ESD Protection Devices: Placement and Return Path

An ESD protection device that is placed correctly protects a product for its whole life. The same device placed a few millimetres away from the connector, with a long ground connection, protects almost nothing, because the discharge finds a lower impedance path through the circuit before the protection device can respond. This article explains what a discharge does to a board and how the layout decides the outcome.

What a Discharge Does

Electrostatic discharge is a very fast current pulse, typically rising in under a nanosecond and lasting tens of nanoseconds, driven by a voltage that can exceed several thousand volts. It reaches a board through a connector, a button, an exposed metal part, an antenna or a seam in the enclosure, and it leaves through whatever path to ground exists.

Two models describe the event in testing. The human body model represents a person touching the board and is used mainly for component handling, while the charged device model represents the board itself being charged and then discharging rapidly when it contacts ground. The second is the more relevant one for a finished assembly, because the discharge current path is determined by the board layout.

The current does not have to flow through a semiconductor to cause damage. It can couple into a nearby trace, create a brief voltage difference across a component, or push a signal line above or below its supply rails. The results range from immediate failure to latent damage that shifts a parameter slightly and causes a field failure months later, which is the outcome that is hardest to detect and most expensive to explain.

Where Protection Belongs

The first layout rule is that a protection device must be between the source of the discharge and the circuit it protects, which in practice means at the point of entry. A connector pin, a test point or an exposed button pad should have its protection device within a few millimetres, before the signal travels anywhere else on the board.

The second rule concerns the return path. The discharge current has to reach ground, and it will take the lowest impedance route available. If the protection device is at the connector but its ground connection is a long thin trace to a distant via, the current will instead flow through the protected circuit, through the device package, or across the board. A short, wide connection to a solid ground plane is what makes the protection device the preferred path.

The third rule is separation. Protected and unprotected nets should not share a return path, and the path the ESD current takes should not run underneath sensitive circuitry. A discharge routed across a ground plane under a microcontroller will inject a voltage difference into that plane, and the sensitive device will see it as a disturbance even if the protection device operated correctly.

The Protection Options

Several device types are used, and they differ in speed, capacitance, clamping voltage and cost.

Transient voltage suppression diodes are the standard choice for signal and power lines. They clamp at a defined voltage and respond in a time short enough for most ESD events, and they are available in low capacitance versions for high speed lines. The capacitance matters: a protection diode on a USB or HDMI pair must not distort the signal, which is why purpose built low capacitance arrays exist.

Metal oxide varistors are slower and have higher capacitance, and they are generally used on power inputs rather than on signal lines.

ESD arrays combine several channels in one package, which is convenient for a connector with many lines, and they often include a steering diode structure that clamps to both rails.

Spark gaps and discharge features are built into the copper itself, as a pointed gap between a trace and a ground pad. They are used where a device cannot be fitted, or as a second line of defence, and their behaviour is less predictable than a semiconductor device.

Series elements such as a small resistor, a ferrite bead or a common mode choke slow the edge that reaches the circuit and can be used with a clamp to reduce the residual stress. They must be placed so that the clamp is at the connector and the series element is between the clamp and the circuit, not the other way round.

Getting the Return Path Right

• Place the protection device as close to the connector or exposed feature as the layout allows, on the same side of the board if possible.
• Connect the device to ground through a short, wide trace and multiple vias into the ground plane, not through a single thin via at the end of a long run.
• Route the input trace to the device first and to the circuit second, so the discharge has a path that does not pass through the protected net.
• Keep the protected trace away from the path the discharge current takes on its way to ground.
• Do not route a protected line under or beside an unprotected line for any distance, since coupling will transfer part of the transient across.
• Keep the ground return for the protection devices separate from the analog ground of a sensitive circuit, joined at a single defined point.
• Where the product has an exposed metal part that is not connected to the circuit, decide deliberately whether it is grounded or floating, and keep that decision consistent.
• Leave creepage and clearance around high voltage nodes that are exposed, since ESD is not the only source of a breakdown.

Two of these rules decide most outcomes. The connection to ground is the one that is most often compromised in a dense layout, and the separation between the protected trace and the discharge path is the one that is most often ignored.

Board Level Measures Beyond Devices

Protection is not only a component question. Several features of the board itself reduce the stress that reaches the circuit.

Guard rings around a sensitive area, connected to a quiet ground, collect charge before it reaches the devices inside. Edge clearance keeps traces away from the board perimeter, where a discharge is most likely to occur. Ground stitching around the outside of the board gives a discharge a low impedance path that avoids the interior. And a well distributed ground plane reduces the voltage difference that a fast current pulse can develop between two points on the board, which is the mechanism behind most board level ESD failures.

Conformal coating is a related measure. A coating is not a barrier against a direct discharge, because the discharge will puncture a thin film, but it does change the coupling and it raises the voltage required to create a surface flash. Where a board will be coated, our article on conformal coating covers how the choice of material and the application method affect coverage over the features that matter.

Testing and Standards

Board level ESD capability is usually verified against an international standard that specifies the waveform, the test levels and the method of applying the discharge. Two methods are used: contact discharge, where the tip touches the product, and air discharge, where a spark jumps to the product. Contact discharge is more repeatable, air discharge is more representative of a real event.

The levels are stated as a voltage, with higher voltages for harsher environments. A consumer product is typically tested to a modest level, while industrial and automotive products are tested to considerably higher ones, and the difference is reflected in the protection design rather than in the firmware.

Testing is done on the finished product rather than on the bare board, because the enclosure, the cable routing and the position of the ground connections all affect the result. That is why a design that passes at the board level can still fail when it is assembled into the enclosure, and why the layout should be pessimistic about the return path rather than assuming the worst case will not happen. Where a product fails, the usual finding is that the discharge current found a path to ground through the circuit because the intended path had higher impedance than the designer believed.

ESD Control in Manufacturing

A board that is protected in the field can still be damaged on the production line, and the measures used in assembly are different from those designed into the circuit.

The basics are well established: wrist straps and heel grounders for operators, conductive or dissipative work surfaces, ionisers where insulating materials cannot be avoided, and packaging that protects boards in transit and in storage. What matters most is that the measures are monitored rather than assumed, because a wrist strap that has failed its continuity check provides no protection at all.

Two areas are commonly overlooked. The first is the board itself when it is handled outside a protected area, such as during a mechanical assembly step or a field service visit. The second is the interface between the board and a plastic part that can hold charge, since a charged insulator discharges into the board when it is moved. Both are addressed by training and by packaging rather than by layout, and both cause failures that are attributed to the design.

FAQ

• Does every connector need ESD protection? Every connector that a user can touch, and every net that leaves the enclosure. An internal board to board connector inside a closed product usually does not.
• How far can a protection device be from the connector? As short as possible. A few millimetres is normal; a long trace makes the device ineffective because the current has already reached the circuit.
• Why does the protection device need several ground vias? To reduce the inductance of the return path. ESD current is fast, so a small inductance creates a large voltage, and that voltage is what damages the circuit.
• Can a spark gap replace a diode? It can be used where a device cannot be fitted, but its trigger voltage varies and it is less repeatable. A TVS diode is preferred where space allows.
• Is ESD a problem for boards that are not handled by users? Yes, during assembly, test and field service. Protection that is designed in also protects against handling damage at those stages.

Summary

ESD protection is a layout problem before it is a component problem. The device has to sit between the discharge and the circuit, the return path has to be a short and low impedance connection to a ground plane, and the discharge current has to be kept away from the traces it could disturb. A protection device that satisfies all three conditions will prevent failures that no amount of firmware hardening can address.

The rules are consistent with everything else in a good layout: short paths, solid references, deliberate separation between noisy and sensitive regions, and no assumption that the current will take the path the schematic suggests. Where a product will be tested to a standard, the protection design should be checked against the capability of the chosen devices and confirmed with the fabricator on the board features that support it, since a guard ring or a spark gap is a copper feature that has to be built to the right tolerance to work.