What Is a Unidirectional TVS Diode? Working, Polarity and Selection

Updated: October 28, 2026

A unidirectional surge-protection (TVS) diode clamps one polarity hard and behaves like a conventional diode the other way. It fits DC rails and signal lines with a defined polarity; AC or symmetric swings usually need bidirectional devices instead.

How does a unidirectional TVS diode work?

A TVS becomes conductive when transient voltage rises. For its usual positive-rail connection, the device is reverse-biased during normal operation. When the applied voltage reaches the relevant avalanche region, current rises and the diode limits voltage according to its current–voltage behavior. The source and circuit impedances determine how much current can flow.

It is a shunt protection element: useful load current normally travels through the supply path, while transient current can be diverted through the TVS. It does not absorb unlimited energy or make the source cease producing a disturbance. A practical design has to support the current path and dispose of the associated heat.

TI’s surge-diode selection note distinguishes working voltage, breakdown, clamp behavior and temperature considerations. Those quantities explain why a TVS should be selected from supported pulse conditions rather than from its name alone. Device behavior is fast, but a blanket sub-nanosecond system-protection claim ignores the mounted loop, source waveform and downstream circuit.

Which terminal connects to a positive DC rail?

The cathode normally connects to the positive node. The anode connects to the selected protection return, so the diode is reverse-biased during ordinary operation. Confirm the package’s marking and pin assignment from the exact datasheet. A cathode band is common on discrete packages, but arrays and special packages can use different terminal arrangements.

Reversing a single-junction unidirectional TVS across a positive supply generally forward-biases it during normal operation. That can draw excessive current and collapse the rail, depending on the source and protection. It does not simply leave an otherwise functioning board with slightly worse surge performance.

The return connection must support the intended transient path. Calling it “ground” does not make its impedance vanish. Placement, connector connection and the protected IC’s reference determine the voltage that matters at the circuit. Verify those nodes together, particularly when chassis, cable shield and circuit ground have different roles. A correct component orientation cannot compensate for an unsuitable return path.

Standoff, breakdown and clamping are different voltages

Each voltage answers a different selection question. Reverse working or standoff voltage, often VRWM, describes the normal blocking range under the stated leakage conditions. Breakdown voltage, VBR, is measured at a specified test current. Clamping voltage, VC, is given at a specified pulse current and waveform. None should be substituted for another without examining the conditions.

Parameter	What it establishes	Common mistake
VRWM	Normal reverse working range at stated leakage conditions	Choosing it from nominal voltage while ignoring tolerance
VBR at IT	Avalanche-region voltage at the test current	Treating it as the maximum voltage during a surge
VC at IPP	Clamp limit under the listed pulse conditions	Applying it at any current or duration
Leakage, IR	Blocking current at a stated voltage and temperature	Calling the device electrically invisible
Capacitance	Loading relevant to the protected node	Using a bulky power TVS indiscriminately on fast signals

Select working voltage against maximum legitimate operation, then check the worst relevant clamp behavior against the downstream circuit’s permitted transient envelope. Include tolerance, temperature and external overshoot. An absolute maximum rating is a limit to respect, not a recommended repetitive operating target. The USB ESD protection guide explains why signal loading needs its own review.

A 12 V rail example that rejects the wrong candidate

Suppose a hypothetical regulated rail can legitimately reach 13.2 V. A candidate has a 15 V working rating, but its supported worst-case clamp at the required current is 24 V. If the downstream IC’s applicable voltage limit is 20 V, that candidate cannot be approved from the 15 V label. It passes the initial working-voltage screen and fails the relevant protection comparison.

Choosing a lower working-voltage device is not automatically the answer: it may leak excessively or conduct during normal high rail voltage. Alternatives can include a lower-clamping protection implementation, a more tolerant input stage, a coordinated second stage or changes to source impedance. Each needs evaluation of the actual event and circuit requirements.

This example deliberately omits a commercial model because it illustrates a decision, not a product recommendation. Do not extrapolate a clamping value from a small breakdown-test current. Use a supported curve or specified value at the relevant conditions, allowing for its limits, and validate the installed node. A datasheet match on one voltage is insufficient evidence of protection.

Unidirectional vs bidirectional TVS: choose by normal voltage

The normal voltage relative to the return matters. A simple unidirectional TVS is asymmetric: reverse avalanche on one polarity and ordinary forward conduction on the other. A bidirectional device is intended to support comparable avalanche protection for both polarities, subject to its actual structure and ratings. onsemi’s TVS application note explains unidirectional and bidirectional protection behavior.

Unidirectional devices often suit positive DC rails. Bidirectional devices can suit nodes that legitimately swing above and below the reference, provided their working range and other characteristics fit. An “AC” description alone is not enough: peak amplitude, bias, signal integrity and the required protected-device voltage all matter. Some interfaces use steering arrays or more specialized structures.

Do not assume a bidirectional replacement improves protection on a unipolar rail. Its response to negative voltage can differ substantially from forward clamping by a unidirectional junction. Conversely, a unidirectional device may load a signal whose legitimate negative swing forward-biases it. Preserve the intended normal signal range and evaluate both transient polarities before substituting.

What happens during a negative surge?

The unidirectional junction can conduct forward. With the cathode on a positive rail and anode on the return, a sufficiently negative rail-to-return voltage forward-biases the diode. The resulting clamp depends on forward current, temperature and the installed loop. It is not universally fixed at −0.7 V during a high-current pulse.

Compare the relevant forward-pulse information with the protected circuit’s negative-voltage tolerance and current paths. Internal IC protection structures may conduct too, so the problem can involve both voltage and injected current. The return path must carry the event without creating unacceptable ground movement or local heating.

A TVS therefore does not replace reverse-battery protection for a sustained wiring fault. Maintaining forward conduction across a reversed source can expose it to far more energy than a short transient. A suitable series blocking or controlled power-path function may be needed. The automotive rectifier guide separates normal forward-current paths from shunt transient suppression and shows why these functions need coordinated selection.

Pulse power is tied to the specified waveform

A pulse-power label is not a continuous rating. For example, the Vishay SMBJ datasheet, document 88392 describes a 600 W class with a specified 10/1000 µs pulse waveform and derating conditions. That number does not establish the same capability for any pulse duration, repetition or ambient temperature.

Peak instantaneous dissipation is approximately P = V × I at that instant. In a purely illustrative rectangular event, 24 V and 10 A for 100 µs give 240 W peak power and about 0.024 J of energy. Real waveforms change with the source, diode behavior and time, so energy is the integral of v(t)i(t), not always that rectangle.

Neither a matching peak number nor a matching energy alone proves survival under a different waveform. Check the supported pulse data, initial temperature, repetitive operation and thermal recovery. Long supply overvoltage and automotive load-dump exposure require the relevant event assessment. The ISO electrical-load scope for vehicle equipment is a reminder that an application’s test requirement must be identified rather than inferred from a generic “surge” label.

Why layout can defeat a suitable diode

Surge-current paths have inductance and resistance. A fast current rise creates an inductive voltage L × di/dt. If the protection current travels through a long connector-to-TVS loop or shares an unsuitable reference path, the downstream circuit can see additional voltage even when the TVS terminals are clamped as expected.

For a hypothetical 5 nH path and 0.8 A/ns current slew, L × di/dt equals 4 V. The example is not a measured PCB parasitic or a forecast for every surge. It shows why a nominal several-volt clamp margin can disappear during a fast edge. Resistance and the measurement reference can add further differences.

Keep the intended transient loop short and provide a controlled return, then measure at both the clamp and the protected circuit with an appropriate setup. The Nexperia automotive ESD handbook discusses protection behavior and implementation. Its examples are application-specific; they do not make every high-power TVS suitable for a high-speed interface. Establish bandwidth, capacitance and the actual threat separately.

Coordinate sustained faults and protective devices

A TVS has a finite thermal and pulse envelope. If an overvoltage source persists, protection may need current limiting, disconnection or another coordinated function. A fuse is not automatically fast enough to save the TVS or the protected load, and a TVS failure mode should not be assumed to be reliably short-circuit in every event.

TI’s hot-swap and OR-ing TVS discussion places the clamp within a larger power-path design. Evaluate source impedance, fault duration and the ratings of switches, tracks and connectors as well as the TVS. State whether the intended system response is ride-through, shutdown or isolation, then verify that response under the applicable conditions.

For purchasing, retain the exact polarity suffix, package, voltage class and datasheet revision. Different suffix conventions are not universal. A replacement selected solely by package marking or pulse-wattage class may change the clamp or capacitance. The component sourcing workflow explains how to attach these requirements to an orderable part.

Can a multimeter establish whether the TVS still works?

A meter provides a limited fault screen. An isolated simple unidirectional TVS may resemble a diode in the forward direction and appear open in reverse if the meter’s test voltage does not reach breakdown. That result does not measure its working-voltage leakage, rated clamp or pulse capability. Bidirectional devices can appear open both ways under the same low-voltage test.

In-circuit parallel paths can obscure a short or create a misleading reading. Remove power and follow the equipment’s isolation and discharge procedure before testing. If a TVS is damaged, investigate the source event or sustained fault rather than replacing it without understanding the cause.

Application acceptance requires a defined waveform and suitable equipment. Record the source, coupling, pulse polarity, initial temperature, repetition and measurement locations. Compare the installed voltage and device stress with the selected limits. A successful clamp test under one event should not be presented as certification against every surge, ESD or automotive disturbance.