DEBUGMISSION 028 · Power switching · LossesIntermediate

The MOSFET Is Getting Hot

A PWM output overheats during a bench test. The test was stopped before further investigation. You have the original board, a set of recorded instrument captures, and three possible explanations. Decide what to measure before recommending a change.

Requirements

Bus · 24 VSpecified ON current · 6 A PWM · 20 kHz at 80%Ambient · 45 °C
  1. 01Investigate
  2. 02Revise & verify
Circuit overview

Original board

Original low-side PWM power stage The 24 volt supply feeds a load, then the drain of Q1. Its source returns to power ground. The MCU drives its gate. This is a functional diagram; load protection is not shown. +24 VPWM LOADDSGQ1N-MOSMCU OUTPUTProbe referencesVGS · gate to sourceVDS · drain to sourceFunctional circuit. Load protection is omitted.

All instrument captures describe this board, regardless of the revision you later choose. This is an authored investigation, not a live circuit simulator.

Circuit and case data
Circuit and case data
Engineering reference
Conduction loss
≤ 1.00 W
Switching loss
≤ 0.20 W
Estimated junction temperature
≤ 110 °C
Supplied component reference

Fictional, board-characterized profiles for this exercise, not selectable real-world part recommendations. Q1 is installed. Resistance is the hot operating-point estimate for each drive.

Hot resistance and effective switching charge
DeviceRDS(on) qualifying VGSR at 3.3 / 5 / 10 VEffective overlap charge / edge · authored modelθJA
Q1 · installed10 V120 / 75 / 55 mΩ 12 nC55 °C/W
Q210 V55 / 25 / 13 mΩ 45 nC110 °C/W
Q34.5 V55 / 30 / 20 mΩ 10 nC45 °C/W

All supplied candidates satisfy the 24 V bus, 6 A operating point and offered gate-drive absolute-maximum requirements. This exercise differentiates conduction, switching and thermal behavior. Candidate switching time uses an authored first-order charge/current model: switchingChargeNc is an effective charge per edge and the listed driver peak current is treated as a constant effective charging/discharging current. It is not a universal Qg/Ig datasheet equation. RDS(on) and θJA values are likewise supplied hot operating/board characterization for this exercise; real release requires device curves, gate resistance, Miller behavior, driver impedance and transient thermal validation.

The RDS(on) qualifying VGS is not VGS(th). Below it, resistance values are authored estimates, not guaranteed RDS(on). θJA applies only to the supplied board conditions.

MEASUREMENT BENCHReady to measure

Original board · prepared instrument captures

Select a measurement target, then measure to inspect the result.

Evidence notebook 0 saved

No evidence saved yet. Measure, inspect the result, then save useful observations.

Current task
Current task

Explain the evidence

What is causing the heating?

Save the load current, gate amplitude, switching overlap and ON-state drop. Use those observations to distinguish the hypotheses.

Your working hypothesis
0 evaluated decisions · Draft and evidence saved locally
Technical references

Loss model and its limits

Conduction: I² · Rhot · duty. Switching: ½ · V · I · (ton + toff) · frequency, with triangular voltage-current overlap. For candidate comparisons, ton + toff = 2 · Qeffective / Idrive. These effective overlap intervals are not total gate rise/fall times.

TJ = ambient + total loss · θJA is a steady-state estimate, not the measured case temperature. An out-of-range prediction rejects the candidate; it does not imply the device can safely reach that temperature.

The worksheet excludes driver dissipation, output-capacitance loss, diode/recovery loss and transient thermal behavior. Qualifying this simplified model is not qualifying real hardware.

Texas Instruments · MOSFET power-loss fundamentals