IRF540NPBF - 100V/33A N-Channel HEXFET Power MOSFET, TO-220, by Infineon

What Is the IRF540NPBF and How Does It Work?

The IRF540NPBF is an N-channel enhancement-mode vertical DMOS power transistor. Apply a positive voltage between gate and source, and a conductive channel forms between drain and source. At Vgs = 10V, the channel resistance drops to 44mΩ maximum - at 10A, that's 4.4W of conduction loss. Manageable with a heatsink. At Vgs = 5V, the channel is only partially formed and Rds(on) can be 200mΩ or higher - at 10A, that's 20W in a TO-220 package with θJA of 62°C/W. Do the math: 20W × 62 = 1240°C rise. The part hits thermal shutdown in under a second.

The device uses Infineon's planar HEXFET structure, which gives it good avalanche ruggedness and reasonable switching speed. Gate charge is 71nC typical at Vgs=10V. Input capacitance is 1960pF at Vds=25V. These numbers determine how fast you can switch it and how much current your gate driver needs to supply. With a proper gate driver (TC4427, IR2104, or a discrete totem-pole), you can switch the IRF540N at 20–100 kHz. Without a gate driver - just a resistor from a microcontroller pin - the gate voltage rise and fall times stretch into microseconds, the MOSFET lingers in the linear region during every transition, and switching losses dominate. The part gets hot. Then it gets hotter because Rds(on) has a positive temperature coefficient. Then it fails.

The TO-220AB package has three pins: gate (pin 1), drain (pin 2 and the metal tab), source (pin 3). The tab is electrically connected to the drain. If you bolt it directly to a grounded heatsink, you're shorting the drain to ground through the heatsink. You need either an insulating thermal pad (silicone or mica) plus a shoulder washer for the screw, or a dedicated isolated heatsink that floats at drain potential. This trips up beginners constantly. Confirmed in hundreds of forum threads across the Infineon community, EEVblog, and Electronics StackExchange.

What Are the Specifications of IRF540NPBF?

Numbers That Actually Matter in a Design

• The 33A headline rating is not what you get in a real circuit. At 25°C case temperature, yes, 33A continuous. But your MOSFET is not at 25°C when it's conducting 33A. At 100°C case, the rating drops to 23A. And that's with an infinite heatsink holding the case at 100°C. In practice, with a reasonable finned heatsink, you're looking at 10–15A continuous for a reliable design. The 33A number on the datasheet front page is the silicon limit with the case held at 25°C by a laboratory cold plate. Your PCB is not a laboratory cold plate. Derate by at least 50% for real-world thermal design.
• Gate drive voltage is everything. Rds(on) = 44mΩ is guaranteed only at Vgs = 10V. At Vgs = 8V, Rds(on) is higher but the datasheet doesn't specify by how much. At Vgs = 6V, you're still above the threshold but nowhere near full enhancement. At Vgs = 5V (Arduino logic level), the MOSFET is barely on - Rds(on) might be 150–300mΩ depending on the specific part and temperature. At 5A, that's 3.75–7.5W dissipation. The TO-220's θJA with no heatsink is 62°C/W, so junction temperature rises 233–465°C above ambient. The part hits 175°C Tj_max and starts degrading within seconds. Every "my IRF540N is overheating at 5A" forum post traces back to this same root cause: insufficient gate drive voltage. The fix is either a gate driver that provides 10–12V Vgs, or switching to the IRL540NPBF (logic-level version, fully enhanced at 5V).
• The body diode is slow and lossy. The intrinsic anti-parallel diode between source and drain has a forward voltage of 1.2V at 16A and a reverse recovery time around 200ns. In a hard-switched half-bridge or synchronous buck converter, that reverse recovery charge gets dumped as loss in the opposing MOSFET every cycle. At 50 kHz with 200ns reverse recovery, the diode is conducting in reverse for 1% of each cycle - and at 16A, that's real power. If you're building a synchronous converter, add an external Schottky diode (or use a MOSFET with a fast body diode like the IRF540Z). For low-side switching of inductive loads (relays, solenoids, motors), the body diode works fine as a freewheeling clamp - it's only an issue in hard-switched topologies.
• The IRF540N is not for linear/audio applications. This part is designed and characterized for switching. In linear mode - where the MOSFET operates partially on, like in a class AB audio amplifier output stage or a linear regulator pass element - the planar HEXFET structure suffers from what's called the Spirito effect. Hot spots form on the die because the threshold voltage has a negative temperature coefficient at low currents, causing current crowding into localized regions. The result is sudden catastrophic failure at power levels well below the 130W rating. The repair community on diyaudio.com and EEVblog has documented this extensively: when repairing audio amplifiers, do not replace the original MOSFETs with IRF540N unless the original was specifically an IRF540N. Use the plain IRF540PBF (Vishay) or a lateral MOSFET (Exicon, Semelab) designed for linear operation. Confirmed in multiple Phoenix Gold, Rockford Fosgate, and car audio amplifier repair threads.
• At 100V, you have margin for 48V systems - but not much else. A 48V battery fully charged is about 54.6V (for Li-ion 13S). Add ringing and inductive kickback on the drain, and you're at 70–80V. The 100V rating gives you headroom. But for anything above 60V nominal, the margin shrinks fast. A 72V system charges to 84V. With switching transients, you're bumping against 100V regularly. For 72V and above, step up to a 150V or 200V part (IRFP250N, IRFP460). The IRF540N is a 48V-and-below MOSFET. Using it at higher voltages is gambling with the avalanche rating.

What Is IRF540NPBF Used For?

DC Motor Drives (24V–48V)

Brushed DC motor H-bridges operating from 24V or 48V battery packs. The 100V rating absorbs inductive flyback from the motor windings when the MOSFETs switch. At 33A, a single IRF540N per leg handles most small-to-medium motor applications: e-bike controllers, electric scooters, robot drive trains, CNC spindle drivers. Use a gate driver IC (IR2104 for half-bridge, HIP4081 for full H-bridge) to get clean 10–12V gate drive. Do not drive the gates directly from a microcontroller. A properly designed 24V/20A H-bridge with four IRF540N MOSFETs, adequate heatsinking, and 10V gate drive will run for years.

DC-DC Converters (Push-Pull, Forward, Flyback)

The IRF540N works well in isolated DC-DC topologies at 20–80 kHz. The 71nC gate charge is moderate - higher than modern trench MOSFETs but adequate for these frequencies. The 44mΩ Rds(on) keeps conduction losses manageable. In a 48V-to-12V forward converter at 100W, conduction losses at 50% duty are roughly 2.5W. With a small heatsink (θSA ~10°C/W), junction temperature stays under 100°C. Push-pull converters at 24V input are the sweet spot: two IRF540N MOSFETs, a center-tapped transformer, and a TL494 or SG3525 controller make a robust, simple design that's been the standard approach for decades.

Battery Protection and Load Switching

Used as a low-side or high-side switch to disconnect loads from 12V–48V batteries. For low-side switching (MOSFET between load and ground), gate drive is straightforward: connect gate to battery positive through a 10kΩ resistor, and pull it to ground with an NPN transistor or open-drain GPIO to turn off. For high-side switching (MOSFET between battery and load), you need a gate voltage higher than the battery voltage plus Vgs(th) - use a dedicated high-side gate driver or a charge pump. This is a common pitfall: connecting an IRF540N as a high-side switch with the gate driven by a 5V microcontroller referenced to ground. The source follows the gate, so with 5V on the gate, the source sits at about 1–2V - the load sees a fraction of the battery voltage while the MOSFET dissipates the rest as heat.

Solenoid and Relay Drivers

Industrial solenoid valves, relay coils, and electromagnetic actuators running on 24V or 48V DC. The IRF540N switches the coil current, and a freewheeling diode across the coil clamps the inductive kickback. The MOSFET's avalanche rating provides additional protection if the diode fails or is too slow. At 100V breakdown, even a 48V coil with significant inductance won't avalanche the MOSFET if the diode is present. Add a TVS diode from drain to source for extra margin in harsh industrial environments where the DC bus may have transients.

Pinball and Arcade Machine Flipper/Hammer Drivers

The IRF540N (and its logic-level cousin IRL540N) is the standard flipper coil driver MOSFET in Sega, Data East, and Stern pinball machines from the 1990s through today. The flipper coil draws a massive current pulse (10–20A) for a few milliseconds to kick the ball, then drops to a lower hold current. The MOSFET sees high peak current but low average power. The original parts in these machines are now 20–30 years old and starting to fail. The repair is well-documented on Pinside and in the pinball repair community. If replacing a blown flipper MOSFET, also check the coil diode - a failed flyback diode is what killed the MOSFET in the first place. Replace both together.

IRF540NPBF vs IRL540NPBF vs IRFZ44NPBF - Which MOSFET for Your Design?

Pick IRF540NPBF when: You're switching 24–48V loads and you have a proper gate driver providing 10–12V gate voltage. You need the 100V rating for voltage margin. The 44mΩ Rds(on) is acceptable for your current levels. This is the default choice for motor drives, DC-DC converters, and industrial switching at moderate voltages.

Pick IRL540NPBF when: You need to drive the MOSFET directly from a 5V microcontroller or logic signal, no gate driver. The "L" stands for logic-level - the gate threshold is 1.0–2.0V, so 5V fully enhances the channel. The trade-off is higher Rds(on) at 58mΩ versus 44mΩ for the standard version. For low-current switching (under 5A) from an Arduino, ESP32, or Raspberry Pi, the IRL540N is the right choice. For higher currents, use the IRF540N with a gate driver - the lower Rds(on) more than pays for the cost of the driver IC.

Pick IRFZ44NPBF when: Your voltage is 24V or below and you want the lowest possible conduction loss. The IRFZ44N has Rds(on) of 17.5mΩ - 60% lower than the IRF540N. At 20A, the IRFZ44N dissipates 7W while the IRF540N dissipates 17.6W. That's the difference between a warm heatsink and a molten one. But the IRFZ44N is only rated for 55V, so it's strictly for 12V and 24V systems. Using it at 48V is beyond its rating.

Frequently Asked Questions About IRF540NPBF

Q1: Can I drive the IRF540NPBF directly from an Arduino/ESP32/Raspberry Pi GPIO pin?

A: No. Arduino and ESP32 GPIOs output 5V or 3.3V respectively. The IRF540NPBF needs 10V gate-to-source for full enhancement with Rds(on)=44mΩ. At 5V gate drive, the MOSFET is partially on with Rds(on) potentially 150–300mΩ - 4–7× higher than spec. At 3.3V, it's barely conducting at all. The solution: use a gate driver IC (TC4427, IR2104, or a simple NPN/PNP totem-pole) to translate the logic signal to 10–12V gate drive, or switch to the logic-level IRL540NPBF which is fully enhanced at 5V. This is the most-asked question about this part across every electronics forum and the answer never changes.

Q2: What's the difference between IRF540, IRF540N, IRF540NPBF, and IRL540NPBF?

A: IRF540 is the original (77mΩ Rds(on)). IRF540N is the newer generation (44mΩ Rds(on)). The "PBF" suffix means lead-free / RoHS compliant - electrically identical to the non-PBF version. IRL540N is the logic-level variant: lower gate threshold (1.0–2.0V), fully enhanced at 5V Vgs, but higher Rds(on) at 58mΩ. The "L" in IRL540N stands for Logic-level. Same package, same pinout, completely different gate drive requirements. An Arduino can drive the IRL540N directly. It cannot drive the IRF540N directly. This distinction is critical and confusing the two is the most common design mistake with this part.

Q3: Why is my IRF540NPBF getting burning hot at only 5A?

A: Three likely causes, in order of probability. One: insufficient gate drive voltage. If Vgs is below 10V, Rds(on) is much higher than 44mΩ. At 5V Vgs, expect 150–300mΩ. At 5A, that's 3.75–7.5W. With no heatsink and θJA=62°C/W, the junction rises 233–465°C above ambient - the part hits thermal limits almost instantly. Two: no heatsink. Even with proper 10V gate drive, 5A × 5A × 44mΩ = 1.1W. With θJA=62°C/W, that's 68°C rise - the part will be too hot to touch (~93°C at 25°C ambient). Add even a small clip-on heatsink. Three: oscillating gate. A long gate trace without a series resistor can ring at several MHz, turning the MOSFET partially on and off at RF frequencies. Add a 10–47Ω gate resistor right at the MOSFET pin. Confirmed on Infineon Community, Electronics StackExchange, and EEVblog in dozens of threads.

Q4: Can I use IRF540NPBF in an audio amplifier?

A: Not recommended. The IRF540N is designed and characterized for switching, not linear operation. In linear mode (class AB amplifier output stage, linear regulator pass element), planar HEXFETs suffer from the Spirito effect: localized hot spots form on the die because the threshold voltage's negative temperature coefficient causes current hogging in small regions. The result is sudden catastrophic failure, often at power levels far below the 130W datasheet rating. If you're repairing an audio amplifier that originally used IRF540 or IRF540N, replace like-for-like. If the original parts were lateral MOSFETs (Exicon, Semelab) or plain IRF540 from Vishay, do not substitute IRF540N. The repair community on diyaudio.com and EEVblog has documented this failure mode extensively. For new audio designs, use purpose-built lateral MOSFETs.

Q5: Does the TO-220 tab need to be insulated from the heatsink?

A: Yes, unless your heatsink is electrically floating and not connected to anything else. The TO-220 tab is internally connected to the drain pin. If you bolt it to a grounded heatsink, you're shorting the drain to ground. Use a silicone thermal pad or mica washer between the tab and the heatsink, plus a plastic shoulder washer to insulate the mounting screw. Apply thermal grease on both sides of the insulator. After mounting, check with a multimeter that there's no continuity between the tab and the heatsink. This trips up beginners regularly and a drain-to-ground short through the heatsink can destroy the MOSFET, the power supply, and whatever the MOSFET was driving.

Q6: What gate resistor value should I use with the IRF540NPBF?

A: 10–47Ω is typical, placed as close to the gate pin as physically possible. The gate resistor serves two purposes: damping parasitic LC ringing between the gate trace inductance and the MOSFET's input capacitance (1960pF), and limiting the peak current that the gate driver must supply. Too high (>100Ω): the gate charges slowly, turn-on and turn-off times stretch into microseconds, and switching losses become dominant. Too low (<10Ω): the gate voltage rings, potentially exceeding the ±20V Vgs maximum during transients. For PWM above 20 kHz, use the lower end (10–22Ω). For on/off switching of DC loads, 47–100Ω is fine. Always add a 10kΩ gate-to-source pulldown resistor to ensure the MOSFET is off during power-up when the gate driver output is high-impedance.

Q7: Is the IRF540NPBF still in production?

A: Yes. Active production, widely available. Infineon has transitioned manufacturing across multiple fabs over the years - originally International Rectifier, then Infineon's own fabs, and more recently contract manufacturing at Sichuan Gen Microelectronics in China. The part number and specifications have remained consistent across all fab transitions. The IRF540N is one of the most produced power MOSFETs in history and will be available for the foreseeable future. Counterfeits are a real concern at this volume level - buy from a source that batch-verifies Rds(on). From our experience, a genuine IRF540NPBF at 25°C will measure within 10% of the 44mΩ spec. A counterfeit will measure 80–200mΩ. The difference is unmistakable with even a basic milliohm meter.

Q8: What is the maximum PWM frequency I can run the IRF540NPBF at?

A: With a proper gate driver supplying at least 500mA peak, 20–100 kHz is practical. The limiting factors are gate charge (71nC) and the thermal budget for switching losses. At 100 kHz with 50% duty, switching losses roughly equal conduction losses. Above 100 kHz, switching losses dominate and efficiency drops rapidly. Below 20 kHz, you're in the audible range and the MOSFET itself may produce an audible whine from magnetostriction in the package and PCB. For most applications, 20–50 kHz is the sweet spot. If you need to switch above 100 kHz, use a modern trench MOSFET with lower gate charge (IRF540Z at 43nC, or a GaN FET for very high frequencies).

Pricing & Availability

Contact ICMASS for current pricing and volume quotes. We stock genuine Infineon IRF540NPBF MOSFETs with batch-tested Rds(on) verification. Also available: IRL540NPBF (logic-level version for direct MCU drive), IRFZ44NPBF (55V/49A for lower voltage applications), and the full HEXFET MOSFET lineup. For gate driver ICs (TC4427, IR2104, IR2110) to pair with your IRF540N order, ask our team for a bundled quote.