MOSFET Power Loss
A MOSFET does not operate without losses. Even when it is fully ON, it has a finite drain-source resistance. During switching, the device also experiences losses while transitioning between the ON and OFF states. Gate-drive power, body-diode conduction and other parasitic effects can add further losses. Understanding these losses is essential when designing power supplies, motor controllers, inverters, Class D amplifiers and other switching circuits.
What Causes MOSFET Power Loss?
The total power dissipated by a MOSFET can come from several different sources.
- Conduction loss
- Turn-on switching loss
- Turn-off switching loss
- Gate-drive loss
- Body-diode conduction loss
- Body-diode reverse-recovery loss
- Output-capacitance loss
- Parasitic and circuit-related losses
The relative importance of each loss depends on the MOSFET, load current, switching frequency, voltage, gate-drive circuit and topology.
Total MOSFET Power Loss
A simplified model of total MOSFET loss can be written as:
Ptotal ≈ Pconduction
+ Pswitching
+ Pgate
+ Pdiode
+ Precovery
+ Pother
This is a practical approximation rather than a complete semiconductor loss model. The manufacturer's datasheet and the actual circuit conditions should be used for accurate design.
Conduction Loss
When a MOSFET is fully ON, it does not behave like a perfect short circuit. It has a finite drain-source resistance called RDS(on).
The approximate conduction loss is:
Pcond = I² × RDS(on)
where:
- Pcond = conduction power loss
- I = MOSFET current
- RDS(on) = drain-source ON resistance
Example of Conduction Loss
Suppose an N-channel MOSFET has:
RDS(on) = 10 mΩ Current = 10 A
The conduction loss is:
Pcond = I² × RDS(on) Pcond = 10² × 0.010 Pcond = 1 W
The MOSFET therefore dissipates approximately 1 W while carrying 10 A, assuming the stated RDS(on) applies under the actual operating conditions.
Why Current Has a Large Effect
Conduction loss increases with the square of current.
Pcond ∝ I²
For example, if current doubles:
I → 2I Pcond → 4Pcond
This is why a MOSFET that operates comfortably at a few amperes can become very hot at substantially higher current even when the increase in current does not appear large.
RDS(on) and Gate Voltage
The specified RDS(on) is only valid under the test conditions shown in the datasheet.
For example, a MOSFET may specify:
RDS(on) = 8 mΩ at VGS = 10 V
This does not mean the MOSFET will have 8 mΩ resistance when driven with a 3.3 V signal.
Always check the RDS(on) specification at the actual gate-drive voltage.
RDS(on) Increases With Temperature
The ON resistance of most power MOSFETs increases as junction temperature rises.
Therefore the conduction loss can increase as the MOSFET heats up.
Temperature ↑
│
▼
RDS(on) ↑
│
▼
Conduction loss ↑
│
▼
Temperature ↑
This temperature dependence must be included in serious thermal designs.
Switching Loss
A MOSFET experiences switching loss while it changes between the OFF and ON states.
During the transition, both drain-source voltage and drain current can be significant at the same time.
OFF → Transition → ON High VDS Low VDS │ │ │\ │ │ \ │ │ \______________________│ Low ID High ID │ _________│ │ / │ / │___________/
The overlap between voltage and current produces instantaneous power dissipation.
Approximate Switching Loss
A simplified switching-loss equation is:
Psw ≈ 1/2 × VDS × ID × (tr + tf) × f
where:
- VDS = drain-source voltage
- ID = drain current
- tr = turn-on transition time
- tf = turn-off transition time
- f = switching frequency
This equation is useful for obtaining an initial estimate. Real switching losses can differ significantly because of capacitances, gate charge, Miller effects, diode recovery, parasitic inductance and the actual switching waveform.
Effect of Switching Frequency
Switching loss increases approximately in proportion to switching frequency when the other conditions remain similar.
Psw ∝ f
For example, increasing a switching frequency from 50 kHz to 100 kHz can approximately double the switching component of the loss if the switching energy per cycle remains approximately constant.
This is one of the major trade-offs in switch-mode power supplies.
Effect of Drain Voltage
Switching loss also increases with the voltage across the MOSFET during the transition.
Psw ∝ VDS
High-voltage switching circuits therefore require particularly careful attention to switching speed, device capacitances, gate drive and thermal design.
Effect of Drain Current
Switching loss also increases with drain current.
Psw ∝ ID
High-current switching applications can therefore experience substantial switching losses even when the MOSFET has a very low RDS(on).
Turn-On Loss
During turn-on, the MOSFET begins to conduct while drain-source voltage is still present.
The gate driver must charge the gate and move the MOSFET through the transition.
A slow turn-on increases the time during which significant voltage and current overlap.
This increases turn-on loss.
Turn-Off Loss
During turn-off, the MOSFET current falls while the drain-source voltage rises.
Again, the overlap between voltage and current produces power dissipation.
A weak gate driver can cause slow turn-off and increase this loss.
Gate-Drive Loss
Although the MOSFET gate draws almost no steady-state DC current, energy is required to charge and discharge the gate repeatedly.
A simplified estimate of gate-drive power is:
Pgate ≈ Qg × Vdrive × f
where:
- Qg = total gate charge
- Vdrive = gate-drive voltage
- f = switching frequency
Example of Gate-Drive Loss
Suppose:
Qg = 100 nC Vdrive = 10 V f = 100 kHz
Then the approximate gate-drive energy consumed per second is:
Pgate ≈ 100 nC × 10 V × 100,000 Pgate ≈ 0.1 W
This represents approximately 0.1 W of gate-drive energy per MOSFET under the simplified assumptions.
The actual driver IC also has its own operating losses.
Gate Charge vs RDS(on)
MOSFET selection often involves a trade-off between low RDS(on) and low gate charge.
| Characteristic | Lower Value Generally Provides |
|---|---|
| RDS(on) | Lower conduction loss |
| Gate charge | Lower gate-drive and switching requirements |
| Output capacitance | Lower capacitive switching loss |
| Reverse-recovery charge | Lower diode-related switching loss |
The MOSFET with the lowest RDS(on) is therefore not automatically the best choice for a high-frequency converter.
Body-Diode Loss
Power MOSFETs contain an intrinsic body diode.
When current flows through this diode, it produces conduction loss.
A simplified estimate is:
Pdiode ≈ VF × I × D
where:
- VF = diode forward voltage
- I = diode current
- D = fraction of time the diode conducts
The exact loss depends on the diode characteristics and current waveform.
Body-Diode Reverse Recovery
When a MOSFET body diode has been conducting and is suddenly reverse biased, stored charge can produce a reverse-recovery current.
This current can increase:
- Switching loss
- Voltage overshoot
- EMI
- Stress on other MOSFETs
Reverse-recovery behavior is particularly important in bridge and hard-switching converter topologies.
Reverse-Recovery Loss
A simplified estimate of reverse-recovery energy can be expressed as:
Erec ≈ V × Qrr
where:
- Erec = approximate recovery energy
- V = voltage associated with the switching event
- Qrr = reverse-recovery charge
The actual energy depends on the circuit and switching waveform, so manufacturer test conditions should be used for accurate calculations.
Output Capacitance Loss
A power MOSFET has drain-source and other parasitic capacitances.
The most commonly specified capacitances include:
- Ciss — input capacitance
- Coss — output capacitance
- Crss — reverse-transfer capacitance
These capacitances must be charged and discharged during switching.
At high switching frequencies, this can become an important source of loss.
Why MOSFET Capacitances Matter
A simplified capacitive energy relationship is:
E ≈ 1/2 × C × V²
This shows that capacitive switching energy increases strongly with voltage.
The actual MOSFET capacitances are nonlinear, so the datasheet's charge-energy curves are often more useful than treating the capacitance as a fixed value.
Switching Loss and Gate Resistance
The gate resistor influences switching speed.
A large gate resistor:
- Reduces gate current
- Slows switching
- May reduce ringing
- May reduce EMI
- Can increase switching loss
A small gate resistor:
- Increases gate current
- Speeds switching
- May reduce switching loss
- Can increase ringing
- Can increase EMI
The correct value is therefore a compromise.
Switching Loss and Gate Driver Strength
A stronger gate driver can charge and discharge the MOSFET gate faster.
Faster transitions generally reduce the time spent in the high-voltage, high-current switching region.
However, extremely fast switching can produce excessive ringing and electromagnetic interference.
The objective is not simply maximum switching speed. The objective is an efficient and controlled switching transition.
Dead-Time Loss
In a half-bridge, dead time is inserted to prevent simultaneous conduction of the high-side and low-side MOSFETs.
During dead time, current may flow through a body diode rather than through the MOSFET channel.
Excessive dead time therefore increases diode conduction loss.
Too little dead time
↓
Shoot-through risk
Too much dead time
↓
More diode conduction
Correct dead time
↓
Efficient switching
Thermal Power Dissipation
All the losses eventually appear as heat.
The MOSFET junction temperature depends on the power dissipated and the thermal resistance from the junction to the surrounding environment.
A simplified relationship is:
TJ = TA + P × θJA
where:
- TJ = junction temperature
- TA = ambient temperature
- P = power dissipation
- θJA = junction-to-ambient thermal resistance
Example Thermal Calculation
Suppose a MOSFET dissipates:
P = 3 Wand the effective thermal resistance is:
θJA = 30 °C/WAt an ambient temperature of 40 °C:
TJ = TA + P × θJA TJ = 40 + (3 × 30) TJ = 130 °CThe MOSFET's maximum junction temperature must be checked against this calculated value, with appropriate design margin.
Thermal Resistance
Common thermal-resistance specifications include:
- θJA — junction to ambient
- θJC — junction to case
- θCS — case to heatsink
The actual thermal path depends on the package, PCB, heatsink, mounting method and airflow.
MOSFET Power Loss in a Low-Side Switch
Consider a simple low-side MOSFET switching a resistive load.
+V │ LOAD │ D N-MOSFET S │ GND
For a low-frequency switching application, conduction loss may dominate.
The designer can estimate:
Pcond ≈ I² × RDS(on)
If switching is frequent, switching loss must also be included.
Example: 10 A MOSFET Switch
Assume:
Load current = 10 A RDS(on) = 8 mΩ Switching frequency = 20 kHz
The conduction loss is approximately:
Pcond = 10² × 0.008 Pcond = 0.8 W
This is only the conduction component. Switching, gate-drive and other losses must also be considered.
Example: High-Frequency Switching
Suppose a MOSFET operates at:
VDS = 100 V ID = 10 A tr = 50 ns tf = 50 ns f = 100 kHz
Using the simplified switching-loss equation:
Psw ≈ 1/2 × VDS × ID × (tr + tf) × f
we obtain an initial estimate of the switching loss. The result should then be compared with the manufacturer's switching-energy data and the actual circuit waveform because the simplified equation does not include all parasitic effects.
Why Datasheet Switching Energy Is Better
MOSFET manufacturers may provide switching-energy specifications such as turn-on and turn-off energy under defined test conditions.
If reliable datasheet values are available and the test conditions closely match the application, they can provide a better estimate than assuming ideal linear voltage and current transitions.
Always check:
- Drain voltage
- Drain current
- Gate-drive voltage
- Gate resistance
- Temperature
- Switching frequency
- Test-circuit topology
MOSFET Losses in a Buck Converter
A buck converter can have several MOSFET-related losses.
- High-side MOSFET conduction loss
- High-side switching loss
- Low-side MOSFET conduction loss
- Low-side switching loss
- Body-diode conduction
- Reverse-recovery loss
- Gate-drive loss
The duty cycle determines how long each MOSFET conducts, so conduction loss must be evaluated over the actual switching waveform rather than assuming continuous full-load current through every device.
MOSFET Losses in a Half-Bridge
A half-bridge contains high-side and low-side switching devices.
+V │ High MOSFET │ ├──── SW │ Low MOSFET │ GND
Losses can arise from:
- High-side conduction
- Low-side conduction
- High-side switching
- Low-side switching
- Body-diode conduction
- Reverse recovery
- Gate drive
- Dead time
Reducing MOSFET Power Loss
Several techniques can reduce MOSFET losses.
- Choose a MOSFET with sufficiently low RDS(on).
- Use an appropriate gate-drive voltage.
- Use a suitable gate driver.
- Minimize unnecessary switching time.
- Choose an appropriate switching frequency.
- Minimize gate-loop inductance.
- Optimize dead time.
- Reduce body-diode conduction when possible.
- Provide adequate cooling.
- Use suitable PCB copper area.
- Control voltage overshoot and ringing.
Lower RDS(on) Is Not Always Better
It may appear that the best MOSFET is always the one with the lowest RDS(on), but this is not necessarily true.
Achieving very low RDS(on) can require a larger die, which can increase gate charge and capacitance.
At high switching frequencies, the additional gate charge and switching loss may outweigh the conduction-loss benefit.
The MOSFET should therefore be selected according to the complete loss budget.
Lower Gate Charge Is Not Always Better
Similarly, a MOSFET with very low gate charge may have higher RDS(on).
For low-frequency, high-current switching, low RDS(on) may be more important.
For high-frequency switching, low gate charge and favorable switching characteristics may become more important.
The correct choice depends on the application.
Importance of Switching Frequency
| Switching Frequency | Typical Design Concern |
|---|---|
| Low | Conduction loss often dominates |
| Moderate | Conduction and switching losses both matter |
| High | Switching, gate and capacitive losses become increasingly important |
These are general trends rather than strict frequency boundaries.
Power Loss and MOSFET Temperature
As MOSFET power dissipation increases, junction temperature increases.
Higher temperature can increase RDS(on), which can increase conduction loss further.
This is why thermal design should be performed using realistic maximum operating conditions.
Higher current ↓ Higher conduction loss ↓ Higher temperature ↓ Higher RDS(on) ↓ Still more conduction loss
Thermal Runaway Considerations
Power MOSFETs generally have a positive temperature coefficient for RDS(on) in their normal fully enhanced region.
As the MOSFET heats up, its resistance increases, which tends to reduce current sharing in parallel devices rather than producing the same type of thermal runaway behavior associated with BJTs.
This does not eliminate thermal failure. A MOSFET can still exceed its maximum junction temperature if total power dissipation is too high.
Parallel MOSFETs and Power Loss
Connecting MOSFETs in parallel can reduce the effective conduction resistance and distribute the load current.
For identical MOSFETs under ideal conditions:
Rtotal ≈ RDS(on) / N
where N is the number of parallel devices.
Real current sharing depends on:
- PCB layout
- Device temperature
- Gate-drive matching
- Source resistance
- Package characteristics
- Device tolerances
MOSFET Efficiency
The efficiency associated with a MOSFET stage can be estimated by comparing the useful output power with the total input power.
Efficiency = Pout / Pin × 100%
Reducing MOSFET losses improves the efficiency of the overall converter or switching system.
Practical MOSFET Loss Calculation Procedure
- Determine the maximum operating current.
- Determine the MOSFET's RDS(on) at the actual gate voltage.
- Account for the increase in RDS(on) with temperature.
- Calculate conduction loss.
- Determine switching frequency.
- Estimate turn-on and turn-off switching loss.
- Calculate or estimate gate-drive loss.
- Consider body-diode conduction.
- Consider reverse-recovery effects.
- Consider output-capacitance and other switching losses.
- Add the significant loss components.
- Calculate junction temperature.
- Verify that the device remains within its ratings.
MOSFET Power Loss Checklist
- What is the maximum drain current?
- What is the actual VGS?
- What is RDS(on) at that VGS?
- What is RDS(on) at the expected temperature?
- What is the switching frequency?
- What are the turn-on and turn-off times?
- What is the total gate charge?
- What is the body-diode behavior?
- Is reverse recovery significant?
- What is the MOSFET junction temperature?
- Is a heatsink required?
- Is the gate driver strong enough?
- Is the PCB layout suitable for the switching speed?
Key Points
- MOSFET power loss is not limited to RDS(on).
- Conduction loss is approximately proportional to I² × RDS(on).
- Switching loss increases with switching frequency.
- Switching loss depends on voltage, current and transition time.
- Gate charge produces additional gate-drive power consumption.
- Body-diode conduction can contribute significant losses in bridge circuits.
- Reverse recovery can increase switching loss and voltage stress.
- MOSFET capacitances can produce significant high-frequency losses.
- RDS(on) normally increases as the MOSFET temperature rises.
- The lowest RDS(on) MOSFET is not necessarily the most efficient at high frequency.
- A strong but controlled gate driver can reduce switching loss.
- Excessive switching speed can increase ringing and EMI.
- Thermal design must account for the total MOSFET loss.
- Datasheet switching-energy information should be used whenever applicable.