Paralleling two rectifier diodes raises the current capability only if the branch currents actually split as intended, and the split is decided by the VF match at temperature, the layout symmetry, and the measurement. This guide covers the motivation, the sharing physics, the layout rules, the derating, and the verification that make a parallel pair behave like one larger diode.
Why Engineers Parallel Diodes
The common reasons are a current demand above a single package’s rating, thermal spreading across the board, and a margin or redundancy strategy that keeps the design alive if one die degrades. A parallel pair of 15 A diodes can carry more than a single 15 A part, but the capability is not the sum of the ratings—it is the sum of what each branch is measured to carry at the working temperature.
The alternative to discrete paralleling is a package that already contains two matched dies. The AMBRB30200CT, a 30 A, 200 V common-cathode Schottky in D2PAK from Good-Ark, is the single-package version of the parallel decision: the two dies share the same thermal base and are characterized together, which removes the branch-level variability that discrete paralleling must manage. The discrete approach remains useful when the current exceeds the dual-package class or when the PCB layout needs the heat spread across the board.
VF Matching and Current Sharing Physics
Two diodes in parallel see the same terminal voltage, and each branch conducts the current that matches its VF at that voltage. If the two parts have different VF curves, the lower-drop branch carries more current, and the imbalance is decided by the spread in VF at the working current, not by the average value.
The temperature coefficient of VF then determines whether the imbalance grows or heals. Where the series-resistance term dominates, VF rises with temperature and the hotter branch conducts less, which is self-balancing. Where the junction term dominates at low current density, VF falls with temperature and the hotter branch conducts more, which can drive thermal runaway in the branch. The sign depends on the device type and the current density, so the sharing analysis reads the datasheet’s VF curve at the actual branch current and temperature—the voltage physics article owns the full mechanism, and the parallel design’s rule is to select parts whose VF spread is small and whose coefficient at the working density is favorable.
The Datasheet Reading for a Parallel Pair.
The reading compares the VF at the working current and at two temperatures, because the sharing depends on both the spread and the slope. The maximum VF column gives the hot bound, the typical curve shows the working value, and the pair’s spread is read at the same current rather than at the datasheet’s test current alone. Where the supplier bins or matches parts, the request for matched or same-lot material is part of the procurement note, and the incoming inspection confirms the lot identity before the parts reach the board.
Layout Symmetry for Thermal Balance
The layout is the second half of the sharing story, because the branch resistance and the branch temperature both shape the split. Each branch should see the same trace length, the same copper area, and the same distance to the load and the source, so that no branch carries an extra voltage drop or runs hotter because of its position.
| Layout factor | Symmetric goal | Why it matters |
|---|---|---|
| Trace length to each anode/cathode | Equal | Equal branch resistance |
| Copper area under each part | Equal | Equal heat spreading |
| Distance to load and source | Equal | Equal parasitic drop |
| Nearby hot components | Avoided on one side only | Equal branch temperature |
| Common sense point | Single connection point | No shared-drop imbalance |
The thermal coupling between the two parts is part of the design: if one diode sits beside a hot inductor and the other does not, their temperatures diverge and the sharing shifts. The symmetric layout keeps both branches at the same temperature, which makes the sharing predictable and the measurement meaningful.
Branch Resistance on the Real Board.
At tens of amperes, a few milliohms of trace or terminal resistance shifts the split more than the datasheet spread alone. The review measures or estimates the branch resistance—the trace, the via, the terminal, and the solder—and the layout’s symmetry goal includes those terms, not just the visible copper. A common point where the two anodes and the two cathodes meet removes the shared-drop imbalance, and the current probes verify the result at the board’s real temperature.
Derating Rules for Parallel Operation
Parallel operation earns a derating review, because the rating of the pair is limited by the branch that carries the most current, not by the average. The review reads the maximum VF spread, the layout imbalance, and the surge behavior, and a common practice is to allow 10–20% below the sum of the single-device ratings as a starting margin, confirmed by the branch measurement at the working temperature.
The surge rating needs special care: IFSM is a non-repetitive survival margin quoted for a defined waveform and starting temperature, and paralleling does not automatically double it. Branch impedance differences and the timing of the surge make the split unpredictable, so the pair is checked against the actual inrush waveform with the derating already applied. The failure mode also matters: if one die fails open, the surviving branch sees the full load, and the design review should state what the system does at that moment.
Verification: Measuring Branch Currents
The measurement is the verdict. Each branch current is measured with a current probe or a low-value shunt at the nominal load, at the hot ambient, and across the duty cycle, and the sharing ratio is read at each condition. A review trigger is commonly set when one branch carries more than the target split by a defined percentage—the trigger value is set by the reliability target and the derating, not by a universal rule.
The thermal image completes the picture: the case temperatures of the two diodes should match when the currents match, and a temperature gap points to a layout imbalance that the current probe alone may miss. The record—branch currents, temperatures, ambient, and the layout revision—is the evidence that the parallel pair behaves as designed, and it becomes the baseline for the production and the field data.
The Production and Field Baseline.
The measured sharing ratio becomes a production check: the incoming lots are sampled for VF spread, and the periodic verification confirms the branches still split as the record says. In the field, the case temperatures of the two diodes are read at the scheduled inspections, and a drift in one branch points to a degraded joint, a changed lot, or a layout change before the part fails. The baseline is the reason the drift is visible, and the field reading closes the loop the bench opened.
Engineering note. The sharing analysis uses the maximum VF spread from the datasheet at the working branch current and temperature, with the thermal chain included; the 10–20% margin is a starting practice that is confirmed by the branch measurement on the build.
Frequently Asked Questions
Why is paralleling not simple addition?
Because the pair’s capability is set by the branch that carries the most current, and the split depends on the VF match, the layout, and the temperature.
How do the diodes share current?
They see the same terminal voltage and each conducts the current matching its VF at that voltage; a lower-drop branch carries more until the coefficients and the layout balance it.
What makes sharing stable?
A small VF spread, a favorable temperature coefficient at the working current density, and a symmetric layout that keeps both branches at the same temperature.
How much derating is needed?
A common starting margin is 10–20% below the sum of the ratings, confirmed by measuring each branch at the working temperature; surge and failure-mode reviews are part of the same check.
How do I verify the pair?
Measure each branch current at nominal load and hot ambient, compare the case temperatures, and record the results as the production baseline.
Conclusion
The parallel pair earns its rating through three gates: matched VF at the working condition, a symmetric layout that balances resistance and temperature, and a branch measurement that proves the split. Run the derating and the surge check with the same discipline, and the pair behaves like one larger diode with a readable margin.
Review the Schottky rectifier diodes category on the Good-Ark site, and send Good-Ark your current demand, ambient, and board layout for a paralleling or dual-package recommendation.