The Bench Reality: Why Transistors Fail and How to Spot It

Transistors rarely fail gracefully. When they die, they usually take out the microcontroller driving them, the load they were switching, or both. Before we get into the math of biasing, you need to understand how and why these silicon switches fail in the real world. Transistor checking isn't just about verifying a part before you solder it; it's about forensic analysis when a circuit board starts smelling like burnt epoxy.

Bench War Story: The Darlington Trap
Setup: I was driving a 12V, 30A automotive fuel pump using a TIP120 Darlington transistor on a custom perfboard for an automated fuel transfer system.
Numbers: Base current was set to 5mA from an ESP32 GPIO (via a 1kΩ resistor), assuming the TIP120's massive hFE (1000+) would easily saturate it.
Outcome: The TIP120 overheated, cracked its TO-220 package open, and shorted the 12V rail directly to the ESP32's GPIO pin, frying the microcontroller.
What went wrong: I ignored the datasheet's saturation voltage. The TIP120 has a massive VCE(sat) of ~2V to 2.5V at high currents. At 30A, that 2V drop equals 60W of thermal dissipation. Without a massive heatsink, the silicon junction exceeded 150°C in under four minutes, causing a catastrophic thermal runaway and collector-emitter short.

That failure highlights the core rule of semiconductor design: never trust a transistor's current gain (hFE) for switching high loads without checking its saturation voltage and thermal limits. When a transistor fails, it typically fails in one of two ways: a short circuit (Collector-to-Emitter or Drain-to-Source) due to thermal overload, or an open circuit (internally blown bond wire) due to overcurrent. Let's look at how to test for both.

Transistor Checking with a Multimeter: Step-by-Step

You don't need a curve tracer to verify a transistor. A standard digital multimeter (DMM) in diode-test mode is all you need. Before testing, note the physical pinout. For a standard TO-92 package (like a 2N3904) with the flat side facing you and legs pointing down, the pins are Emitter (E), Base (B), and Collector (C) from left to right. For a TO-220 package (like a TIP120 or IRFZ44N) with the metal tab facing away from you and legs pointing down, the pins are Base/Gate (1), Collector/Drain (2), and Emitter/Source (3).

Always desolder at least the base/gate leg from the circuit before testing. In-circuit transistor checking is notoriously unreliable due to parallel current paths through other components.

Testing a Bipolar Junction Transistor (BJT)

  1. Set your DMM to Diode Test mode. The display should show an open loop (OL) or a flashing 1 when probes are separated.
  2. Identify the Base. Place the red probe on the Base and the black probe on the Emitter. A good NPN silicon transistor will read between 0.550V and 0.750V. Reverse the probes; it should read OL.
  3. Check Base-to-Collector. Red on Base, black on Collector. You should see the same 0.550V–0.750V forward voltage drop. Reverse for OL.
  4. Check Collector-to-Emitter. Place probes across C and E in both directions. Both must read OL. If you read a short (0.000V or a beep), the transistor is blown.
  5. PNP Verification: For a PNP transistor (like a 2N3906), swap the probe colors. Black on Base, red on E/should yield the 0.6V drop.

Testing an N-Channel MOSFET

MOSFETs are voltage-controlled and have an internal body diode, which changes the testing procedure. According to semiconductor fundamentals outlined by All About Circuits, checking the gate oxide integrity is just as critical as checking the channel.

  1. Discharge the Gate. Touch the Source and Gate pins together briefly with your finger or a probe to drain any stored gate capacitance.
  2. Check the Body Diode. Red probe on Source, black probe on Drain. You should read the body diode forward voltage (usually around 0.4V to 0.6V). Reverse probes (Red on Drain, Black on Source); it should read OL.
  3. Check for Gate Shorts. Measure between Gate and Source, and Gate and Drain. Both must read OL. Any low resistance here means the gate oxide is punctured.
  4. The Charge Test. While keeping the black probe on the Source, touch the red probe to the Gate to charge it via the DMM's internal battery. Move the red probe back to the Drain. The DMM should now read a low voltage drop (the MOSFET is turned on). Short Gate to Source again to turn it off.

Operation Regions and Biasing for the Job

Selecting and biasing a transistor depends entirely on what job you need it to do. If you are building an audio amplifier, you want the transistor in the Active (Linear) region. If you are switching a relay or a motor with a microcontroller, you want it in the Saturation region (for BJTs) or the Triode/Ohmic region (for MOSFETs). Driving a switching transistor in the active region is the #1 cause of the thermal failures described in the war story above.

Transistor Operation Regions and Typical Bench Values
Region BJT (NPN) State MOSFET (N-Ch) State Typical V_CE / V_DS Primary Application
Cutoff V_BE < 0.6V (I_B = 0) V_GS < V_th (I_D = 0) Equal to Supply Voltage (V_CC) Switch OFF (Open circuit)
Active / Linear V_BE ≈ 0.7V, V_CE > 0.3V V_GS > V_th, V_DS > V_GS - V_th 1.0V to (V_CC - 1V) Signal amplification, linear regulators
Saturation / Triode V_BE ≈ 0.7V, I_B forced high V_GS >> V_th, V_DS < V_GS - V_th < 0.3V (BJT) or I_D * R_DS(on) (MOSFET) Switch ON (Closed circuit, PWM)

To bias a BJT for switching, you must use forced beta. Never use the datasheet's hFE (which is measured in the active region). For a switching circuit, assume a forced beta of 10. If your load requires 500mA of collector current ($I_C$), you must supply at least 50mA of base current ($I_B$) to guarantee the transistor enters deep saturation, minimizing $V_{CE(sat)}$ and heat generation. For deeper insights into calculating exact biasing networks, refer to the ON Semiconductor TIP120 datasheet and its associated application notes on thermal derating.

A Complete Application Circuit: Driving a 12V Relay

Let's build a robust, real-world switching circuit. We want to drive a standard 12V automotive-style relay (coil resistance = 100Ω, meaning $I_C$ = 120mA) using a 3.3V logic GPIO from an ESP32 or Raspberry Pi Pico. We will use a standard 2N2222 NPN BJT.

Component Selection and Biasing Math

  • Q1: 2N2222 (NPN BJT, TO-92 or TO-18). Max $I_C$ is 800mA, so 120mA is well within limits.
  • Forced Beta: We will use a forced beta of 10 to ensure hard saturation.
  • Required Base Current ($I_B$): $I_C / 10 = 120mA / 10 = 12mA$.
  • Base Resistor ($R_B$): The GPIO outputs 3.3V. The Base-Emitter junction drops ~0.7V. The voltage across $R_B$ is $3.3V - 0.7V = 2.6V$. Using Ohm's Law: $R = V / I = 2.6V / 0.012A = 216Ω$. Standard value: 220Ω.
  • Resistor Wattage: $P = I^2R = (0.012)^2 * 220 = 0.031W$. A standard 1/4W (0.25W) resistor is more than sufficient.
  • Flyback Diode (D1): 1N4007 or 1N4148. Placed in reverse bias across the relay coil (cathode to 12V, anode to the transistor's collector). This clamps the inductive kickback spike when the transistor turns off, preventing the collector voltage from spiking to 100V+ and punching through the transistor's junction.
Wiring Checklist:
1. ESP32 GPIO → 220Ω Resistor → 2N2222 Base.
2. 2N2222 Emitter → System Ground.
3. 12V Supply → Relay Coil Pin 1.
4. Relay Coil Pin 2 → 2N2222 Collector.
5. 1N4007 Anode → 2N2222 Collector; 1N4007 Cathode → 12V Supply.

Safe Default Part Numbers for Your Component Drawer

Stop buying random transistor kits from Amazon that lack proper datasheets. Stock your bench with these proven, industry-standard defaults. These parts are available from reputable distributors like Mouser, Digi-Key, or Arrow, ensuring you aren't getting counterfeit silicon that will fail under load.

Bench Default Transistors and Ratings
Part Number Type & Package Max V_CE / V_DS Max I_C / I_D Best Use Case
2N3904 NPN BJT (TO-92) 40V 200mA Low-power switching, logic level translation, LED drivers.
2N3906 PNP BJT (TO-92) -40V -200mA High-side switching, complementary pairs with 2N3904.
2N2222 / PN2222 NPN BJT (TO-92/TO-18) 40V 600mA Medium-power switching, relay drivers, small motor control.
2N7000 N-Ch MOSFET (TO-92) 60V 200mA Logic-level low-power switching, I2C pull-ups, signal routing.
IRLZ44N N-Ch MOSFET (TO-220) 55V 47A High-current loads, PWM motor control, solenoid drivers. (Logic-level gate).
TIP120 NPN Darlington (TO-220) 60V 5A High-gain applications where base current is severely limited. Requires heavy heatsinking due to high VCE(sat).

When upgrading from a BJT to a MOSFET for high-current switching (like the fuel pump scenario from the intro), the IRLZ44N is the ultimate bench savior. Because it is a logic-level MOSFET, its $R_{DS(on)}$ is fully specified at a $V_{GS}$ of 5V (and even 4V). At 30A, an IRLZ44N with an $R_{DS(on)}$ of roughly 0.022Ω will dissipate only $I^2R = (30)^2 * 0.022 = 19.8W$. While that still requires a heatsink, it is a massive 66% reduction in waste heat compared to the TIP120's 60W, and it won't thermally destroy your microcontroller if you forget the thermal paste.

Proper transistor checking and biasing isn't about memorizing physics equations; it's about respecting the thermal and electrical limits of the silicon. Keep your DMM in diode mode, calculate your forced beta, and always use a flyback diode on inductive loads.