An electric vehicle is not one power stage; it is a map of them. A traction inverter, a DC-DC converter, an on-board charger, a set of auxiliary loads — each stage runs at a different voltage, frequency, and duty, and each has a device chemistry that fits it best. The recurring question “MOSFET or IGBT?” is really a question about where on that map a part sits. This article draws the EV power-stage map, explains why IGBTs still rule the traction inverter at 400-800 V while MOSFETs own the DC-DC and auxiliary stages, and follows the 800 V trend that is redrawing the boundaries — then closes with the verification a buyer should run before an EV power device is designed in.
The EV Power-Stage Map: Every Part Has a Home
The cleanest way to think about EV power electronics is as a map with a few distinct territories, each defined by its voltage, frequency, and current. The traction inverter is the largest and highest-voltage stage, driving the motor with hundreds of amps at switching frequencies in the low tens of kilohertz. The DC-DC converter steps the high-voltage traction battery down to the 12-14 V auxiliary rail, running at much higher frequencies. The on-board charger rectifies and conditions grid power. And a layer of auxiliary loads — pumps, fans, lighting, the low-voltage electronics — draws from the 12 V rail.
Each territory rewards a different device. The traction inverter wants a part that can block high voltage and carry high current with low conduction loss at a modest switching frequency, which is the IGBT’s home turf. The DC-DC and auxiliary stages want a part that can switch fast at lower voltage to keep a small, efficient converter, which is the MOSFET’s strength. The mistake is to ask “which device is better” as if the whole vehicle were one load; the map makes the answer per-stage. The EV traction device comparison and the MOSFET selection guide develop the two families; this article assigns them to stages.
Traction Inverter: Why IGBT Still Rules at 400-800 V
The traction inverter is where the biggest power flows, and it is where IGBTs remain the workhorse. The reason is a trade between voltage capability, conduction loss, and switching loss that lands in the IGBT’s favor at the traction operating point.
At 400-800 V and hundreds of amps, the inverter wants low conduction loss at high current and the ruggedness to handle the motor’s stall and surge events. An IGBT carries a saturation voltage that is low relative to its high blocking voltage, and it does so with a manufacturing cost and ruggedness profile that has made it the default for the traction role. The switching frequency is modest — a few to low tens of kilohertz — which is where the IGBT’s turn-off tail is an acceptable cost rather than a fatal one. The IGBT guide and the 1200 V IGBT motor-drive article develop the loss and packaging logic that keeps IGBTs in the traction role.
The comparison is not a permanent verdict. A MOSFET can block 400-800 V and switch faster, but at the traction current its on-resistance rises with voltage, pushing conduction loss up and the thermal budget with it. The trade is why the map, not the part name, decides the answer: at the traction inverter’s voltage and current, the IGBT’s conduction and ruggedness advantages outweigh its slower switching, and that is the logic this article is meant to make explicit. The IGBT vs MOSFET comparison gives the general crossover; the stage map here applies it to the vehicle.
DC-DC and Auxiliary: The MOSFET Grove
Move down the map to the DC-DC converter and the auxiliary loads, and the answer flips. These stages run at lower voltage and higher frequency, and the MOSFET’s fast switching and low switching loss make it the clear choice.
The DC-DC converter steps the 400-800 V traction bus down to the 12-14 V auxiliary rail, and it switches at high frequency to keep the magnetics small. At that frequency the MOSFET’s fast body-diode and low switching loss dominate; the converter’s efficiency depends on switching the device cleanly, and a MOSFET’s lower stored charge and faster edge win over an IGBT’s turn-off tail. The auxiliary stage — the low-voltage loads, lighting, fans, and the 12 V electronics — is the MOSFET’s simplest home, where low on-resistance at modest voltage and easy gate drive are all that matter. The synchronous rectifier MOSFET article and the MOSFET selection guide cover the loss logic of this stage.
The divide is clean enough to state as a rule of thumb: if the stage blocks high voltage, carries high current, and switches slowly, the IGBT has the edge; if the stage switches fast at lower voltage, the MOSFET wins. The EV is the map that makes the rule concrete, and the same stage logic applies to the charger and the auxiliary loads. Each stage selects its device by its own operating point rather than by a single vehicle-wide verdict.
The 800 V Trend and Its Device Consequences
The map is not static, and the 800 V trend is redrawing its boundaries. Automakers are moving traction systems from 400 V to 800 V to cut charging time and reduce cable and motor copper losses. The shift has real consequences for the device map.
At 800 V, the traction inverter’s blocking requirement rises, and the IGBT’s already strong high-voltage position becomes more entrenched — an 800 V inverter is even more clearly an IGBT (or SiC) application than a 400 V one, because the MOSFET’s on-resistance penalty grows with voltage. The DC-DC converter and charger also see the higher bus, which pushes their input-side devices toward higher voltage ratings and, where switching loss matters, toward silicon carbide. The 800 V EV charger article and the SiC selection guide cover the wide-bandgap side of the trend.
The design consequence is that the 800 V move does not simply replace one device with another; it reassigns each stage. The inverter stays IGBT or moves to SiC MOSFET for efficiency; the DC-DC and charger see higher voltage ratings; and the thermal and insulation margins tighten across all stages. A designer planning an 800 V platform should re-read the map for the new bus voltage rather than carry over 400 V device choices, and the SiC MOSFET article gives the specific checks for the higher-voltage stages.
The thermal and reliability picture completes the map, because the two device families meet in one place even though they live on opposite sides of it.
Despite living on opposite sides of the map, the IGBT in the inverter and the MOSFET in the DC-DC stage meet in one place: the thermal budget. Both devices dissipate heat, both are limited by a junction temperature, and both must survive the mission profile of the vehicle. The thermal discipline is the shared language that lets the two device families be compared honestly.
The thermal picture starts with loss, because loss is what heats the junction. In the inverter, the IGBT’s conduction and switching losses heat a module that is typically liquid- or heavy-heatsink-cooled; in the DC-DC stage, the MOSFET’s conduction and switching losses heat a smaller, air- or chassis-cooled package. In both, the junction temperature equals the ambient plus the loss times the thermal resistance, and the design must keep that junction below the derated limit across the worst case. The rectifier thermal design guide and the thermal management worked case develop the same junction-temperature method that both stages share, even though the packages differ.
The reliability side is where the map and the thermal budget come together. A traction IGBT that runs at high junction temperature for years faces power-cycling fatigue of the solder and bond wires; a DC-DC MOSFET that switches fast faces a different set of stress signatures. Both are captured by the automotive reliability framework — the thermal and power-cycling life curves, the AEC-Q101 qualification, and the mission profile. The thermal cycling article and the automotive rectifier mission-profile article connect the thermal and reliability math that both device families live under.
The practical point is that the device choice is never purely electrical. An IGBT may win the conduction comparison on paper, but if its package cannot shed the heat in the module layout, the stage fails thermally; a MOSFET may win the switching comparison, but if its on-resistance drives the junction past the limit at the operating current, it is not viable either. Reading the map means checking every stage’s thermal closure, not just its electrical ratings. The two families are different chemistries, but they are judged by the same junction-temperature and reliability standards, which is what makes the EV stage map a coherent design rather than a collection of unrelated parts.
Sourcing an EV Power Device: Verification Checklist
The map decides which device type a stage needs, but the buyer still has to verify the specific part. EV power devices carry high stakes — a traction or charger failure is a warranty event — so the sourcing decision deserves a verification checklist.
The checklist starts with the datasheet ratings against the stage’s worst case: the blocking voltage against the highest bus transient, the continuous and surge current against the motor or load duty, and the thermal ratings against the mission profile. Then comes the qualification and traceability: an automotive part should carry AEC-Q101 evidence, and the report should show sample sizes, conditions, failure modes, and lot traceability rather than a bare stamp. Finally, the second-source and supply questions: is the part available from more than one qualified source, and does the supplier’s quality and documentation support a long production life? The power semiconductor sourcing strategy and the automotive MOSFET design-in guide give the full verification and qualification flow.
The verification closes the map. The buyer who knows which stage a part serves, which device chemistry that stage rewards, and how the 800 V trend changes the assignment can then verify the specific part against the datasheet, the qualification, and the supply chain. The power MOSFET category and the IGBT category list the families the map lands on, and the verification checklist is what turns a stage assignment into a design that ships. The EV is a map of power stages; each part has a home; and the trade between MOSFET and IGBT is decided stage by stage, not by a single headline answer.