the hot condition, because the circuit’s peaks can shift with temperature.
The measured peak is the number that matters; if it is within 20–30% of the rating, the class is correct.
A note on safety: the rectifier node is live in a mains-fed supply, so the measurement uses an isolated or differential probe and the bench is treated as a live-work area. The measurement is only valid if it does not change the circuit—a grounded scope probe on the wrong node changes exactly what it is measuring.
Voltage ratings are also temperature-dependent in practice: leakage grows with junction temperature and can interact with the blocking stress in a hot enclosure, and some ratings are stated at a defined temperature. Read the rating column with its temperature condition, and keep the margin at the hot condition, not the 25 °C one.
The margin also protects against misreading the application: a “48 V” rail in a datasheet rarely means 48 V at the diode—tolerance, ripple, and transients add tens of volts. The class selection is based on the measured peak, and the nominal label is only the starting point.
Measurement note. The peak-to-RMS relationship (peak ≈ 1.414 × RMS for a sine wave) is a fixed electrical relationship, and the 20–30% margin is an engineering habit rather than a standard requirement. Confirm the actual worst-case peak on the scope at the hot condition before finalizing the class; the margin rule is explained here as the selection gate, and each application article links back to this one rather than restating it.
Frequently Asked Questions
What is the difference between VRRM, VRMS, and VDC?
VRRM is the repetitive peak reverse voltage the part blocks each cycle; VRMS is the RMS AC input it can handle; VDC is the continuous DC blocking capability. Each answers a different question, and mixing them reads the wrong one.
Why do I need 20–30% margin over the reverse peak?
To absorb tolerance, temperature effects, aging, and the difference between simulated and measured peaks. The datasheet states the rating; the margin is the designer’s job.
What reverse voltage does a 230 V mains bridge see?
Each diode blocks the peak of the sine input, about 325 V for 230 V RMS, plus whatever switching or surge transients the application adds. Select a 600 V-class bridge for mains input in practice.
Can I use a 100 V rectifier on a 100 V peak?
Only with zero margin, which leaves no room for transients or tolerance. Apply 20–30% margin and take the next class—a 150 V part for a 100 V peak is the honest choice.
How do I measure the real reverse peak?
Use a differential or isolated probe across the diode, capture the worst input and load condition, read the peak including overshoot, and repeat at the hot condition. Compare that number against VRRM with margin.
Conclusion
Voltage selection is three columns and one rule: read VRRM, VRMS, and VDC for what each measures, find the worst real peak on the scope, and apply 20–30% margin before choosing the class. The margin is cheap; the field return it prevents is not.
Compare blocking ratings across the standard bridge rectifiers and related categories on the Good-Ark site, and submit your measured reverse peak and transient conditions to Good-Ark for a class confirmation.