Transient voltage is a brief, high-energy spike or dip in electrical potential that lasts from microseconds to a few milliseconds, superimposed on the normal operating voltage. When a transient hits your board, it fundamentally changes the dielectric stress on your insulation, forces semiconductors into avalanche breakdown, and induces false triggering in sensitive logic circuits. Hobbyists and junior engineers frequently confuse transients with steady-state overvoltage or voltage sags and swells; while a grid swell might last for seconds and show up on a standard multimeter, a transient is over in microseconds, leaving behind nothing but a punctured silicon junction and a bricked microcontroller as evidence.
The Physics of the Spike (and the Math Behind It)
To understand transients, we have to look at energy storage. Think of it like water hammer in plumbing: when you slam a valve shut, the kinetic energy of the moving water has nowhere to go, creating a massive, destructive pressure spike that rattles your pipes. In electronics, current flowing through an inductor is that moving water, and the switch (transistor or relay contacts) is the valve.
Let us run a worked numeric example using a standard Omron G2R-1-24VDC relay. The coil resistance is 650Ω, meaning the steady-state current at 24V is roughly 37mA. The coil inductance measures at approximately 50mH. When the driving NPN transistor switches off, the current must drop from 37mA to 0A. If the transistor turns off in 1 microsecond (1µs), we calculate the rate of current change (di/dt):
- di: 0.037 A
- dt: 0.000001 s
- di/dt: 37,000 A/s
Using the inductor voltage equation V = L × (di/dt), we multiply the inductance (0.050 H) by the rate of change (37,000 A/s). The result is a 1,850V transient spike generated on a nominal 24V circuit. Without a clamping mechanism, this 1.85kV spike will instantly punch through the collector-emitter junction of your driving transistor.
Where You Meet Transient Voltage in Practice
Transients are not just a theoretical problem; they are a daily reality on the bench and in the field. According to power quality research by Fluke, transients are the leading cause of premature degradation in industrial control systems. You will encounter them in two primary domains:
The Mains AC Side
On the grid side, transients are violent and high-energy. Nearby lightning strikes can induce 6kV to 20kV spikes on utility lines. Utility capacitor bank switching and large industrial motor starting also generate severe transients that travel through your service entrance and into your branch circuits.
The Low Voltage DC Side
On the bench, transients are usually self-inflicted. Inductive kickback from relays, solenoids, and stepper motors is the most common culprit. Electrostatic discharge (ESD) from human touch can inject up to 15kV into an exposed I/O pin in dry winter air. In automotive 12V systems, a "load dump" (when the battery disconnects while the alternator is charging) can generate transients hitting 40V to 60V for hundreds of milliseconds, easily destroying unprotected infotainment or sensor modules.
Autopsy of a Bricked ESP32 (Real-World Scenario)
To see how this theory destroys hardware, let us walk through a real-world failure from the workbench.
- The Setup: A builder uses an ESP32 DevKit v1 to control a 12V, 2A pneumatic solenoid valve. The ESP32 GPIO 26 drives the gate of an IRLZ44N logic-level MOSFET. The builder omits a flyback diode to save board space.
- The Numbers: The solenoid coil inductance is measured at 15mH. The steady-state current is 2A. The IRLZ44N MOSFET is driven hard and switches off in roughly 1µs.
- The Outcome: The first time the solenoid disengages, the ESP32 immediately resets. On the third cycle, GPIO 26 becomes permanently shorted to the 3.3V rail, and the onboard AMS1117 voltage regulator gets hot to the touch. The board is dead.
- What Went Wrong: The unclamped theoretical peak voltage is
V = 0.015H × (2A / 0.000001s) = 30,000V. In reality, parasitic capacitance and the MOSFET's avalanche breakdown clamp this to a few hundred volts. However, this massive voltage spike couples through the MOSFET's drain-gate capacitance (the Miller effect) directly back into the ESP32's GPIO pin. The spike vastly exceeds the ESP32's absolute maximum rating of 3.6V, puncturing the silicon dioxide gate oxide inside the microcontroller and permanently shorting the pin to VCC.
Sizing and Selecting Transient Protection
Fixing transient issues requires matching the protection component to the specific threat profile. You cannot use a slow component to catch a fast spike. Below is a comparison of the three most common transient suppression technologies, referencing standard component families like Littelfuse TVS diodes and EPCOS MOVs.
| Technology | Response Time | Energy Handling | Best Use Case | Example Part |
|---|---|---|---|---|
| TVS Diode | Picoseconds (<1ns) | Low to Medium | Protecting sensitive logic, GPIO pins, and data lines from ESD and fast inductive spikes. | Littelfuse SMAJ5.0A |
| MOV (Metal Oxide Varistor) | Nanoseconds (25-50ns) | High | Mains AC surge suppression, power supply inputs, and diverting lightning-induced energy. | EPCOS B72214S |
| RC Snubber | Determined by RC time constant | Medium | Damping high-frequency ringing across relay contacts and triac-switched AC motors. | Custom (e.g., 100Ω + 100nF X2) |
Selection Framework: If you are protecting a microcontroller I/O pin, always choose a TVS diode. If you are protecting the AC mains input of a power supply, use an MOV. If you are trying to stop relay contacts from pitting and arcing, use an RC snubber.
Frequently Asked Questions
Can I just use a standard 1N4007 rectifier diode as a flyback diode?
For slow-switching relays (like a mechanical switch turning off once a second), a 1N4007 is fine. However, if you are using PWM to control a solenoid or motor at high frequencies, the 1N4007's slow reverse recovery time (up to 30µs) means it will not clamp the transient fast enough, and it will overheat. For PWM applications, always use a fast-recovery diode or a Schottky diode like the 1N5819.
Why does my multimeter not show the transient voltage?
Standard digital multimeters sample voltage a few times per second and average the readings. A transient lasting 5 microseconds is entirely invisible to the ADC sampling cycle. To see a transient, you must use an oscilloscope with the trigger mode set to "Single Shot" or "Normal" with a high-frequency probe.
Do TVS diodes degrade over time like MOVs?
Yes, but differently. MOVs degrade significantly with every large surge they absorb, eventually failing short-circuit. TVS diodes are designed to handle thousands of ESD strikes, but if subjected to a massive energy event beyond their peak pulse power rating (e.g., a direct lightning strike on a data line), they will fail catastrophically and short out, protecting the downstream circuit but requiring replacement.
