Troubleshooting Dead MOSFETs in Power Supplies

You crack open a power supply that’s gone silent, and there it is—a MOSFET that looks like it’s been through a small war. Maybe it’s cracked, maybe it’s blackened, or maybe it just sits there looking innocent while the supply refuses to do anything useful. I’ve been in this game for over a decade, and let me tell you: dead MOSFETs in power supplies are rarely the whole story. They’re almost always a symptom of something deeper.

The real trick isn’t just swapping the part. It’s understanding why it died in the first place. Otherwise, you’ll be back in an hour watching the replacement go up in smoke. Troubleshooting dead MOSFETs requires patience, a solid methodology, and the willingness to admit that sometimes—honestly—you need to take a step back and look at the whole circuit.

Seriously, half the battle is knowing what questions to ask before you reach for the soldering iron. Let’s dig in.

Why MOSFETs Die: The Anatomy of a Power Supply Failure

MOSFET failure in power supplies isn’t random. It follows predictable patterns, and once you recognize those patterns, the troubleshooting process becomes almost mechanical. The most common culprits fall into a few neat categories: overvoltage, overcurrent, excessive heat, and gate drive issues. Each leaves its own signature.

Look—most engineers I’ve worked with focus on the MOSFET itself and forget about the ecosystem around it. The gate driver, the snubber network, the transformer, even the layout of the PCB. A dead MOSFET in a switched-mode power supply is like a check engine light. It tells you something went wrong, but it doesn’t tell you where.

The Big Four Causes You Need to Know

Let’s break this down into the absolute fundamentals. If you can identify which of these four categories caused the failure, you’re already ninety percent of the way to a fix.

- Overvoltage stress: The drain-to-source voltage exceeded the rating, often from reflected voltage spikes or poor snubber design. - Overcurrent stress: The current through the channel exceeded safe limits, typically from shorted loads or transformer saturation. - Thermal runaway: The junction temperature climbed beyond what the package could handle, often due to insufficient heatsinking or high switching losses. - Gate drive failure: The gate voltage was too low, too high, or oscillated in a way that kept the device in the linear region far too long.

Each of these has a distinct set of symptoms. Overvoltage tends to leave visible burn marks near the drain pin. Overcurrent often blows the source pin or the bond wires. Thermal failure usually shows up as cracks in the package or discoloration on the backside. Gate drive issues? They’re sneaky. The MOSFET might look fine but test completely dead on the meter.

Why Thermal Management Matters More Than You Think

I can’t tell you how many times I’ve seen a dead MOSFET that was perfectly fine electrically but died because the thermal grease was applied poorly. It sounds ridiculous until you measure the case temperature and realize it’s running at 120 degrees Celsius under normal load.

Heat is the silent killer. Every degree above the rated junction temperature reduces the device’s lifetime—sometimes exponentially. And in power supplies, the thermal loop is often overlooked. The heatsink might be undersized, or maybe there’s a dust blockage in the airflow. Sometimes the mounting torque on the screw is wrong, creating an air gap that acts like a thermal insulator.

Here’s the thing: if you replace a failed MOSFET without checking the thermal interface, you’re rolling the dice. And the house always wins.

Step-by-Step: Practical Troubleshooting of Dead MOSFETs

Alright, you’ve got a dead MOSFET in your power supply and you’re ready to figure out the root cause. Let’s walk through the process the way I do it in the lab, step by step. No shortcuts, no guesswork—just methodical investigation.

Visual Inspection: What the Board Tells You

Before you touch a single probe, look at the board. Really look. Pop it under a magnifying lamp and scan every component near the failed MOSFET. You’re looking for cracked solder joints, bulging capacitors, burnt resistors, or any sign of arcing. I’ve found more faults with my eyes than with any oscilloscope.

Pay special attention to the gate drive circuit. If there’s a resistor between the gate driver and the MOSFET gate, check it. They often open up under stress. Same goes for the gate-to-source zener diode, if one exists. A shorted zener can pull the gate voltage down and kill switching performance.

Bulging electrolytic capacitors near the primary side? Red flag. They cause ripple, which causes stress, which kills MOSFETs. It’s a chain reaction.

Electrical Testing: Getting Concrete Data

Now we get serious. Grab your multimeter and set it to diode mode. Measure between the gate and source, then between the drain and source on the dead MOSFET. A healthy MOSFET should show open circuit (or a diode drop from source to drain) in most cases. A short between any two pins confirms failure.

But don’t stop there. Check the output rectifiers, the transformer windings for shorts, and the bulk capacitor for proper ESR. I’ve wasted hours chasing a dead MOSFET only to find that the main filter cap had gone open circuit, causing massive ripple that took out the switch.

If you have an oscilloscope, and honestly you should, probe the gate drive signal while the supply is powered through a current-limited bench supply. Look for clean square waves. Ringing, excessive overshoot, or missing pulses all point to gate drive issues that will kill any MOSFET.

The Gate Driver Is Your Best Friend and Worst Enemy

The gate driver is where most people make their biggest mistake. They assume it’s fine because the output looks okay on a scope. But here’s the kicker: the gate driver needs to source enough current to charge the MOSFET’s gate capacitance quickly enough. If the driver is weak, the MOSFET spends too much time in the linear region and cooks itself.

Check the gate driver IC’s supply voltage, its input signals, and its output waveform under load. A gate driver that’s oscillating or has a slow rise time will kill a MOSFET within seconds. I’ve seen it happen on a prototype board where the trace length between the driver and the gate was just a few millimeters too long, adding inductance that caused ringing.

Common Symptoms: Recognizing a MOSFET in Distress

Sometimes you catch a failing MOSFET before it actually dies. That’s the ideal scenario, but it requires knowing what to look for. Power supplies don’t go from perfect to dead in one step. There’s a degradation phase.

Intermittent Operation and Thermal Throttling

If the power supply works fine for a while then shuts down, especially under load, thermal issues are likely. A MOSFET in power supplies that’s approaching its limits will heat up, increase its on-resistance, generate more heat, and eventually trigger over-temperature protection—or fail completely.

Use a thermal camera if you have one. If not, an infrared thermometer works well. Compare the temperature of the suspect device to a known good unit. A difference of more than twenty degrees under the same load is a concern.

Audible Noise and Ripple

Listen to the supply. A healthy power supply is nearly silent. If you hear hissing, buzzing, or whining, that’s often the transformer or the inductor vibrating due to irregular current waveforms caused by a dying MOSFET. Check the output ripple with a scope. Excessive ripple, especially at the switching frequency, points directly to switching issues.

Preventing MOSFET Failure: Design Lessons That Stick

Once you’ve fixed the dead MOSFET and resolved the root cause, it’s time to think about prevention. The best repair is the one you never have to repeat. Here are the lessons I’ve learned the hard way.

Snubber Networks Are Not Optional

Many designers treat snubber networks as an afterthought. Big mistake. A properly designed RC snubber across the drain-to-source of the MOSFET can reduce voltage spikes by thirty to fifty percent. That’s often the difference between a reliable supply and one that fails after six months.

If the original design didn’t have a snubber, and you’re seeing repetitive failures, consider adding one. Measure the ringing frequency with a scope, then calculate the appropriate capacitor and resistor values. It’s not hard, and it pays off immediately.

Derating Is Your Insurance Policy

Here’s a rule I live by: never run a MOSFET at more than eighty percent of its rated voltage or seventy percent of its rated current in a power supply. Yes, the datasheet says it can handle 600 volts and 20 amps. That doesn’t mean you should push it there in real-world conditions.

Temperature derating is even more critical. If your MOSFET is rated for 150 degrees junction temperature, design for a maximum of 110 degrees under worst-case conditions. That margin absorbs component tolerances, aging, and unexpected load variations.