Testing a MOSFET with a Multimeter: The Capacitance Sign, Diode Check, and Short Detection

What a multimeter can and cannot prove about a MOSFET: the body-diode check, the gate limit that kills parts, and what only a curve tracer shows.

Request a Quote

Testing a MOSFET with a Multimeter: The Capacitance Sign, Diode Check, and Short Detection

Testing a MOSFET with a multimeter feels like testing a capacitor, and that feeling is the clue. A good MOSFET holds a charge on its gate like a capacitor, the body diode can restore that charge, and a shorted part reads like a diode in both directions — the fingerprint of the failed device. This article explains why the gate-capacitance test works, maps gate, drain, and source on common packages, shows the body-diode check and its limits, works the short and leak detection the meter can prove, and closes by separating the MOSFET test from the BJT test.

Last updated: October 11, 2026

Why a MOSFET Test Feels Like a Capacitor Test

The first surprise of MOSFET testing is that a good part does not read like a diode at all; it reads like a tiny capacitor. The reason is the gate structure. The MOSFET’s gate is a metal layer on a thin insulating oxide above the semiconductor; between gate and body, the structure is literally a capacitor — the oxide is the dielectric, the gate is one plate, and the body is the other.

When a meter touches the gate, it charges that capacitor; the meter’s test voltage deposits charge on the gate, and the part stores it. A good MOSFET therefore shows a charging response — the meter reading rises or falls as the gate capacitor charges, and it discharges back when the probe is removed. This capacitance sign is the healthy fingerprint, and it is the basis of the whole test method. The MOSFET datasheet article and the power MOSFET selection guide develop the structure this article tests; the NPN transistor test article covers the sibling family, and the final section here separates the two meter behaviors.

The practical test habit is to look for the charging sign, not the zero reading. A meter that charges the gate then lets it discharge is seeing a healthy MOSFET; a meter that reads open with no charge or discharge is seeing a broken gate path; a meter that reads like a diode both ways is seeing a shorted part. The capacitance sign is the anchor the rest of the tests read against.

Pin Mapping: Gate, Drain, Source on Common Packages

The gate-capacitance test only works if the pins are identified correctly, and the pin mapping is the first skill. The three terminals — gate, drain, source — appear on common packages in recognizable patterns.

On a typical through-hole or D2PAK package, the gate is the terminal that charges like a capacitor, the drain and source are the terminals of the body. On an axial-flavored package the body terminals map by the package convention, and on the SMD packages the pin layout follows the datasheet drawing. The reliable mapping habit is to identify the gate by its capacitance behavior first — the terminal that charges and discharges — and then identify the drain and source relative to it, using the body diode as the reference. The MOSFET datasheet article and the MOSFET polarity article give the pin conventions and the polarity meaning of the body.

The mapping mistake costs the whole test: a meter touching the wrong terminals reads a capacitor or a diode that is not the part, and the verdict is nonsense. The mapping is the safety rail for everything that follows, and it is worth the thirty seconds it takes to confirm before testing.

Good-Ark power MOSFETs in common packages used in the multimeter testing guide — Power MOSFET devices in common packages whose gate, drain, and source terminals are mapped in the multimeter testing guide, from the Good-Ark power MOSFET category

The expected reads are compact enough for a table, and the table is the bench reference:

Test	Healthy N-channel	Broken / shorted	What the meter shows
Gate to body	Charges then discharges	No charge, or low-resistance leak	Capacitance sign appears and decays
Drain to source (forward)	Body diode conducts	Conducts both ways	One-way read is healthy
Drain to source (reverse)	Body diode blocks	Conducts both ways	Both-way read is a short
Gate to drain	Charges like a capacitor	Low resistance	The short is visible on the gate

The table is the method in one view: each probe placement has a healthy read and a broken read, and the broken read is usually the opposite of healthy rather than a subtler variant. The power MOSFET selection guide and the MOSFET polarity article give the N-channel and P-channel directions that the table’s polarity column assumes.

What the Meter Can and Cannot Prove

A diode-mode reading proves a small set of things and says nothing about the rest. Knowing which column a reading falls into is what stops a “good” part from going back onto a board.

The meter can prove	The meter cannot prove
The body diode is present and oriented the right way	What the on-resistance is at your gate drive voltage
Gate-source is shorted, or drain-source is shorted	Whether the gate oxide is degraded but still holding
Gross gate leakage at the meter’s small test voltage	Gate leakage at the rated gate voltage
The package pins connect to the die at all	Whether the thermal path is intact, since a delaminated die still measures fine

A real part shows why the second column is the one that matters after a fault. onsemi specifies the NTMFS5C430NL at ±20 V maximum gate-to-source, 900 A of pulsed drain current for a 10 µs pulse, 493 mJ of single-pulse avalanche energy, a 1.4 °C/W junction-to-case resistance, and a 175 °C junction limit. None of those four numbers can be confirmed with a two-lead meter on a bench.

The gate limit deserves its own paragraph, because it is where bench work damages parts. A multimeter in diode or resistance mode sources a few hundred microamps at a couple of volts — far below the ±20 V limit — so the instrument itself will not harm the gate. The operator will. An ungrounded hand can deliver kilovolts, and the gate oxide is the thinnest structure in the device, which is why a part that measured perfectly can be dead by the time it reaches the board.

Two practical boundary conditions follow. A part that survives a fault may still have spent avalanche energy against its 493 mJ budget, so a device that reads normally after a short-circuit event is not automatically a device you should reinstall. And because the junction-to-case path is 1.4 °C/W while the ambient path on a 650 mm² pad is 40 °C/W, the same reading on the meter says nothing about which of those two paths your board will actually provide.

The Body-Diode Check and Its Limits

The second reading is the body diode. Every MOSFET has a body diode between the drain and the source — the junction that forms between the semiconductor and the base — and it is the most diode-like reading the part gives.

On an N-channel part, the body diode conducts from source to drain: the meter in diode mode reads conduction in one direction and blocking in the other. That one-way reading is a healthy sign. The body-diode check is fast and useful, but it has limits that the test must respect. First, the body diode’s VF is a fraction of a volt to a couple of volts, and the meter’s diode mode must be able to read it. Second, the body diode is not the whole MOSFET: a part with a healthy body diode can still have a broken gate or a failed channel, because the two structures fail independently. Third, the body-diode reading is polarity-sensitive, so the N-channel and P-channel parts read in opposite directions. The MOSFET polarity article and the BJT test article cover the direction and the family distinction.

The body-diode check is a screen, not a verdict. A part with no body diode is definitely broken; a part with a healthy body diode needs the gate-capacitance and short checks to complete the story. The three tests together are the full meter method, and the limits of each keep the verdict honest.

Detecting Shorts and Leaks: What the Meter Can Prove

The short detection is where the meter earns its keep, and the shorted MOSFET has a distinctive fingerprint: it reads like a diode in both directions. The explanation is that a short between the gate and the channel, or between the drain and the source, turns the part into a conductive path that the meter sees as conduction either way.

The leak detection is the subtler read. A leaky gate, a damaged oxide, or a degraded channel shows as a resistive path where a healthy part reads charging or open. The ohms-continuity check on the gate, held against the known behavior, exposes the leak: a gate that should read as a capacitor but reads as a low resistance is leaking. The MOSFET testing and the power discrete package testing article cover the continuity and leak checks in the production-inspection context, and this article applies them to the bench.

The honest boundary of the meter test is that it proves open-circuit, continuity, and diode behavior, not dynamic performance. A part that passes the meter is a structurally intact part; it is not proven to switch at speed or carry its rated current. The meter is the first gate of the diagnosis, and the dynamic checks belong to the bench that can apply load.

Schottky rectifier devices in the same discrete package family whose diode-mode behavior is compared in the MOSFET test guide, from the Schottky category

The in-circuit caveat completes the bench picture. Testing a MOSFET while it is still soldered to a board reads the parallel path of the surrounding circuit, not the bare part, so the capacitor sign and the body-diode read can be masked by the neighboring components. The practical rule is to lift one lead or desolder the part before trusting a verdict, or to treat the in-circuit read as a screen that points to the suspect part rather than a proof. The power discrete package testing article documents the in-circuit and out-of-circuit distinction in the production-inspection context, and it is the same caution the repair bench follows when a loaded board reads inconclusively.

MOSFET vs BJT Meter Tests: When Ohms Confuses

The final distinction is the one that confuses the most benches: the MOSFET test versus the BJT test. The two families are tested differently, and using the wrong method produces a wrong verdict.

The BJT is tested as a pair of junctions: base, collector, and emitter read in diode mode with the base-collector and base-emitter junctions behaving like diodes. The MOSFET is tested as a capacitor plus a body diode: gate, drain, and source read with the gate as a capacitive terminal and the drain-source as a diode. The confusion arises in ohms mode, where a multimeter’s continuity test can mislead: the BJT’s junctions read conductive in one direction, and the MOSFET’s body diode also reads one-way, but the gate terminal has no diode equivalent in the BJT. The BJT test article and the MOSFET selection guide document the two families; the bench habit is to name the family before the test.

The rule that keeps ohms from confusing: if the part has a terminal that charges like a capacitor, test it as a MOSFET; if it shows two diode junctions and no gate capacitance, test it as a BJT. The meter method follows the structure, and the structure names the verdict.

The MOSFET meter test is a three-part method built on one physical fact: the gate is a capacitor, the body is a diode, and a short reads like conduction both ways. Map the pins, watch the charging sign, check the body diode with its limits, and read the shorts and leaks the meter can prove. Do not confuse the gate-capacitance part with the BJT’s junction pair, and the multimeter becomes a fast, honest first gate for every MOSFET that lands on the bench. The power MOSFET category supplies the parts whose structure this guide tests, and the MOSFET selection guide completes the family context.

FAQs

Can a multimeter damage a MOSFET gate?

Not through its own test current, which is small and well below the gate rating. Static discharge from the operator is the real risk, so ground yourself and handle the part by the body rather than the pins.

What does a good diode-mode reading actually mean?

It confirms the body diode exists and is oriented correctly, and that the drain and source are not shorted. It does not confirm the on-resistance at your gate voltage, the gate leakage at rated voltage, or that the thermal path is intact.

Is a MOSFET that survived a short circuit safe to reuse?

Not automatically. Avalanche events spend part of a finite energy budget, and a device can measure normally afterwards while having less margin left than it started with. Replace parts that have seen a fault unless you can test them properly.

Why does the meter reading change when I touch the gate?

The gate is a capacitor, and the meter applies a small charge when it probes. That stored charge changes the channel state, so the reading drifts until the gate is discharged again. It is normal behaviour for a healthy device, not a fault.