Bookkeeping Your MOSFET Losses, With Formulas!

MOSFET losses come in many forms, and principally the three that we care about are conduction, switching, and gate losses. These three losses combine to produce heat in the MOSFET, which could potentially lead to failure of a device. No one wants to manually desolder MOSFETs as they can have large bond pads that are difficult to pull off of the PCB. So before you start grabbing random MOSFETs for a test circuit board, make sure you select the right MOSFETs for your system.

In this article, we will look at the formulas for MOSFET switching losses and how they relate to the device properties and specifications. These devices have a relatively small number of specifications that matter for highly efficient switching, and we will review some of those in this article.

Main MOSFET Loss Mechanisms

The primary MOSFET loss mechanisms in switching circuits, particularly circuits driven with fast PWM signals, are as follows:

• Conduction losses due to resistance between drain and Source terminals

• Switching losses due to incomplete conduction when transitioning from off to on

• Gate losses incurred when the driver circuit charges up the gate

The total loss in the mosfet is the sum of these three mechanisms.

These three loss mechanisms are best examined in a SPICE simulation, where the MOSFET model is described as an equivalent circuit that includes capacitances, and ideally stray inductances of the package terminals. Now let's examine each of the three mechanisms and how they relate to device parameters.

Conduction Losses

The primary parameter determining the conduction loss is the power dissipated in the drain-source channel when the device is modulated on. In other words, the conduction loss is proportional to the on-state drain-source resistance and the channel current:

Power loss due to conduction in the channel

The conduction losses are primarily reduced by reducing the on-state channel resistance. Power MOSFETs tend to prefer lowest possible channel resistance when the MOSFET is run in DC. When run in AC, the channel resistance being as low as possible does not guarantee reliability. While it does reduce conduction loss, a lower channel resistance tends to correspond with higher gate capacitance. This will affect the gate losses as detailed below.

Gate Charging Losses

During switching, this loss mechanism tends to receive less attention than the standard switching loss mechanism detailed below. Gate charging losses are incurred during the process of charging up the gate. The terminal leading into the gate has some parasitic resistance which limits the turn on and turn off times to lower limits, meaning the MOSFET cannot switch on arbitrarily fast. To a first order approximation, the loss is proportional to the total gate charge:

Power loss during gate charging

This should illustrate why simply going to a smaller on-state resistance is not always the magic bullet when working with switching MOSFETs. Smaller on-state resistance requires a physically larger channel, which then incurs larger total gate capacitance. This then limits the switching rate, or equivalently requires a larger gate drive current in order to turn on in a reasonable time period such that switching loss is minimized. This essentially means that some of the reduced conduction loss is just shifted over to the gate drive circuit, which has its own losses due to its drive circuit channel resistances.

Switching Loss

Switching losses in MOSFETs are the power losses that occur as a MOSFET is switching between on and off states. Switching losses occur during the transition from off to on states, during which time the drain source resistance is changing. As current is brought into the MOSFET, it interacts with the changing resistance and momentarily incurs strong losses.

The power loss in the MOSFET channel during switching is given as follows:

Power loss during switching

The total loss outlined in these equations represents an approximation based on a linear model for changes in the MOSFET conductance during switching. The conductance change in the channel is dynamic, and thus the power loss is dynamic.

Finally, the terminal capacitances and gate drive current matter as well; higher gate drive will speed up the transition time because the terminal capacitances can charge up faster, thus leading to lower switching loss. This should illustrate why you cannot just use any gate driver to toggle every MOSFET, some gate drivers need to source a lot of current and thus will incur their own set of losses independent of the MOSFET. These gate drivers tend to be more reliable than the MOSFETs they are driving, so we might prefer to have losses occur in the gate driver rather than the MOSFETs.

Gate Drive Losses

The gate driver itself incurs some loss due to the driving current traveling through its own circuitry, which tends to be built from complementary BJTs in push-pull configuration. An example circuit is shown below

Example gate drive circuit from Infineon.

Technically, you can calculate the gate drive circuit losses using the upper and lower resistance values for sourcing and sinking gate drive current. Sometimes, these source and sink currents are different, so slightly different losses are incurred on the source edge versus the sink edge. Some gate drivers may provide data on this so that the losses can be accounted for in a switching power system.

Anytime you need to simulate power systems with switching MOSFETs, use the complete set of circuit simulation features in PSpice from Cadence. PSpice users can access a powerful SPICE simulator as well as specialty design capabilities like model creation, graphing and analysis tools, and much more.

Subscribe to our newsletter for the latest updates. If you’re looking to learn more about how Cadence has the solution for you, talk to our team of experts.