Nexperia – How to use a board-mounted MOSFET to switch a current as high as 200A

Today’s MOSFETs have such low on-resistance that they stay cool even when handling very high currents. In many cases a heat sink is not required, since the copper of the PCB will dissipate the relatively small amount of heat generated by the MOSFET. This ability to surface-mount power devices has made new packages such as the LFPAK and Power-SO8 very popular.

Clearly, however, there is a limit to the amount of power that a surface- mount MOSFET can handle in practice. This Design Note explores this limit by describing a test system based on a good example of the new generation of surface-mount MOSFETs: Nexperia’s PSMN0R7-25YLD. Could this tiny surface-mount device, which has a board footprint of 4.5mm x 5.1mm, handle a continuous current of 200A?

This part is listed by Nexperia as a 25V N-channel logic-level MOSFET. It is housed in an LFPAK56 package, also described as a Power-SO8, and its sturdy design features an internal copper clip which reduces thermal resistance and increases the device’s tolerance of high transient currents, as shown in Figure 1.


Fig. 1: Internal construction of the Nexperia LFPAK package

The datasheet shows that this tiny device can dissipate 158W; the low on-resistance, a maximum 0.72mΩ at a gate-source voltage of 10V, suggests it should be capable of handling a current of several hundred Amps.

The value of 158W only holds true, however, if the mounting base – that is, the solder tab – is at 25°C. What does that imply? Dissipating 158W without a large rise in temperature calls for a huge heat-sink, probably assisted by water cooling: this is unsuitable for production designs.

The datasheet also shows that as the temperature at the mounting base rises, the power dissipation must be reduced. For example, at 75°C the maximum power dissipation must be derated to 95W.

Temperature also affects on-resistance: at the PSMN0R7-25YLD’s maximum junction temperature of 150°C, maximum on-resistance rises to 1.15mΩ. With these values, it is possible to calculate the current that will generate 95W and therefore the maximum current Power = I2R. This can be expressed as:

I2 = P/R
I2 = 95/0.00115
I = 287A

While this is an absolute maximum figure, it suggests that the target of 200A for a real-world design is within reach. The question then arises: what kind of PCB can carry 200A? Clearly not a conventional one. The thickest PCB copper available is 0.5mm (14oz). An online PCB trace- width calculator shows that even with 0.5mm copper, a trace 20mm wide is needed to keep the temperature rise of the PCB to as little as 20°C.

To demonstrate the ability to handle 200A, a test board was made which used solid copper bus bars 25mm wide x 3mm thick. Two pieces, one for the source and one for the drain, were glued to a third to form a rigid structure. The glue used was self-shimming, and thus provided an electrical insulating barrier, as shown in Figure 2.

To achieve the lowest possible thermal resistance, the source and drain of the PSMN0R7-25YLD were soldered directly to the copper. The thick copper spreads the heat over a large area, enabling the MOSFET to be cooled by natural convection.


Fig. 2: Copper base boards for testing the PSMN0R7-25YLD’s ability to handle 200A

MOSFETs are designed to operate as switches, dissipating very little power when fully on or fully off. In the intermediate state, however, current and voltage rise simultaneously, and it is therefore essential to switch as fast as possible if thermal dissipation is to be kept at a low level.

In the test system, a TC4422CPA from Microchip was used to drive the MOSFET’s gate. The TC4422CPA driver can provide up to 9A to charge the gate’s capacitance and thus achieve fast switching times.

Nichrome wire was used as a resistive load. The loop inductance was kept low by routing the flow and return cables close together to reduce the loop area.

Two DC-DC converters provided 100A each to the Nichrome wire resistors, which began to glow red after a few seconds. The clamp-on current meter showed that the current was slightly more than 200A. A thermocouple attached to the copper close to the MOSFET showed the temperature rising slowly, before settling at around 60°C.

In practice, of course, surface-mount MOSFETs in production designs are soldered to far thinner copper layers than were used in this test board. But the test demonstrates that the high current rating of Nexperia’s LFPAK package is valuable because it enables devices to survive short-term overload conditions such as the start-up phase of a motor or a locked rotor fault.